The Cognitive Underpinnings of Creative Thought

MISSING IMAGE

Material Information

Title:
The Cognitive Underpinnings of Creative Thought a Latent Variable Analysis Exploring the Roles of Intelligence and Executive Functions in Associative, Divergent, and Convergent Thinking Processes
Physical Description:
1 online resource (170 p.)
Language:
english
Creator:
Lee, Christine S.
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Educational Psychology, Human Development and Organizational Studies in Education
Committee Chair:
Therriault, David
Committee Members:
Ashton, Patricia T
Miller, M David
Abrams, Lise

Subjects

Subjects / Keywords:
cognition -- creative -- problem -- solving
Human Development and Organizational Studies in Education -- Dissertations, Academic -- UF
Genre:
Educational Psychology thesis, Ph.D.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
The importance of creative thinking is widely acknowledged ineducational and professional domains; however, the cognitive mechanismsunderlying creativity are currently not well understood. Recently, researchersin the field of creative cognition are taking an individual differencesapproach to exploring the role of intelligence and executive functions increative thought. Findings from contemporary studies of creativity are indicatingthat intelligence plays a stronger role in creative thinking than what haspreviously been reported. The purpose of this study was to contribute to thefield of creative cognition by systematically examining the relationships amongvarious cognitive factors related to intelligence (processing speed and workingmemory) and three types of creative processes (associative, divergent, andconvergent). In addition, the contribution of students’ beliefs of creativityon their performance on creative thinking tasks was also examined.   Structuralequation modeling was used to test the relationships among working memory,processing speed, and intelligence, as well as associative processing,divergent thinking, and convergent thinking. Two hundred and sixty fiveparticipants were recruited from a large southeastern university using anonline research participant pool. Participants were required to complete abattery of paper-and-pencil and computer-based tasks that assessed a range ofelementary to executive cognitive abilities linked to intelligence and creativethinking as well as a theory of creativity survey. Results showed evidence for theassociative basis in both divergent and convergent thinking processes. Findingsalso supported recent work that points to intelligence exerting a significantinfluence on creative thinking, including the associative, divergent, andconvergent creative processes explored in this study. No significantdifferences were found between students’ beliefs of creativity and theirperformance on the creative thinking tasks. Overall, findings from this study shedslight on how associative, divergent, and convergent creative processes are related,as well as the role of intelligence and executive functions in creativethinking. Recasting creativity as a higher-order cognitive process has importantimplications for future approaches to studying creativity within individuals.In addition, results from this study can be used to inform theoreticalframeworks of creative cognition aimed to better understand the fundamentalcognitive processes in creative thought.
Statement of Responsibility:
by Christine S. Lee.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
General Note:
Description based on online resource; title from PDF title page.
General Note:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
General Note:
Adviser: Therriault, David.
General Note:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-12-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044899:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 THE COGNITIVE UNDERPINNINGS OF CREATIVE THOUGHT: A LATENT VARIABLE ANALYSIS EXPLORING THE ROLES OF INTELLIGENCE AND EXECUTIVE FUNCTIONS IN ASSOCIATIVE, DIVERGENT, AND CONVERGENT THINKING PROCESSES By CHRISTINE S. LEE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Christine S. Lee

PAGE 3

3 To mom and dad, thank you for your unwavering support and encouragement over the years

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor, Dr. David J. Therriault, whose mentoring and guidance has made my graduate journey an incredible learning experience. I would also like to thank my dissertation committee of Dr. David Miller, Dr. Lise Abrams, and Dr. Patri cia Ashton for their support over the past three years as I moved from an idea to a completed research study. I am also extremely grateful to all the professors in Educational Psychology and Research Evaluation and Methodology who have imparted their wisdom, knowledge, and insights over the years. I am thankful for m y co lleagues Dr. Jenni Schelble, Dr. Ye Wang and Yujeong Park who have provided intellectual feedback and have been an invaluable support system. Finally, I would like to thank my family and f riends, who have patiently listened to my ideas, consistently provided encouragements, and celebrat ed every milestone along the journey I am tremendously grateful for all of these people who have played integral roles in my personal, academic, and profess ional growth.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENT S .................................................................................................. 4 LIST OF FIGURES ........................................................................................................ 10 ABSTRACT ................................................................................................................... 11 CHAPTER 1 INTRODUCTI ON .................................................................................................... 13 Cognition and Creative Problem Solving ................................................................ 13 Aims of this Study ................................................................................................... 14 2 REVIEW O F LITERATURE .................................................................................... 15 Creative Problem Solving ....................................................................................... 15 Cognitive Processes Involved in Creativity ............................................................. 18 Divergent Thinking ........................................................................................... 23 Convergent Thinking ........................................................................................ 29 Associative Processing .................................................................................... 33 The Relationships among Intell igence, Working Memory, Processing Speed, and Creativity ....................................................................................................... 40 Intelligence ....................................................................................................... 41 Executive Processes ........................................................................................ 44 Working Memory and Short Term Memory ....................................................... 46 Processing Speed ............................................................................................ 50 Implicit Theories and Creative Cognition ................................................................ 53 3 PURPOSE OF THE STUDY ................................................................................... 59 Exploring the Relationships among Working Memory, Processing Speed, Intelligence, Associative Fluency, Divergent Thinking, and Convergent Thinking ............................................................................................................... 59 Examining the Role of Implicit Theories of Creativity in Creative Thinking ............. 59 4 HYPOTHESES ....................................................................................................... 61 The Relationships Among Intelligence, Working Memory, Processing Speed, Associative Fluency, Divergent Thinking, and Convergent Thinking ................... 61 The Role of Implicit Beliefs of Creativity on Creative Thinking Performance .......... 63 5 METHOD ................................................................................................................ 65

PAGE 6

6 Research Design .................................................................................................... 65 Data Collection ....................................................................................................... 66 Participants ....................................................................................................... 66 Materials ........................................................................................................... 67 Creative thinking tasks ............................................................................... 67 Intelligence tests ........................................................................................ 73 Working memory tests ............................................................................... 75 Processing speed tasks ............................................................................. 77 Implicit theory of creativity scale ................................................................ 78 Procedure ......................................................................................................... 81 Data Analysis .......................................................................................................... 81 Examining the Role of Intelligence and Executive Functions in Creative Thinking ......................................................................................................... 81 Examining the Role of Implicit Beliefs of Creativity in Creative Thinking .......... 83 6 PILOT STUDY ........................................................................................................ 84 Methods .................................................................................................................. 84 Participants ....................................................................................................... 84 Measures .......................................................................................................... 84 Results of Pilot Study .............................................................................................. 85 Correlations among Observed Variables .......................................................... 85 Results of Path Analys is Exploring the Role of Working Memory and Intelligence on Associative Processing, Divergent Thinking, and Convergent Thinking ..................................................................................... 86 7 RESULTS ............................................................................................................... 89 Original Models (Using Parcel Indicators) ............................................................... 93 Model 1a: The Relationships among Working Memory, Processing Speed, and Intelligence ............................................................................................. 93 Model 1a CFA ............................................................................................ 93 Model 1a SEM ........................................................................................... 94 Model 1b: The Relationships among Associative Fluency, Divergent Thinking, and Convergent Thinking ............................................................... 96 Model 1b CFA ............................................................................................ 96 Model 1b SEM ........................................................................................... 97 Model 1: The Relationships among Processing Speed, Working Memory, Intelligence, Associative Fluency, Divergent Thinking, and Convergent Thinking ......................................................................................................... 98 Model 1 CFA .............................................................................................. 98 Model 1 SEM ............................................................................................. 98 Model 2: The Relationships among Processing Speed, Working Memory, Intelligence, Associative Fluency, Divergent Thinking, and Convergent Thinking ....................................................................................................... 101 Original Models (Using Item Level Indicators) ...................................................... 1 04

PAGE 7

7 Model 3a: The Relationships among Processing Speed, Working Memory, and Intelligence ........................................................................................... 104 Mo del 3a CFA .......................................................................................... 104 Model 3a SEM ......................................................................................... 105 Model 3b: The Relationships among Associative Fluency, Divergent Thinking, and Convergent Thinking ............................................................. 107 Model 3b CFA .......................................................................................... 107 Model 3b SEM ......................................................................................... 109 Model 3: The Relationships among Processing Speed, Working Memory, Fluid Intelligence, Associative Fluency, Divergent Thinking, and Convergent Thinking (Item Level Indicators) ............................................... 111 Model 3 CFA ............................................................................................ 111 Model 3 SEM ........................................................................................... 113 Model 4: The Rel ationships among Intelligence ( g ) as a Higher Order Factor, Associative Fluency, Divergent Thinking, and Convergent Thinking ....................................................................................................... 114 Model 4 CFA ............................................................................................ 114 Model 4 SEM ........................................................................................... 115 Model 5: The Relationships among Intelligence ( g ) as a Higher Order Factor, Associative Fluency, Divergent Thinking, and Convergent Thinking (Using Item Level Indicators) ........................................................ 116 Model 5 CFA ............................................................................................ 116 Model 5 SEM ........................................................................................... 119 The Role of Entity versus Incremental Beliefs of Creativity in Associative Fluency, Divergent Thinking, and Convergent Thinking .................................... 123 8 DISCUSSION ....................................................................................................... 127 Working Memory Predicts Processing Speed and Fluid Intelligence .................... 127 Associative Fluency Predicts Divergent and Convergent Thinking ....................... 130 Intelligence Predicts Associative Fluency, and Indirectly Predicts Divergent and Convergent Thinking ......................................................................................... 135 The Role of Working Memory in Associative Fluency, Divergent Thinking, and Convergent Thinking ......................................................................................... 140 The Role of Implicit Theories in Associative, Divergent, and Convergent Thinking Processes ........................................................................................... 142 Implications and Future Directions ........................................................................ 144 APPENDIX A DIVERGENT THINKING TEST RUBRIC .............................................................. 148 B ABBREVIATED TORRANCE TEST FOR ADULTS .............................................. 149 C ITEMS ON THE REMOTE ASSOCIATES TEST .................................................. 150 D INSIGHT PROBLEMS .......................................................................................... 151

PAGE 8

8 E THEORY OF CREATIVITY SCALE ...................................................................... 152 LIST OF REFERENCES ............................................................................................. 153 BIOGRAPHICAL SKETCH .......................................................................................... 170

PAGE 9

9 LIST OF TABLES Table page 2 1 Summary of definitions of divergent thinking, convergent thinking, and associative processing ....................................................................................... 40 5 1 Demographic breakdown of participants ............................................................ 67 6 1 Correlations among fluency, top 2 ratings, and s napshot ratings for wooden pencil, wire coat hanger, and consequences divergent thinking tests ................ 85 6 2 Correl ations among working memory tasks, intelligence tests, divergent thinking and convergent thinking tests ................................................................ 86 7 1 Descriptive statistics of observed variables ........................................................ 89 7 2 Description of latent variables and indicators ..................................................... 90 7 3 Correlations between observed variables ........................................................... 91 7 4 Summary of chi square ( 2) test, degrees of freedom, and goodness of fit indices of final models ...................................................................................... 122 7 5 Descriptive statistics of entity and incremental groups on creativity tasks ........ 124 7 6 Results of t tests between the entity and incremental beliefs groups ............... 126

PAGE 10

10 LIST OF FIGURES Figure page 4 1 Proposed structural equation model of the relationships among working memory (WM), processing speed (PS), intelligence (IQ), associative fluency (AF), divergent thinking (DT), and convergent thinking (CT) .............................. 63 6 1 Relationships among working memory (WM1, WM2), intelligence, associativ e fluency, divergent t h inking, and convergent t hinking. ......................................... 87 7 1 Relationships among working memory, processing speed, and intelligence ...... 96 7 2 Relationships among associative fluency, divergent thinking, and convergent thinking ............................................................................................................... 97 7 3 Model 1 SEM: IQ predicts associative fluency, divergent thinking, and convergent thinking .......................................................................................... 100 7 4 Model 1 SEM (insignificant paths removed) ..................................................... 100 7 5 Model 2: Relationships among working memory, processing speed, intelligence, associative fluency, divergent thinking, and convergent thinking .. 102 7 6 Model 3a: Relationships among working memory, processing speed, and intelligence (item level indicators) ..................................................................... 106 7 7 Model 3b Relationships among associative fluency, divergent thinking, and convergent thinking (item level indicators) ....................................................... 110 7 8 Model 3: Relationships among working memory, processing speed, intelligence, associative fluency, divergent thinking, and convergent thinking (item level indicators) ....................................................................................... 112 7 9 Model 4: Relationships among higher order g associative fluency, divergent thinking, and convergent thinking (item level indicators) .................................. 116 7 10 Model 5: Relationships among higher order g associative fluency, divergent thinking, and convergent thinking (item level indicators) .................................. 121

PAGE 11

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE COGNITIVE UNDERPINNINGS OF CREATIVE THOUGHT: A LATENT VARIABLE ANALYSIS EXPLORING THE ROLES OF INTELLIGENCE AND EXECUTIVE FUNCTIONS IN ASSOCIATIVE, DIVERGENT, AND CONVERGENT THINKING PROCESSES By Christine S. L ee December 2012 Chair: David J. Therriault Major: Educational Psychology The importance of creative thinking is widely acknowledged in educational and professional domains ; however, the cognitive mechanisms underlying creativity are currently not well understood. Recently, res earchers in the field of creative cognition are taking an individual differences approach to exploring the role of intelligence and executive functions in creative thought. Findings from contemporary studies of creativity are indicating that intelligence plays a stronger role in creative thinking than what has previously been reported. The purpose of this study was to contribute to the field of creative cognition by systematically investigating the relationships among various cognitive f actors r elated to intelligence ( i.e. processing speed and working memory) and three types of creative processes ( i.e. associative, divergent, and convergent) In addition, the contribution of students beliefs of creativity on their performance on creative thinking tasks was also examined. Structural equation modeling (SEM) was used to test the relationships among working memory, processing speed, and intelligence, as well as associative fluency

PAGE 12

12 divergent thi nking, and convergent thinking. Two hundred and sixty five participants were recruited from a large southeastern university using an online research participant pool. Participants were required to complete a battery of tasks that assessed a range of elementary to executive cog nitive abilities linke d to intelligence and creativ e thinking as well as a theory of creativity survey. Results showed evidence for the associative basis in both divergent and convergent thinking processes Findings supported recent work that points to intelligence exerting a significant influence on creative thinking including the associative, divergent, and convergent creative processes explored in this study There was a lack of significant differences between students beliefs of c reativity and their performance on the creative thinking tasks. Overall, f indings from this study provide a framework for the relationships among associative, divergent, and convergent creative processes as well as the role of intelligence and executive f unctions in creative thinking Recasting creativity as a higher order cognitive process has important implications for future approaches to studying individual differences in creativity Finally results from this study can be used to inform theoretical fr ameworks of creative cognition to better understand the fundamental cognitive processes in creative thought.

PAGE 13

13 CHAPTER 1 INTRODUCTION Cognition and Creative Problem Solving Creative problem solving involves the generation and application of novel approaches to complex problems that lead to powerful ideas and high quality innovations. The outcomes of creativity include major discoveries, paradigm shifts, and advancements in a field (Batey & Furnham, 2006; Runco 2007). In educational contexts, creative thinking cannot be overestimated; it is relevant for generating effective solutions for daily tasks, developing groundbreaking ideas for research and teaching, and designing innovative curricula and programs. Although the importance of creative thinking is acknowledge in instructional, learning, and professional contexts, creativity remains a construct that is not well understood in our education system and actively debated in the psychological literature (Dietrich & Kanso 2010; Plucker, Beghetto, & Dow, 2004). A relatively recent approach to examining the cognitive underpinnings of creativity is being led by researchers in the field of creative cognition who are showing that cre ativity involves both divergent and convergent thinking processes (e.g., Brophy, 2000; Finke, Ward, & Smith, 1992; Kaufman, 2009; Ward, Smith, & Vaid, 1997; Weisberg, 1995). In addition, contemporary creativity research shows that associative processes (Be nedek, Konen, & Neubauer, 2012; Mednick, 1962), fluid (e.g., Silvia, 2008a, b) and crystallized (e.g., Cho et al., 2010) intelligence, and executive functions (e.g., Gilhooly et al., 2007; Nusbaum & Silvia, 2011; Sub et al., 2002) also play a central role in creative cognition. Taken together, modern creativity research is delineating

PAGE 14

14 specific creative processes and reexamining the relationship between creative thinking and higher order cognition. Aims of this Study The broad goals of this study are twofol d. The first aim was to contribute to explore the relationships among various cognitive abilities and processes involved in creative thinking. Structural equation modeling (SEM) was used to understand the roles of working memory, proc essing speed, and intelligence in three cognitive processes of creativity: associative fluency, divergent thinking, and convergent thinking Secondly this study examined how peoples implicit theories of creativity, or their beliefs about whether creativity is a fixed or malleable ability, relate to their creative performance across various creative thinking tasks.

PAGE 15

15 CHAPTER 2 REVIEW OF LITERATURE Creative Problem Solving Creative problem solving is generally defined as seeking original ways to reach goals when the means to do so are not readily apparent (Brophy, 1998, p. 123). The creative problem solving process involves moving through ambiguous problem solving states, as more than one pathway to several possible solutions exist (Anderson, 1993; Finke et al., 1992). Creative problems requires simultaneously considering and evaluating feasible if then conditions in order to develop, determine, and follow a series of steps that will lead to a high quality solution (Anderson, 1993; Mumford et al., 1991). In contrast to traditional openended and ill defined problems that also have more than one correct solution, creative problem solving requires a novel goal state (Barnes, 1978; Finke et al., 1992; Treffinger, 1995). Therefore, a unique aspect of creative problem solving is that the solutions are judged based on degree of originality in addition to quality and appropriateness. The first models of creative processes were strongly influenced by introspective case studies of eminent individuals. The prominent methodology used by creativity researchers of this time (e. g., Albert, 1975; Cattell, 1971 ; Ellis, 1904; Galton, 1962) involved indepth studies of people who made notewor thy creative contributions to society (Becker, 1995). Self reports from individuals who were acknowledged for their creative achievements included accounts from Samuel Taylor Coleridge, author of the poem titled Kubla Khan the famous composer Wolfgang M ozart Albert Einstein, and Friedrich August von Kekule, who discovered the ring structure of benzene ( Mednick, 1962; Rothenberg & Greenberg, 1976). A common theme across the case studies was

PAGE 16

16 that creative ideas appeared with no conscious effort, revealed in dreams, altered states of consciousness, or remote analogies that emerged into their awareness (Weisberg, 1986). For example, mathematician and physicist, Henri Poincar, reported that ideas rose in crowds; I felt them collide until pairs interlocked so to speak, making a stable combination. By next morning I had established the existence of a class of Fuchsian function s. (as cited in Mednick, 1962, p. 220). As Plato had argued that an artist is only able to create from what is inspired by their m use (Sternber g & Lubart, 1996; Weiner, 2000), these early accounts from inventors, scientists, and artists suggested that creativity was an experience characterized by a sudden burst of illumination that preceded a moment of insight in which a creative solution was realized. Drawing from these reports, Wallas (1926) proposed a four stage model that served as the basis for many of the early systematic methods of studying the creative process. His four stage model of scientific insight includes a preparation, inc ubation, illumination, and verification stage ( Wallas, 1926). Briefly described, the process begins with the identification of a problem, a need, or a deficiency, then (a) preparation ensues, which involves putting forth intense effort in an attempt to sol ve a problem, such as engaging in reading, discussing, and exploring, followed by (b) incubation, a stage in which the individual temporarily gives up on the problem, and undergoes a period of unconscious mental activity, until (c) illumination is experienced as a new idea or solution suddenly and unexpectedly appears, and finally ends with (d) verification, or the evaluation and modification of the idea or solution. According to Wallas, once the four steps of the creative process were completed, a person c ould return to any of the earlier phases during the creative problem solving process.

PAGE 17

17 Elaborating upon Wallas (1926) linear four stage model, researchers have developed more complex models of creative problem solving that reflect the dynamic processes inv olved in creativity. These models include additional stages, specify different levels and subprocesses, and account for the multidirectional interactions between the levels and stages of the creative problem solving process (Isaksen, Dorval, & Treffinger, 2000; Mumford et al., 1991). The more recent theories and models of creative problem solving include Isaksen et al.s (2000) four component, eight stage creative problem solving framework, Runco and Chands (1995) twotier creative problem sol ving model, and Mumford and colleagues (1991) eight process creative solving model. Some discrepancies between these models exist (e.g., labels of any given stage); however, the three creative problem solving models share many of the following similaritie s in regard to the essential elements of the creative problem solving process. All three models include a problem identification and development stage (i.e., Understanding the problem stage, Isaksen et al., 2000; Problem construction stage, Mumford et al., 1991; Problem finding stage, Runco & Chand, 1995), an idea generation stage (i. e., Generating ideas stage, Isaksen et al., 2000; Category search stage, Mumford et al., 1991; Ideation stage, Runco & Chand, 1995), and an implementation and evaluation stage (i.e., Planning Approach stage, Isaksen et al., 2000; Idea Evaluation and Idea Implementation stages, Mumford et al., 1991; Evaluation stage, Runco & Chand, 1995). In contrast to Wallas (1926) model, the newer models of creative problem solving consist of multiple stages and portray the creative problem solving process in a cyclical and multidirectional, rather than a linear

PAGE 18

18 fashion. These recent creative problem solving models are believed to better reflect the complex patterns of thinking that occur in real life creative problem solving (Isaksen et al., 2000; Runco & Chand, 1995). In a review of studies conducted on creative problem solving, Runco (2004) provided an extensive list of studies examining the processes involved in redefi ning problems (e.g., Getzels & Csikszentmihalyi, 1975; Mumford et al., 1991), generating ideas (e.g., Guilford, 1950), synthesizing and combining old and new ideas (Finke et al., 1992; Ward et al., 1997), reorganizing information (Baughman & Mumford, 1995) and evaluating ideas (Mumford, Supinkski, et al., 1996). As research in the field of creativity continues to develop, findings from studies are increasingly highlighting the multidimensional and multidirectional nature of creative thinking. This study ai ms to contribute to this work by examining the role of factors that explain individual differences in cognition, including working memory processing speed, and intelligence, in three cognitive processes of creative thinking: associative fluency, divergent thinking, and convergent thinking. The role that implicit beliefs of creativity have in peoples performance on associative, divergent, and convergent thinking tasks will also be explored. Cognitive Processes Involved in Creativity The phenomenon known as creativity has been recognized since the beginning of civilization. Dating back to the Renaissance period, spanning roughly through the 15th to 17th century, creativity was present in the cultural movement characterized by a resurgence of learning and expression based on intellectual freedom, and a break from government and religious authority (Weiner, 2000). The artistic contributions of Leonardo da Vinci and Michelangelo, new scientific methods practiced by Galileo, and

PAGE 19

19 radical political views of Machiav elli exemplify the flood of new ideas in this era of religious, political, scientific and artistic transformation (Weiner, 2000). The Age of Enlightenment followed in the 18th century, a period in which creativity was linked to the concept of human reasoning. Progress through discovery was believed to be possible through human investigation, experimentation, and interpretation. The works of Descartes, discoveries by Sir Issaac Newton, rise of capitalism and the Industrial Revolution marked the growing confidence in humans ability to produce more in quantity as well as quality (Weiner, 2000). Continuing into the 19th century, early creativity research reflected a philosophical and generalist perspective, relying heavily on case studies or personal observat ions of highly influential creative individuals such as Charles Darwin and Francis Galton (Becker, 1995). The philosophers of this era were concerned with uncovering the basic nature of creativity by examining characteristics of eminent, creative individuals (Becker, 1995 ; Weisberg, 1986). The scientific study of creativity can be traced back to 1879, with Wundts founding of the first psychological laboratory in Leipzig. Some of the earliest notions of creativity can also be found in Spearmans (1930) Crea tive Mind in which he defined creativity as the seeing and formation of new relationships to create something that did not previously exist. Interest in the phenomenon of creativity was also present among psychoanalytic theorists, who conducted some of the ea rliest studies on human thought and behavior. Psychoanalytic theorists primarily focused on the causes behind human pathology, and creativity was also viewed from this medical model (MacKinnon, 1975). The leading psychoanaly tic theorist, Sigmund Freud, who one of the first in western civilization to undertake the study of humans ability to create, viewed creative

PAGE 20

20 production as the manifestation of the conflicts in ones internal drives, fueling the egos ability to produce something useful for oneself a nd for society (Taylor, 1975). From this psychoanalytical perspective, processes involved in creativity, including learning, thinking, solving, and curiosity were viewed as means for reducing primal drives and reducing internal tensions (Getzels & Csikszentmihalyi, 1975). The development of a new field of psychology, called humanistic psychology, offered a more optimistic and favorable perspective of the study of human behavior. The view that humans existed primarily to reduce innate conflicts was replac ed with the belief that humans have a need to create and therefore, voluntarily seek out new experiences that provided opportunities to work with novel ideas, exercise initiative, and express a persons unique character (Getzels & Csikszentmihalyi, 1975). Early Gestalt psychologists, whose philosophy is often represented by the phrase the whole is greater than the sum of its parts, adopted a holistic perspective that characterized creative thinking as a deep analysis of a field in order to identify gaps that needed to be resolved (Duncker, 1926, 1945; Kohler, 1927; Wertheimer, 1945). Once the gaps were identified, creativity was required to restructure the field (e. g., changing functional meanings, grouping, and organizing) in order to restore harmony (a greater whole ) Gestalt psychologists also believed that creativity involved actions that produce new ideas or insights through imagination rather than r eason and logic (Taylor, 1975). T h e Piagetian developmental theory model was also replacing earlier psychoanal ytic and behaviorist frameworks, and as the field of psychology broadened to include theories of humans as active organisms initiating experiences with their world, creativity incr easingly became a topic of interest (Getzel & Csikszentmihalyi, 1975).

PAGE 21

21 Many researchers cite Guilford s 1950 presidential address to the American Psychological Association as time that creativity came under the lens of scientific inquiry At this farewell address, Guilford advocated for the scientific community to apply experimental methods to the study creativity via divergent thinking tests Further discussed below, Guilfords distinction between divergent and convergent thinking has had a surprisingly la rge and long lasting influence in creativity research. Guil ford (1950, 1957, 1966, 1967) believed that creative ability was an inherent part of human nature, and he promoted a psychometric approach to examining the mental functions that contribute to creat ive production among everyday people. This general ist view and approach to studying creativity was in stark contrast to the popular views of creativity as something of a genius quality. In his address, Guil ford posed the following questions, Why is creati ve productivity a relatively infrequent phenomenon?, Why is there so little apparent correlation between education and creative productiveness?, Why do we not produce a larger number of creative geniuses than we do, under supposedly enlightened, moder n educational practices? (Guil ford, 1950, p. 444), highlighting the need of a better understanding of creativity, especially in educational practices. The challenges presented to researchers in Guilford s address is often cited as an example of zeitgeist (i. e., spirit of the times), as they were compatible with the assumptions and needs of that era (Getzel & Csikszentmihalyi, 1975; Weiner, 2000). During this period in history, the United States was entering a new era of economic surplus following World War II. Scientific research was pursued with zest, resulting in discoveries including penicillin, space exploration, and atomic explosives (Barron, 1988; Getzel & Csikszentmihalyi, 1975). These developments created a sense of endless

PAGE 22

22 possibilities in humans intelligence and technological advancements, spurring social interest in creativity research which was increasingly viewed as an essential part of moving towards a greater future (Barron, 1988; Getzel & Csikszentmihalyi, 1975; Weiner, 2000). This ideal was in line with Guilfords argument that a psychometric approach to studying creative potential among everyday people should be taken in lieu of studying only exceptionally creative individuals who are rare and difficult to investigate. There was a surge of creativity research in the 50s and 60s, exploring many different facets of creativity. While only about 16 journal articles containing the word creativity existed at the time of Guilfords address, the annual count rose to 56 by 1959 and to approximately 328 in 1999 (Sternberg & Dess, 2001). The continued interest in creativity is further illustrated by the launching of two journals devoted to creativity research ( Journal of Creative Behavior 1967 and Creativity Research Journal 1989), and the National Science Foundation now funds research on a wide range of creativity research topics (Sternberg & Dess, 2001). Creativity research has spread into diverse avenues of study, including applications of frameworks from intelligence (e.g., Cattell 1971), personality (e.g., Eysenck, 1995; MacKinnon 1962, 1975; MacKinnon & Hall, 1971), cognition (e.g., Finke et al., 1992; Osborn, 1953), social (e.g., Amabile, 1996) and cultural ( Csikszentmihalyi, 1999) research The purpose of this study is to contr ibute to contemporary research in the field of cognitive psychology applied to creativity, by exploring the relationships among the cognitive underpinnings of creative thinking and intelligence. In the following sections, the major constructs under investi gation will be discussed in detail.

PAGE 23

23 Divergent Thinking Divergent thinking is an inductive, ideational process that involves generating a broad range of solutions or as many ideas as possible to a given stimulus. It is often contrast ed with convergent thinking, a deductive process that involves systematically applying rules, organizing ideas, and evaluating options, in order to arrive at a single, correct solution (Brophy, 1998; Guilford, 1967; Kaufman et al., 2011). Guilford (1967) distinguished between di vergent thinking and convergent thinking in his structure of intellect (SI) model, emphasizing divergent thinking as a critical process underlying creativity, and categorizing convergent thinking as a process relevant for success on traditional academic and cognitive tests. Accordingly, Guilford developed a series of pencil andpaper divergent thinking tests, including the Consequences Test (e.g., Imagine everyone lost the ability to read and write. What would happen as a consequence?), and the Unusual Uses Test (e.g., Think of as many unusual uses as possible for a wooden pencil) (Guilford, 1967; Guilford, Merrifield, & Wilson, 1958). Since Guilford, a battery of standardized divergent thinking tests has been developed, such as the Torrance (1966) Tes ts of Creative Thinking (TTCT), another widely used divergent thinking test that includes verbal and figural components. A common feature among divergent thinking tests is the requirement to emit a large number of responses to a series of ill defined problems that are scored along various indicators of creativity including, fluency (i.e., the number of ideas) originality (i.e., uniqueness or novelty of the ideas) and flexibility (i.e., the number of different types of categories) (Batey & Furnham, 2006; Dietrich & Kanso, 2010; Plucker & Runco, 1998). Divergent thinking tests were originally developed to measure individual differences in divergent thinking ability (contrasted with convergent

PAGE 24

24 thinking, a noncreative process) but have since become the primary psychometric measure of creativity, dominating theoretical and empirical work in this field (Dietrich & Kanso, 2010) The contributions of this psychometric revolution opened the door of creative investigation to the masses, by providing a standardized method to gather and score creativity data (Sternberg & Lubart, 1996, p. 861). To many researchers surprise, the use of pencil andpaper divergent thinking tests adopted in the 1950s continues to be the most commonly used method for assessing creat ivity today (Barron & Harrington, 1981; Batey & Fur nham, 2006 ; Dietrich & Kanso, 2010; Plucker & Runco, 1998); t he majority of current approaches to assessing creativity are the same or very similar to the methods proposed over fifty years ago ( Runco, 2003 2007). In a review of creativity measurement, Plucker and Runco (1998) state, psychometric studies of creativity conducted in the previous few decades form the foundation of current understandings of creativity (p. 36). The methods for measur ing cre ativity suggested by the pioneers of creativity research (e.g., Guilford 1950; Mednick, 1962; Torrance, 1966; Wallach & Kogan, 1965) are fairly straightforward, contributing to their appeal for experimental studies. For example, a series of Wallach and Kogan s (1965) creative thinking tasks require only one judge to score answers to the divergent thinking tests based on a fluency score (i.e., tally of the number of ideas generated) and an originality score of 1 for unique ideas and 0 for all others. In another example, t he scoring me thod for the Torrance Tests is based on a normative samples pool of common ideas, and points are given for ideas outside of the norm (Torrance, 2008). Scoring methods vary slightly across

PAGE 25

25 different divergent thinking tests, but all of them follow a clear, standardized procedure that awards points for statistically uncommon responses which are summed to produce a final creativity score. Although widely used in creativity research, reports on the validity of diverg ent thinking tes ts are mixed. A qualified statement by Bar ron and Harrington (1981) in their review of creativity research represents the controversy surrounding divergent thinking tests some divergent thinking tests, administered under some conditions and scored by some sets of criteria, do measure abilities related to creative achievement and behavior in some domains. (p. 447, italics added). Runco and Chand (1995) classify three camps regarding the relationship between divergent thinking and creativity. One camp v iews divergent thinking as synonymous with creativity, a second camp take takes the completely opposing view that divergent thinking has little to do with creative thinking, and the last camp take an intermediary position, viewing divergent thinking as an estimate of one of many indicators of creativit y. Researcher have argued that a downside to using divergent thinking tests is that creativity i s been reduced to the objectively defined statistical rarity of a response with regard to all the responses of a subject population (Sternberg & Lubart, 1996, p. 861). Since the beginning of the divergent thinking testing movement, many researchers have criticized that such tests oversimplify the concepts of fluency and flexibility to merely generating a large amount of different ideas to unrealistic situations. In his book Genius and the Process of Creative Thought Cattell (1971) argues that a persons fluency and flexibility in generating ideas needs to be taken into account with the interactions among abilities, personality traits, and interests to fully understand the creative person. In

PAGE 26

26 addition, he argues that creativity is a cumulative process larg ely dependent on the frequency and degree of intensity a person returns to a task in order to reach a creative achievement. Cattell (1971) cites eminent creative achievements such as Edisons discovery of the electric light bulb and Keplers realization that planets move in ellipses rather than circles, as examples of the cumulative process of creativity that requires not only fluency and flexibility in the development of ideas, but also important personality and motivation factors. Along the same line, i n a widely cited review of psychometric research in creativity, Wallach (1971) stated that, subjects vary widely and systematically in their attainments yet little if any of that systematic variation is captured by individual differences on ideational flu ency tests (p. 60). Finally, s ome researchers argue that divergent thinking tests only tap into lower level creative abilities, differentiating between little c (everyday creative acts) and Big C (creative acts that have a significant impact within a field) and even making a case for mini c (novel and personally meaningful interpretation of experiences, actions, and events) (Beghetto & Kaufman, 2007; Simonton, 2000). In a recent review of creative research, Runco (2004) points out that while psychom etric pencil and paper tests have been widely used among ordinary individuals, such studies have not been replicated with eminent individuals who have received recognition for their creative contributions. Furthermore, psychometric studies of creativity us ing divergent thinking tests conform to a normal distribution, whereas studies of renowned creative achievements represent a skewed distribution of a select group of people (Batey & Furnham, 2006). Studies interested in the latter group have employed more qualitative methods such as c ase studies (e.g., Ellis, 1904; Galton,

PAGE 27

27 1962) to investigate creativity. Therefore, there exists a discrepancy regarding the population targeted and method employed by creativity researchers that surrounds the age old issue of how creativity is defined (e. g., little c versus Big C). On the other hand, e vi dence for the validity of divergent thinking tests include positive and statistically significant relationships between divergent thinking test scores and various indicators o f real life creative behaviors at the elementary, junior high school, undergraduate, and graduate levels (Barron & Harrington, 1981) Some validation studies have also shown that divergent thinking tests are highly correlated with real world creativity inc luding number of patents gained, producing plays and novels, and founding new businesses or professional organizations (Plucker, 1999; Runco, 1991; Torrance, 1968, 1969, 1972). Studies conducted among adult professionals such as Air Force captains (Barron, 1955) and sales women (Wallace, 196 1) also showed that performance on divergent think ing tests correlated strongly with rated originality and problem solving in real life situations ( r = .30, p < .01 for scores on the Unusual Uses test, and r = .36, p < .01 for scores on the Consequences test). More recent evidence for the predictive validity of divergent thinking tests has also been found. For instance, Hong, Milgram, and Gorsky (1995) found that 60 second grade students performance on measures of creative thinking, assessed by ideational fluency on the Unusual Uses and Pattern Meaning tasks (adapted from Wallach & Kogans 1965 original study) and two problem solving tasks including the Box and the Chair task s, wer e significantly related so scores on the Tel Aviv Activities Inventory: Primary Grades which measures creative activities outside of school in domains including art, music, sport, drama, literature, and dance ( r = .41) In addition, originality

PAGE 28

28 scores were not related to s cores on the Informa tion and Vocabulary subtests of the Wechsler Intelligence Scale for Children (WISC) (Wechsler, 1974) providing evidence for the discriminant validity of the divergent thinking tests Similar results have been shown in studies among early childhood (e.g., Hong & Milgram, 1991; Milgram, 1983; Milgram et al., 1978; Milgram & Arad, 1981) and adolescent populations (e.g., Hong, Milgram, & Whiston, 1993; Hong, Whiston, & Milgram, 1993; Milgram, 1990; Milgram et al., 1978) Finally, Plucker (1999) used SEM to reanalyze Torrances (1968) 20year data from a longitudinal study of over 200 elementary students The data set included students divergent thinking scores on the Torrrance Test of Creative T hinking, IQ scores from several tests of intelligence and real life creativity scores, represented by an estimate of the quantity of publicly recognized creative achievemensts (e. g., inventions, awards, published articles) (Plucker, 1999) Results showed that divergent thinking strongly predicted creative achievement s ( r = .60, p < .001), even when accounting for the role of intelligence in creative achievements. In addition, divergent thinking explained almost half of the variance in adult creative achievement, contributing more than three times that of intelligence. Divergent thinking and intelligence were correlated at r = .20 (Plucke, 1999). Altogether, these findings point to the predictive power of divergent thinking tests, especially in comparison to intelligence tests. Divergent thinking tests continue to be the most widely used measure for assessing creative thinking (Batey & Furnham, 2006; Runco, 2010). Nevertheless, the sole use of divergent thinking tasks to assess and make subsequent all inclusive conclusions about an individuals creative potential is not wholly accepted.

PAGE 29

29 Convergent Thinking Convergent thinking was popularized by Guilford (1967) who presented it as the antithesis of creative thought (in contrast to divergent thinking). Convergent thinking tests measure cognitive processes such as discerning whic h ideas are most appropriate or of highest quality with the objective of arriving at a single, correct solution (Guilford, 196 7; Brophy, 2000). More recently, researchers are arguing that convergent thinking complements divergent thinking in the cr eative process ( e.g., Broph y, 2000, Dietrich & Kanso, 2010; Finke et al., 1992; Cropley, 2006). However, c ompared to divergent thinking, much less attention has been paid to the convergent processes in creative thought Creativity tasks that engage conver gent thinking processes include the Remote Associates Test (RAT, Mednick, 1962) as well as various insight problems, such as Dunckers (1945) candle problem. The process of finding the solution to these problems is often referred to as thinking outside of the box, as the problem solver is required to break away from obvious responses and view the probl em from a different perspective within a novel search space where the solution typically resides (Wiley, 1998). For example, the RAT item s consist of a tria d of cue words that are not directly related to each other, but rather, are related to a common associate fourth word, either through semantic association, synonymy, or formation of a compound word (e.g., bird, tie, pen black). Often, the most common ass ociates to each cue word (the words that first come to mind) are not related to the other cue words. Therefore, identifying the correct associate word requires the problem solver to suppress the strongest associates, and search for remote associates of t he three cue words (Mednick, 1962).

PAGE 30

30 Arguably, the process of generating r emote associates is also involved in divergent thinking, as greater performance on divergent thinking tests require the suppression of common responses in order to generate more novel and unusual responses. However, what distinguishes convergent tests of creativity from divergent thinking tests is that success on tests such as the RAT is determined by whether or not the problem solver has identified the single, correct solution ther efore, making the RAT and insight problems convergent in nature. The use and interpretation of peoples performance on the RAT and insight problems vary across studies. Mednick (1962) developed the RAT based on his theory of associative processing in creativity to assess individual differences in associative hierarchies ( discussed in more detail below); however, the RAT has been used in several studies to assess creativity generally, and some researchers have even conceptualized the RAT as a creative probl em solving task (e.g., Cushen & Wiley, 2012). Recently, the RAT has also appeared in studies involving memory (e.g., Storm, Angello, & Bjork, 2011). Within the creativity literature, the distinction between divergent thinking tests and convergent creativit y has not been thoroughly acknowledged or empirically examined. This gap in the literature is important to address, given that existing creativity tests are arguably tapping into different cognitive processes It is important to pull apart the distinct cog nitive mechanisms involved in creative thinking in order to better understand this complex construct. Studies have shown that the RAT correlates highly with IQ tests and it has been suggested that the RAT measures cognitive processes similar to those measured by intelligence tests (Kaufman et al., 2011; Ward, 1971). In fact, early objections to

PAGE 31

31 interpreting the RAT scores as a measure of creativity have been raised on both concep tual and empirical grounds. T here is mixed evidence regarding whether differences in performance on the RAT relate to individual differences in associative abilities as originality proposed by Mednick, or executive functions more closely related to intelligence (Mendelsohn, 1976). In support of the latter, several studies showed that RAT performance was not related to the number of associations produced for a given stimuli (Yahav, 1965), weakly related to pairedassociate learning ( r = .19 between measure of associative mediation and RAT scor e) (Greenberg, 1966), and unrelated to associative processes in a concept formation task (Jacobson et al., 1968). Taft and Rossiter (1966) also raised the issue of the convergent nature of the RAT test items. They showed that with few exceptions, convergent thinking tests, including school achievement, performance on verbal IQ, quantitative IQ, progressive matrices, speed and accuracy, and number series tasks, correlated considerably higher with the RAT ( r = .57, .60, .46, .38, .27, .41, and .40, p < .01, r espectively) compared to the scores on tests of divergent thinking including ideational fluency, word fluency, and total fluency, flexibility, and originality scores on unusual uses and consequences tests ( r = .15, .43, .15, .15, and .27, p < .01, respecti vely). Similar studies have also shown moderately positive correlations between scores on the RAT test and conventional measures of intelligence ( r = .2 to .5) (Laughlin, Doherty, & Dunn, 1968). An alternative interpretation of the cognitive mechanisms involved in solving the RAT highlight the role of attentional processes in making remote associations. Mendelsohn (1976) showed evidence that when controlling for verbal intelligence, performance on the RAT is related to individual differences in attentional processes,

PAGE 32

32 assessed by the employment of multiple category sets in solving anagrams (Mendelsohn, 1976). In his study, participants were assigned to either an informed condition, in which they were told that the anagrams fell into an animal, food, or misce llaneous category, or an uninformed condition, in which they were simply instructed to solve the 30 anagrams. Mendelsohn proposed that individuals who scored high in the informed condition were able to do so based on the assumption that the memory search a nd retrieval processes involved in solving the anagrams also required the ability to simultaneously and effectively maintain the three search strategies (i.e., consider animals, foods, and miscellaneous words). Based on his findings that high scores in the informed anagram condition related to high scores on the RAT, Mendelsohn (1976) argued that attentional processes such as the ability to receive and store in accessible form a broad range of information from the environment would serve to increase the ra nge of elements, including unusual peripheral, or incidental elements (p. 363) and facilitate the ability to find the single, correct mediating link (or remote association) to solve the RAT problem. Despite the early controversies regarding the cognitive mechanisms and abilities assessed by the RAT, the majority of the creativity literature has used and interpreted the RAT as a measure of general creative ability based on Mednicks (1962) theory of the associative basis of creativity. Not until recently has the RAT been recategorized as a convergent thinking test of creativity (e.g., Benedek et al., 2012; Kaufman et al., 2011; Nielsen, Pickett, & Simonton, 2008). In addition, convergent thinking has recently received more attention in the creativity liter ature, and is increasingly being highlighted as an important aspect of the creative problem solving process (rather than a process

PAGE 33

33 that is opposite or counterproductive to creativity) (Brophy, 2000; Cropley, 2006). Especially in the creative problem solving literature, convergent processes, such as ability to choose the best and most useful ideas and to notice when adjustments in ideas generat ed or paths taken are necessary, have been cited as an important part of effectively judging and adapting ideas generated in order to achieve a high quality novel and appropriate solution (Brophy, 1998; Kaufman et al., 2011; Treffinger et al., 2002). Associative Processing The study of associative processing in creativity emphasizes the recombination of existing elements into novel products. Associative processing involves the activation of mental networks made up of related (or associated) concepts and ideas. The traditional associative view of creative thinking suggests that any given stimulus will cue an idea, which will then cue another idea, until this spread of activation reaches distally related ideas. It is believed that creative ideas occur when more uncommon elements lower in the associative hierarchy are generated. There are several explanations regarding the complex interactions among individual difference factors that in fluence associative processing, beginning with the cognitive mechanisms that initiate the activation of associative elements to those that lead to the achievement of a creative outcome. Medni cks (1962) theory of associative processes has been influential in viewing associative abilities in creative thinking from an individual differences perspective. According to Mednick (1962), creative individuals have associative hierarchies (the gradient of associative response strength for available associations) that are flat allowing them to make many associations between remote ideas, form associative elements into novel combinations, and generate creative and useful solutions. Conversely, less creativ e individuals have steep associative hierarchies that

PAGE 34

34 result in fewer and more common associations (Mednick, 1962). More specifically, for individuals with steep associative hierarchies, the associative strengths of stereotyped responses (e.g., table cha ir) are dominant, such that once conventional responses to the stimulus have been generated, the associative strengths of less common responses are weak. Ultimately, this results in accessing a smaller number of unusual ideas, leavi ng the individual with a set of common ideas from which to develop creative solutions. Individuals who have flatter associative hierarchies may have the same strongest response to the conventional associates, but the associative strengths to more remote associates decline more st eadily, allowing them to access more uncommon ideas (Mednick, 1962). In support of Mednicks (1962) theory, a study examining individual differences in associative processing showed that there was a negative and significant relationship between peoples judgment of the associative distance between two stimuli words and their originality score on a series of creativity tests ( r = .22, p < .05) (Rossman & Fink, 2010). In other words, more creative individuals judged the associative distance between two unrelated words to be smaller than less creative individuals ( Rossman & Fink, 2010). Similarly, studies have shown that shorter associative pathways between unrelated words were related to magical ideation and paranormal beliefs (Gianotti et al., 2001; Mohr et al., 2001). Results from these studies suggest that individuals with flexible associative networks are more likely to generate and combine remote and distal associations, increasing the likelihood of generating novel solutions. Eysenck (1995) propo sed that lower levels of cognitive inhibition (poorer ability to inhibit irrelevant information) support associative processing. He argued that cognitive

PAGE 35

35 disinhibition and allusive or over inclusive thinking (e.g., considering seemingly unrelated informati on) contri butes to cognitive flexibility so that accessible ideas in long term memory cue a broader range of uncommon (and potentially novel) ideas in a persons semantic network (Eysenck, 1995). Similarly, Martindale (1995) suggested that creative associations are made during primary processing, characteri zed by defocused attention and free associative thinking. However, Martindale (1999) also argued that as opposed to being in a permanent state of defocused attention, creative people have a tendency to os cillate between primary and secondary processes, the latter of which are characterized by logical and analytic thinking. This echoes other theories of creativity, including Wallas (1926) classic four stage model, as well as Finke, Ward, & Smiths (1992) Geneplore Model which consists of a generative (primary processes) and exploratory (secondary processes) phase. Csikszentmihalyis (1999) seminal work also showed that creative individuals maneuver through a broad range of dimensions such as moments of extraversion and introversion, and masculinity and femininity. Finally, frameworks of creative problem solving processes distinguish among a series of divergent (e.g., fluency, elaboration) and convergent (e.g., evaluation, synthesis) processes (e.g., Mumford et al., 1991; Treffinger et al., 2000; 2002) In theory, associative processing is involved in both divergent and convergent thought processes. For example, the activation and retrieval of remote associations is likely to support divergent processes where the goal is to generate many unusual solutions (e.g., Think of as many uses for a brick as possible). On the other hand, the ability to initiate a wider associative spread and access remote concepts is also likely to promote success on a convergent creat ive thinking task such as the RAT where the

PAGE 36

36 goal is to identify a solution that is distally r elated from the original stimulus There are several explanations attempting to delineate the underlying cognitive mechanisms of divergent and convergent thinking, many of which relate to associative processes. For instance, it has been proposed that insight or progress towards a solution is guided by implicit spreading activation (Bowers, Farvolden, & Mermigis, 1995), incremental steps in which the problem solver i s building upon existing knowledge (Ward, 1995; Weisberg, 1995), unconscious activity that can be disrupted by simultaneously engaging in explicit tasks such as verbalization during the problem solving process (Dominowski, 1995), and attentional processes as well as associational and ideational fluency (Mendelsohn, 1976). However, not until recently has the role of associative processes in creativity been directly tested. A study conducted by Benedek et al. (2012) aimed to address this gap in the literatu re by examining the role of associative processing, which they defined as the ability to fluently retrieve and combine remote associations, with respect to divergent thinking and intelligence. Four association tasks were developed to measured associative f luency (i.e., ability to generate a large number of associations), associative flexibility (i.e., ability to create a n associationchain in which each word generated is associated only to the word that precedes it), dissociative ability (i.e., ability to g enerate lists of unrelated words), and associative combination (i.e., ability to generate a word that is associated with a pair of unrelated words). Performance on the four association tasks were significantly correlated with performance on the Alternative Uses divergent thinking test ( r = .55 to .62, p < .01), and with intelligence ( r = .09 to .20, p < .05). Results from a regression analysis showed that the four association measures

PAGE 37

37 explained almost half ( R2 = .47) of the variance in divergent thinking ability, with p < .001) and associative p < .05). Results from SEM also showed that these two p < .001 for dissociative abiltiy p < .05 for associative combination), but had no significant relationship to intelligence. Associative flexibility was the only associative p < .05). Overall, results from this study contribute to our understanding in two important ways. Firstly, this is the first known study to empirically explore associative processes by separately investigating four different aspec ts of associative ability. S econdly, this study provi des some insight into how associative processes relate to divergent thinking and intelligence. Benedek et al.s (2012) findings indicate that whereas associative flexibi lity predicts intelligence, di ssociative ability and associative combination is predict ive of creativity. A possible reason given for the lack of predictive power in associative fluency for both creativity and intelligence was that a substantial part of the variance in associative fluency may have been accounted for by the latter three more complex associative processes (Benedek et al., 2012). Kaufman and colleagues (2009) explored the roles of associative learning, working memory, and processing speed in intelligence. They defined associative learning as the ability to remember and voluntarily recall specific associations betw een stimuli (p. 374). A difference worth noting is that associative learning in this study was operationalized as a more complex, executive process compared to the four associative processes explored in Benedek et al. s (2012) study. The associative learning task

PAGE 38

38 required participants to engage in explicit and voluntary storage and retrieval of associations, in contrast to other types of association tasks where the solutions are experienced via a more implicit and sud den process (Kaufman et al., 2009). It was proposed that associative learning would predict intelligence, based on early and recent research that indicated that intelligent behavior involved memory patterns of associations among stimuli (e.g., Binet & Simon, 1916; Williams & Pearlberg, 2006). As expected, results from SEM showed that associative learning, working memory, and processing speed all made statistically significant contributions to intelligence ( = .33, .34, p < .01, .25, p < .05, respectively). Results from this study suggest that associate learning may be a possible contributor to the variance in intelligence. Finally, research in problem solving has shown that at people engage in a preliminary task analysis before engaging in the problem sol ving activity at hand. This task analysis involves searches of long term memory through episodic and semantic knowledge for relevant information or ideas that can be applied to the task (Ericsson & Simon, 1993). Findings from a think aloud study conducted by Gilhooly et al. (2007) showed evidence that participants engaged in unmediated associative processes that paralleled the preliminary task analysis processes reported in the problem solving literature. Analyses of the think aloud protocols showed that on an unusual uses divergent thinking test, the initial processes verbalized by participants consisted largely of unmediated retrieval from long term memory (e.g., A tire could be used as a flotation aid). More complex processes including scanning the properties of the item to trigger uses (e.g., Bricks are heavy), generating a broad range of possible uses (e.g., Artwork, weapon, tool, and disassembling the items (e.g., R emove the shoelaces

PAGE 39

39 from the shoe) occurred later in task (Gilhooly et al., 2007 p. 615). Furthermore, results showed that unmediated retrieval w as strongly correlated with fluency score s (i.e., number of ideas generated) ( r = .73, p < .01) whereas the latter processes, including generating broad uses and disassembling items, were more strongly related to novelty or originality scores ( r = .42, .62, p < .01, respectively). Multiple regression analyses also showed that both memory search and disassembly made significant contributions to predicting the novelty score on the unusual uses t asks ( = .37, .57, p < .01, respectively). Overall, results from this study showed that a memory search in which unmediated ideas are fluently retrieved is a precursor to more complex cognitive processes involved in generating and developing novel ideas. These findings support the hypothesized model in the present study in which the construct associative fluency, assessed by semantic category and letter fluency tasks, is specified as a predictor of divergent and convergent thinking two m ore complex cognit ive processes related to generating novel ideas and identifying remote associations that occur later in the creative thinking process Altogether, research is needed to identify the different creative thinking processes reported in the literature, including the many associative processes that have been theorized to underlie creative thinking. In this study, the relationships among associative flue ncy, divergent thinking, and convergent thinking were explored. Table 2 1 summarizes the definitions of the three creative thinking processes provided in the review of the literature described above.

PAGE 40

40 Table 2 1. Summary of definitions of divergent thinking, convergent thinking, and associative processing Divergent thinking Convergent thinking Associative processing Brainstorming (Osborn, 1953) Generating many possibilities (Guilford, 1967) Goal oriented reasoning in different directions that takes into account a variety of aspects (Dorner and Kreuzig, 1983) Idea generation phase of creative process (Finke, Ward, & Smith, 1992) Unstructured thinking; spontaneous processes with multiple explanations (Finke, 1996) The nature of tests where going off in multiple directions to obtain multiple answers increases the scores (Goff & Torrance, 2002) Producing many different answers (Runco, 2007) Synthesis and analytic ability; breaking down symbolic structures in perceptual and conceptual domains (Guilford, 1950) Anticipating the functional characteristics of generated ideas (Guilford, 1967) E valuation and selection of adaptive ideas (Sternberg, 1985) Systematic reasoning directed towards on correct answer (Dorner and Kreuzig, 1983) Exploration phase; structured aspects of creative thinking that involves conscious deliberate control (Finke, W ard, & Smith, 1992) Problem solving that is goal directed and requires a single correct answer (Runco, 2007) Working towards concrete goals (Nielson, Pickett, Simonton, 2008) Building connections between elements of associative network (Mednick, 1962) Ability to access and retrieve broad range of elements (e.g., words needed for a solution) by maintaining and bringing into continuity separate sequences of mental activity (Mendelsohn, 1976) Formation of uncommon associations as a result of defocused attention and lowered cognitive inhibition (Eysenck, 1995) Ability to oscillate between primary (e.g., analogical, freeassociative) and secondary (e.g., logical, analytic) processes (Martindale, 1995) Involved in the ability to remember and voluntarily recall specific associations between stimuli (e.g., learn multiple responseoutcome contingencies) (Kaufman et al., 2011) Ability to fluidly retrieve and combine remote associations (Benedek et al., 2012) The Relationships among Intelligence, Working Me mory, Processing Speed, and Creativity Spearman (1904) developed a series of cognitive tests that were found to positively correlate with one another, and since then, many tests of individual differences have been developed that load highly on a general factor (Spearmans g ) in psychometric studies of intelligence (e.g., Ravens Matrices; Cattells Culture Fair Test)

PAGE 41

41 (Carroll, 1993, Jensen, 1998). Although the construct of general intelligence ( g ) has been well established, the underlying mechanisms of g and the inter relationships among them, are less clear. Currently, research shows working memory and processing speed as the best subs t rate s of g (Conway et al., 2002; Jensen, 1998). Sternberg (1985) offered two widely used definitions to contrast creati v ity and intelligence. He defined creativity as the process of bringing into being something novel and useful, and intelligence as the ability to purposively adapt to, shape, and select environments (Sternberg, 1985) (as cited in Sternberg & OHara, 1999, p. 251). In other words, creativity is traditionally defined as the development of a novel and useful product, and intelligence is defined as a general reasoning or learning ability assessed by complex cognitive tasks. Until recently, researchers of cr eativity and intelligence represented the two constructs as consisting of distinct sets of abilities. Classic studies (e. g., Getzels & Jackson, 1962; Wallach & Kogan, 1965) exploring the relationship between creativity and intelligence showed evidence tha t separable sets of cognitive abilities, career aspirations, and interests existed between a group of students who scored high on traditio nal measures of intelligence compared to another group of students who scored high on measures of creativity. In addit ion, correlations between intelligence and creativity scores were small, ranging between .09 and .26 (Wallach & Kogan, 1965) and r = .17 in a more recent metaa nalysis (Kim, 2005). Intelligence Intelligence is a multifactorial construct consisting of vario us cognitive abilities that are traditio nally measured using standardized intelligent quotient (IQ) tests (Barron & Harrington, 1981; Batey & Furnham, 2006). Many views on the relationship between creativity and intelligence exist. Early models of intellec tual abilities generally placed

PAGE 42

42 creativity as a subset of intelligence (e.g., Cattells (1971) model of fluid and crystallized intelligence ; Structure of Intellect Model, Guilford, 1967; Berlin Intelligence Structure, Jager, 1982, 1984). Sternberg and Lubarts (1996) Investment Theory describes six components of creativity, specifying intelligence as one of the six subsets that make up creativity. Other empirical studies showed moderate relationships between creativity and intelligence measures, suggesting that intelligence and creativity are separate constructs with overlapping features (e.g., Cox, 1926; MacKinnon, 1965). Finally, some researchers proposed a nonlinear relationship between creativity and intelligence (e.g., Barron, 1963; MacKinnon, 1978). Ac cording to what researchers have labeled the threshold theory, creativity and intelligence are moderately ( r = .30 .40) related up to an IQ of approximately 120 (Barron, 1963; MacKinnon, 1978) However, at higher levels of IQ (beyond 120) there is consi derable variance in creative performance, and int elligence explains little of this variance in creative achievement (Barron, 1963; Guilford, 1981; MacKinnon, 1978). One explanation for the threshold theory is that intelligence is a necessary, but not suff icient, condition for creative achievement. S tudies have tested the threshold theory in different domains, including mathematics (MacKinnon, 1978), military (Barron, 1963), and ar chitects (MacKinnon, 1961), and findings regarding the threshold theory are mixed. Overall, creativity researchers continue to view creativity and intelligence as unitary constructs that are modestly related at best (Batey & Furnham, 2006; Kaufman, 2009; Kim, 2008 ; Runco, 2007). Exactly how these two constructs are related remains an area of contentious debate. A methodological limitation of the earlier studies of creativity and intelligence was that conclusions were almost entirely based on correlational analyses among various

PAGE 43

43 measures of intelligence and creativity. Recent studies using more advanced statistical methods to reexamine the relationship between intelligence and creativity indicate that these two constructs may be more strongly related than previously believed (e.g., Nusbaum & Silvia, 2011; Plucker, 1999; Silvia, 2008 a, b ). For example, in a latent variable reanalysis of Wallach and Kogans (1965) classic study, Silvia (2008) found that creativity and intelligence were more highly correlated than reported in the original study. Compared to a correlation of r = .09 repo rted by Wallach and Kogan (1965), a latent creativity factor based on 10 different creativity scores was significantly related to a latent intelligence factor based on 10 intel ligence and achievement scores at = .20 Similarly, Batey, Furnham, and Safiul lina (2010) found divergent thinking test scores were significantly related to fluid intelligence ( r = .26, p < .01) and fluid intelligence significantly predicted divergent thinking performance using multiple regression analysis ( = .28, p < .01). Nusbau m and Silvia (2011) also showed that a higher order intelligence latent variable significantly predicted a higher order creativity latent variable using SEM ( = .45, p < .001). Vincent, Decker, and Mumford (2002) found a strong positive relationship between creativity and intelligence ( r = .73, p < .05) and in their structural equation model exploring the relationships among intelligence, divergent thinking and expertise, they found that intelligence significantly predicted divergent thinking ( = .23, p < .05). Finally, in an investigation of the relationships among fluid and crystallized intelligence and creativity, Sligh, Conners, and Roskos Ewoldsen (2005) showed that in a high IQ group, fluid intelligence was correlated with creativity which as assesse d by divergent thinking tests ( r = .39, p < .05). In sum, recent research points to intelligence playing an important role in creative thought, specifically in

PAGE 44

44 divergent thinking ability. An aim of this study is to extend these findings by examining the role of intelligence in not only divergent thinking, but associative fluency and convergent thinking as well. Executive Processes The historical roots of executive functions research is in neuropsychological studies of patients with frontal lobe damage; however, executive processes are also commonly discussed in the context of intelligence and working memory research. A challenge to studying executive functions is that these processes are manifested in the operations of other cognitive processes that are not directly tapping into executive processes (Miyake et al., 2000). Partly for this reason, the term executive functions has been applied using various definitions in the psychological literature ( Ilkowska & Engle, 2010; Miyake et al., 2000). In a review of the literature on executive processing, Miyake et al. (2000) identif ied shifting of mental sets, monitoring and updating of infor mation in working memory, and inhibition of dominant responses, as three of the most frequently cited processes related to executive functions. Results from a study using SEM showed support for three independent shifting, updating, and inhibition factors that were proposed to make up executive functioning The three separate factors were significantly related to each other ( r ranging from .42 to .63) and the fit indices of the three factor model ( 2 = 20.29, df = 24, CFI = 1.00, SRMR = .047) showed that t his model fit the data better than alternative two factor or nested models (e.g., 2 = 29.35, df = 25, CFI = .95, SRMR = .057) (Miyake et al., 2000). Generally represented by some form of these three processes, executive functions is an umbrella term that refers to important higher order cognition including the monitoring and regulation of cognitive

PAGE 45

45 processes, employment of strategies, searching for information, and judging and decision making during complex tasks ( Ilkowska & Engle, 2010). Recent qualitati ve research exploring the cognitive processes involved in divergent thinking (a core cognitive process believed to underlie creativity) suggests that divergent thinking involves several important executive functions Think aloud studies of dive rgent thinki ng show that higher order processes including strategy selection, category fluency, mental disassembling of figures, alternating between ideation and evaluation, and breaking set in the face of interference (Gilhooly et al., 2007; Khandwalla, 1993; Ruscio, Whitney, & Amabile, 1996), support divergent thinking processes For example, Gilhooly et al., (2007) found that participants completing an Alternative Uses task engaged in a range of strategies that involved som e degree of executive control. They found t hat s uccessful divergent thinkers exhibited higher rates of inhibiting common responses and deliberate sw itching of retrieval cues (Gilho oly et al., 2007), processes believed to engage the central executive component in working memory (Baddeley & Andrade, 2000). In addition, empirical studies of patients with frontal lobe damage also indicate that executive functions are involved in letter fluenc y and semantic category fluency tasks (Martin et al., 1994; Phillips, 1997). As discussed above, these fluency t asks engage memory searches aimed to retrieve a broad range of associated ideas and information from episodic and semantic memory. Therefore, there is strong indication that executive functions, such as working memory, are likely to predict performance on associative processing tasks.

PAGE 46

46 In light of studies that show intelligence as a significant predictor of creativity, as well as studies that indicate the important role of executive functions in associative abilities, the relationship between intelligence and creativity warrants reinvestigation. In addition, resear ch is needed to identify executive processes which may explain the role of higher order cognition in creative thinking This line of research is important for a better understanding of how various cognitive abilities that account for individual differences in analytical reasoning may also contribute to creativity As such, examining the possible roles of executive processes including working memory and processing speed in creative thinking is another aim of this study. The constructs w orking memory and processing speed will be discussed in more detail below. Working Memory and Short Term Memory Working memory involves the simultaneous storing and processing of information while engaging in complex mental activities such as the acquisition of new knowledge, reading comprehension, and problem solving (Baddeley 1992; Baddeley & Andrade, 2000; Baddeley & Logie, 1999; Ericsson & Kintsch, 1995). In contrast to the unitary storage model of short term memory, the working memory model is a multicomponent system, proposed to consist of a storage and an executive attention control component (Baddeley, 2000; Engle et al., 1999). Therefore, working memory capacity is believed to influence how successfully people are able to maintain memory representations (storage component) while simultaneously processing other information, overcoming distractions, and/or shifting attention (attention and central executive component) (Baddeley, 1992; Conway et al., 2003). Seve ral studies have shown that working memory is related to general intelligence ( g ) (Ackerman, Beier, & Boyle, 2002; Colom et al., 2005; Kyllonen &

PAGE 47

47 Christal, 1990; Sub et al., 2002) and fluid intelligence (Conway et al., 2002; Engle & Kane, 2004; Engle et al ., 1999; Unsworth & Engle, 2005). Studies examining the relationship between working memory and g have shown large structural coefficients of .70 (Ackerman et al. 2002), .96 (Colom et al., 2004) and .86 (Colom & Shih, 2004) between these two constructs. S imilarly, studies examining the relationship between working memory and fluid intelligence ( gf ) have shown structural coefficients of .59 (Engle et al., 1999) and .60 (Conway et al., 2002). Due to its limiting capacity, working memory, which consists of th e ability to maintain, update, and process information, is believed to be an important source of individual differences in general reasoning and learning ability (Conway et al., 2007). In support of this view, studies have shown that individuals with higher working memory capacity show better performance on tasks that require maintaining controlled attention while simultaneously overcoming interference (Conway, Cowan, & Bunting, 2001; Conway et al., 2002; Engle et al., 1999; Kane, Bleckley, Conway, & Engle, 2001). There are mixed findings regarding the relationships among short term memory, working memory, and intelligence (e.g., Ackerman, Beier, & Boyle, 2005; Conway et al., 2002; Engle et al., 1999). It has been suggested that working memory capacity cons ists of domain general central executive processes, whereas short term memory consists of lower order domainspecific coding and storage processes (Kane et al., 2004). Studies supporting this view showed that the common variance between working memory and short term memory reflected storage capacities, and the residual w orking memory variance reflect executive processes (e.g., Engle et al., 1999; Conway et al., 2002). For example, Engle et al. (1999) showed that short term memory and working memory were

PAGE 48

48 sig nificantly correlated ( r = .68), and when g f was added to the model, WM significantly predicted g f (.59) whereas short term memory did not ( .13). Based on these findings, an controlled view of wor king memory has been proposed (Conway et al., 2002), sugges ting that the executive component of working memory which allows a person to attend to goal relevant aspects of a task ( while also inhibit ing distractions) relates strongly to intelligence. In contrast other studies, such as the study conducted by Kane et al. (2004), have shown that bot h the executive attention component of working memory (.52) and the domainspecific processing factor of short term memory (.54) predicted fluid intelligence. Similarly, a recent SEM study by Colom et al. (2005) also show ed that when the working memory l atent variable is composed of storage and processing components, its relationship with general intelligence is strong (.89). However, when the storage component of working memory is partial l ed out (and a short term memory l atent variable is added to the model), the relationship of working memory to intelligence weakens (.79) and short term memory also predicts general intelligence (.58). Taking into account the short term, working memory, and intelligence debate, there is a general consensus that working memory capacity consists of important executive and attentional capacities that relates strongly to fluid intelligence. The important role of memory in creative thought is not a new notion. In The Nature of Human Intelligence, Guilford (1967) acknowledges that memory provides and stores information that serves a fundamental role in all problem solving and creative processes. Early studies exploring the relationship between memory and creativity typically employed simple memory recall and span tasks coupled with divergent thinking

PAGE 49

49 tests of creativity showing that memory predicted the ability to generate a large number of ideas (e. g., Pollert et al., 1969). Only a handful of studies have indirectly investigated the relationship between working memory and creativity (e.g., Daneman, 1991; DeYoung, Flanders, & Peterson, 2008; Sub et al., 2002;). Findings showed that working memory capacity is related to verbal fluency (Daneman, 1991), divergent thinking (Sub et al., 2002), and insight problem solving (DeYoung et al., 2008). Sub et al. (2002) tested several structural equation models relating working memory (spec ified as a storage and processing latent variable and a supervision latent variable) to intelligence ( g at the apex specified by speed, memory, creativity, and reasoning latent variables). Results showed that the storag e, processing, and coordination worki ng memory latent variable and the supervision working memory latent variable predicted the creativity factor ( .39 and .21, respectively). It has also been suggested that individuals with high working memory capacity are more successful at overcoming interference caused by automatic, unoriginal responses, and therefore, are more successful at using strategies to generate novel responses on divergent thinking tasks (Nusbaum & Silvia, 2011). Although the role of working memory capacity in various cognitive activities (e.g., academic, re ading comprehension, problem solving) has been studied, its relationship with creative thinking remains largely unexplored. This gap in the literature is important to address, especially in light of recent studies that show that working memory and creative thinking share higher order processes that reflect facets of general intelligence. Complex cognitive processes including the retrieval and processing of information, identification of useful strategies, supervision of mental processes, handling of complex conceptual relationships, and

PAGE 50

50 evaluation of ideas, are cited in both working memory and creativity literature (e. g., Baddeley, 1992; Cattell, 1971; Runco, 2004; Runco & Chand, 1995; Sub et al., 2002). In addition, researchers have proposed that attentional capacities influence performance on complex creativity tasks that necessitate attention to a wide range of cues (Mendelsohn, 1976; Rastogi & Sharma, 2010). Overall, a large body of research investigating the relationship between working memory and intelligence (see Ackerman, Beier & Boyle, 2005, for a review) as well as the relationship between creativity and intelligence exists (for reviews, see Barron & Harrington, 1981; Batey & Furnham, 2006) ; however, research investig ating the relationship between working memory and creativity has received little attention. Altogether, there is indication that working memory influences creative thinking and future research is needed to develop a better understanding of the role of wor king memory in creative thinking, including associative, divergent, and convergent processes. Processing Speed Processing speed or mental speed refers to how quickly a person can process information. Theoretical models of intelligence (e.g., Carrolls (19 93) threestratum theory of intelligence; HornCattell theory of intelligence, Horn & Cattell, 1966) identify processing speed as an important component of general i ntelligence. Processing speed typically measured by timed tasks such as reaction time, computing, and coding speed tasks (Jensen, 2006). Studies have shown that quicker processing speed predicts various abilities and performance on a broad range of tasks. For example, processing speed has been linked to problem solving ability (Carroll, 1993), phonological awareness, spatial and verbal memory, and reading comprehension (Jensen, 2006; Kail & Salthouse, 1994). In many studies of developmental changes in cognition, speed of

PAGE 51

51 processing has also been shown to mediate the relationships between agerela ted changes in working memory and general intelligence (e.g., Coyle et al., 2011; Fry & Hale, 1996). Modest correlations between various speed tasks and intelligence have been found (e.g., r = .30 to .40 ); however it has also been noted that processing s peed seldom accounts for more than 10% of the variance in g ( Jensen, 2006). In addition, the relationship between processing speed and intelligence tends to increase in studies that used complex processi ng speed tasks confounding the speed of processing with executive processes in accounting for the variance in g (Conway et al., 2002) Research shows that processing speed is significantly related to working memory capacity (Kail & Salthouse, 1994) and general intelligence (Deary, 2001; Fry & Hale, 1996) However, similar to the mixed findings regarding the relationships among short term memory, working memory, and intelligence (general and fluid) the nature of the relationships among processing speed, working memory, and g is also unclear. It is possibl e that working memory and processing speed independently contribute to g (Kaufman et al., 2009). Alternatively, some studies show that processing speed mediates the relationship between working memory and general intelligence (e.g., Fry & Hale, 1996; Jensen 1998; Kail & Salthouse, 1994). Then again, other studies show that working memory is the primary predictor of g (e.g., Conway et al., 2002; Engle et al., 1999; Kyllonen & Christal, 1990), even when controlling for processing speed. A study by Conway et al. (2002) explored the relationships among all four constructs (i.e., short term memory, working memory, processing speed, and fluid intelligence) Using SEM results from their study showed evidence for the strong predictive role of working memory (.60), and the weaker roles of short term memory (.18) and processing speed

PAGE 52

52 (.07) in predicting g f Overall, compared to other cognitive tests, measures of processing speed tend to load less strongly on general intelligence (Deary, 2001); therefore, although the link between processing speed and g is acknowledged, there is a general agreement that processing speed is not the central mechanism under lying intelligence. Similar to working memory, t he relationship between processing speed and creative thinking has r eceived little attention. Martindale (1999) proposed that creative people selectively control their speed of information processing. According to his theory, defocused attention at the beginning of the creative problem solving stage allows people to attend to many different stimuli, slowing down the speed of processing. As they move to later stages of problem solving, their attention becomes more focused on specific ideas or solution paths, allowing for quicker information processing (Martindale, 1999). Rin dermann and Neu bauer (2004) tested an indirect speedfactor model to explain the relationships among processing speed, intelligence, creativity, and academic performance. They proposed that processing speed is a basic mental ability that influences hig her mental abilities, which in turn predict real life achievements such as school performance. Results showed that intelligence and creativity mediated the relationships between processing speed and academic performance (sum of indirect effects: = .31). Inte lligence and creativity directly predicted school performance ( = .54 and .26, respectively) (Rindermann & Neubauer; 2004) From their study evidence for the direct role of processing speed in intelligence and creativity was shown; however, it was also noted that another important source of individual differences in human cognitive performance, working memory, was not accounted for in the model.

PAGE 53

53 Creative thinking tasks involve the quick processing and recording of information, and it is expected that mental speed, which has been linked to intelligence and working memory would also contribute positively to performance on creative thinking tas ks. In this study, the indirect role of processing speed through intelligence in associative fluency, divergent thinking, and convergent thinking will be explored. Implicit Theories and Creative Cognition Dweck and colleagues (1999) proposed that peoples approaches and responses to learning are rooted in the implicit beliefs they hold regarding intellectual abilities. She proposed two belie f frameworks for understanding peoples theories of intelligence: entity and incremental beliefs People who hold an entity theory of intelligence believe that intelligence is fixed or unchangeable, whereas people who hold an incremental theory of intelligence believe that intelligence is malleable or changeable (Dweck & Leggett, 1988; Dweck, 1999). Research has demonstrat ed that even when intellectual ability is controlled peoples implicit theories of intelligen ce shape the way they approach learning, impacting their motivation, goal orientations, and responses to challenges and feedback (e.g., Dweck & Leggett, 1988; He yman & Dweck, 1992). People who hold an entity theory of intelligence, often as a result of experiencing failures and/or receiving trait based praise (Dweck, 1999) tend to put forth low effort when faced with challenges, withdraw effort when receiving negative feedback, hold performance goals to showcase ability or to avoid negative judgments, and make negative attributions for the failures These individuals are also more likely to develop learned helplessness (e.g., attributing failures to factors out o f ones control) (Dweck, 1999) In contrast, those who endorse an incremental theory of intelligence tend to put forth greater effort to overcome challenges, hold mastery goals, focus on increasing

PAGE 54

54 understanding, make positive attributions, and demonstrate adaptive behaviors (e.g., developing strategies) when faced with failure (Dweck & Leggett, 1988; Hong et al., 1999). Furthermore, experimental studies have shown that teaching students about the incremental theory of intelligence leads to greater academic achievement (measured by higher grades) when compared to control groups (Aronson, Fried, & Good, 2002; Good, Aronson, & Inzlicht, 2003; Hong et al., 1995). Blackwell, Trzesniewski, and Dweck (2007) also found that implicit theories of intelligence impacte d adolescents math achievement over the two years of junior high school, indicating that beliefs of intelligence have long term effects on academic achievement outcomes, especially in the context of a challenging transition in which academic achievements have more serious consequences. In addition to intelligence, people may also hold implicit theories of creativity. To date the influence of peoples beliefs regarding the nature creative ability (as a fix ed or malleable trait) on their creative thinking has not been explored. Although the implicit theory of intelligence has not been investigated within the context of creativity research rela ted studies of peoples beliefs about creativity indicate that people hold entity like beliefs (Sternberg & Lubart, 1996). For example, a common stereotype surrounding creativity is the genius view, based on the belief that creative t hinking processes and acts are carried out by exceptional individuals who are divinely inspired or granted inexplicable sparks of insight. Stories of eminent individuals, such as Friedrich August von Kekules dream or the proverbial apple that fell on Isaac Newtons head, perpetuate this belief (Sternberg & Lubart, 1996; Weisberg, 1995).

PAGE 55

55 The genius view of creativity is not supported by contemporary research in creative cognition. A growing body of literature on the nature of human cognition indicates that insights may not be as arbitrary as was previously b elieved, but instead result from a series of highly structured cognitive processes More specifically, cognitive researchers explain that new ideas and solutions that appear to com e about in flashes of insight are actually incremental in nature, involving a series of systematic, constructive steps towards a novel solution (Finke et al., 1992; Ward, 1995; Ward et al., 1999). This transfer of knowledge from old to new experiences was suggested in some early studies in which subjects were exposed to a problem followed by a new problem that contained similar features of the first (e.g. Gick & Holyaock, 1980, 1983). Findings showed that subjects exhibited analogical transfer, abstracting key features from the initial problem to solve the second problem. This indicated that people use problem representations from past experiences ( consisting of goals, information relevant to defining and solving the problem, and strategies ) to apply to new problems (Gick & Holyaok, 1980, 1983; Holyoak & Koh, 1987) Along this line of argument, Thomas Ward (1995) claim ed that there is always something old about new ideas. He believed that when people develop new ideas, those ideas are heavily influenced by the individuals existing knowledge. Furthermore, Ward (1995) claim ed that the new ideas are structured in specific ways so that the old knowledge from which t hey come from can be traced back to specific category representations, a phenomenon he refers to as structured imagination (Ward, 1994, 1995). Support for structured imagination can be found in studies that have shown that an event remembered is altered ov er successive recall tasks, suggesting that the

PAGE 56

56 process of retrieving a memory involves an imaginative reconstruction of the past (Stein, 1989). Real life examples of creative acts also illustrate structured imagination, such as the invention of railway cars which some argue were modeled after the stagecoach, a common vehicle of the time (Ward, 1994, 1995). Ward (1994) conducted a hallmark study on the role of category structures in creative thought, in which a group of college students were asked to genera te a novel exemplar of some known category that would be appropriate to an imaginary setting (e. g., drawing animals from a different planet). Ward (1994) found that subjects drawings and description of the imagined creatures were highly structured (i.e., contained many properties such as eyes, legs, and similar patterns of the earth animals category), and attribute correlations (e. g., feathers, wings, and beak present together) were also evident. Furthermore, the constraining effects of existing category properties were present even in the condition in which participants were explicitly instructed to use their wildest imagination (Ward, 1994). These findings provided support for the commonalities between creative and noncreative thought. Ward (1994, 1995) concluded that because novelty and imagination seem to be rooted within a definite structural framework, an assessment of both old and new ideas need to be made in order to have a complete understanding of the creative process. In addition to old ideas s erving as the bases for new ones, creative cognitive researchers believe that old knowledge categories are constructed, extended, modified, and combined in response to new situations we encounter every day (Ward et al., 1997; Ward et al., 1999). Expanding upon Wards (1994, 1995) concept of structured imagination, the process of using preexisting concepts and extending them to develop

PAGE 57

57 novel combinations is a cognitive activity that creativity cognitive researchers call conceptual expansion (Ward et al., 19 97). Several studies have illustrate d the basic premises of conceptual expansion, showing that creative ideas develop as extensions of familia r concepts (e.g., Cacciari, Levorato, & Cicogna, 1997; Sifonis, 1995). In addition, conceptual expansion has been found to occur in higher frequency when subjects access knowledge in more abstract contexts (Ward et al., 1999). For example, in Wards (1994) study discussed above, when participants were asked to draw animals from a different planet, subjects showed a te ndency to take the path of least resistance (as demonstrated in their continuing with relatively unmodified ideas from existing categories). However, when the task was modified to specify unique conditions on the planets (e.g., ground covered with molten l ava), subjects used more general knowledge, demonstrated more flexibility in the kinds of information retrieved, and synthesized v arious types of information keeping the unusual environmental conditions in mind (Ward, 1994). This finding indicates that when problem contexts prevent subjects from relying upon common, existing exemplars, subjects draw upon a broader range of knowledge, engaging in a more flexible search and retrieval of information which leads to more innovative ideas. Weisberg (1995) took a similar standpoint that the development of creative products involves ordinar y cognitive processes. He argued that there has been too much emphasis on the negative influence of past experiences on creative thinking, stating that creative thinking moves beyond what has been done only slowly, and when it does, it is more as a modification of the past than rejection of it (Weisberg, 1988, p. 71). From this perspective, creativity is viewed as a combination of

PAGE 58

58 basic and more complex cognitive processes by wh ich old and new ideas are combined. Having made a case for the incremental nature of creative cognition, the literature on im plicit theories of intelligence suggests that it is possible that people who endorse a genius or entity view of creativity ar e likely to demonstrate less effort, focus on performance goals, lack intrinsic motivation, and employ less effective strategies in the face of challenges, such as the ambiguity that characterizes many of the creativity tests. On the other hand, if people hold an incremental view of creativity, they may engage in the reverse set of behaviors, which is likely to support greater creative thinking performance. Currently, research exploring peoples theories of creativity and the role that these beliefs play in creative thinking performance has not been conducted. Combining the work of Dweck and colleagues (1999) with the findings from the field of creative cognition, a second aim of this study is to explore how entity versus incremental beliefs of creativity influences performance on associative processing, divergent thinking, and convergent thinking tasks.

PAGE 59

59 CHAPTER 3 PURPOSE OF THE STUDY Exploring the Relationships among Working Memory, Processing Speed, Intelligenc e, Associative Fluency, Divergent Thinking, and Convergent Thinking In this study, I aimed to answer the following question regarding the relationship between intelligence and creative thinking: What are t he roles of intelligence and executive functions, including working memory and processing speed, in associative fluency, divergent thinking, and convergent thinking? Based on Mednicks (1962) theory of the associative basis of creativity and recent evidenc e of specific associative processes in creative thinking (e.g., Benedek et al., 2012), I examined whether associative fluency predicts two distinct creative processes: divergent thinking and convergent thinking. In addition, drawing from recent research that indicates an important role of intelligence in creativity (e.g., Batey et al., 2010; Cho et al., 2010; Silvia, 2008a, b; Vincent et al., 2002), the direct effect of intelligence on associative fluency, and indirect effects on divergent thinking and con vergent thinking, w as also explored. Finally, executive process es including working memory and processing speed, were also included in the model to e xamine their relative contributions on associative fluency, divergent thinking, and convergent thinking. T aken together, this study tests a theoretical model of the relationships among working memory, processing speed, intelligence, and three creative thinking processes. Examining the Role of Implicit Theories of Creativity in Creative Thinking I also examined the following question regarding the role of implicit theories of creativity in creative thinking: Are students theories of creativity ( i.e., entity versus incremental beliefs) related to performance on associative, divergent, and convergent creative th inking tests?

PAGE 60

60 Drawing from Dweck and colleagues (1995, 1999) work on implicit theories, I was interested in exploring peoples implicit beliefs of creativity, and the role that different beliefs may have in their creative thinking performance. More specif ically, I examined if performance on associative fluency, divergent thinking, and convergent thinking tasks differed as a function of holding entity versus incremental beliefs theories of creativity. In sum, by taking into account various cognitive abiliti es associated with intelligence, including working memory and processing speed, as well as thr ee cr eative thinking processes including associative fluency, divergent thinking, and convergent thinking ), this research explored the cognitive abilities and pro cesses that are involved in creative thi nking. The goal s of this study were to understand how three types of creative thinking processes are related, reveal the unique contributions of intelligence and executive functions on creative thinking, a nd examine how implicit theories of creativity influence creative thinking performance.

PAGE 61

61 CHAPTER 4 HYPOTHESES T he Relationships A mong I ntelligence, Working Memory, Processing Speed, Associative Fluency, Divergent Thinking, and Convergent Thinking In the propos ed model (Figure 4 1), Working Memory (WM) and Processing Speed (PS) were specified as exogenous latent variables. Intelligence (IQ), Associative Fluency (AF), Divergent Thinking (DT), and Convergent Thinking (CT) were specified as endogenous latent variables. Working memory was specified to have a direct relationship to intelligence and an indirect relationship to associativ e fluency through intelligence. The direct path from WM to IQ was expected based on research showing strong relationships between thes e two constructs (e.g., Conway et al., 2002, 2003; Engle & Kane, 2004; Engle et al., 1999). This relationship is explained by the shared attentioncontrol mechanism in working memory (Conway et al., 2003). In addition, recent research indicates that intell igence and executive functions play a significant role in creative thinking (e.g., Gilhooly et al., 2007; Nusbaum & Silvia, 2011; Silvia, 2008a, b ; Sub et al., 2002). T herefore, WM, an executive process that is an established predictor of intelligence, was specified to indirectly contribute to AF, DT and CT, through IQ in the proposed model. PS was specified as a direct predictor of IQ. Given prior research (Jensen, 2006; Kail & Salthouse, 1994), it is expected that PS will also have a positive direct relat ionship to IQ. Currently, research examining the relationship between PS and creative thinking processes (i. e., AF, DT, and CT) is lacking, with the exception of the speedfactor model tested by Rindermann and Neubauer (2004) However, drawing from research findings that shown an important role of processes closely related to

PAGE 62

62 processing speed (e.g., acti vation of semantic networks, fluent retrieval of ideas) in creativity, it is reasonable to expect that a persons mental speed in retrieving and generating information contributes positively to cr eative thought. Therefore, PS was also specified to indirectly contribute to AF, DT, and CT through IQ. IQ wa s specified as a direct predictor of AF, DT and CT. Based on research that suggests that associative abilities are related to intelligence (Benedek et al., 2012; Kaufman et al., 2009; Mendel sohn, 1976) IQ was specified to predict AF measured by letter and semantic category tasks The proposed relationship between IQ and DT was based on recent studies t hat showed a significant influence of IQ on DT abilities (Batey et al., 2010; Gilhooly et al., 2007; Nusbaum & Silvia, 2011). Finally, IQ was also specified as a direct predictor of CT based on research that suggest overlapping cognitive processes in both intelligenc e and convergent thinking tests of creativity because of the shared analytic and evaluative processes involved in convergent tests ( Kaufman, 2009 ; Mendelsohn, 1976). Finally, based on theoretical and empirical evidence of the associative bases of creative thinking, AF was specified as a direct predictor of DT and CT ( Benedek et al., 2012; Eysenck, 1995; Mednick, 1962). In summary, the following hypotheses were examined: 1. Working memory will have a direct relationship to intelligence. 2. Working memory will have an indirect relationship to associative fluency through intelligence, and an indirect relationship to divergent thinking and convergent thinking through intelligence and associative fluency. 3. Processing speed will have a direct relationship to intelligence. 4. Processing speed will have an indirect relationship to associative fluency through intelligence, and an indirect relationship to divergent thinking and convergent thinking through intelligence and associative fluency.

PAGE 63

63 5. Intelligenc e will have a direct relationship to associati ve fluency, divergent thinking, and convergent thinking. 6. Intelligence will also have an indirect relationship to divergent thinking and convergent thinking through associative fluency 7. Associative fluency will have a direct relationship to divergent thinking and convergent thinking. Figure 4 1 Proposed structural equation model of the relationships among working memory (WM) processing speed (PS) intelligence (IQ) associative fluency (AF), divergent thinkin g (DT), and convergent thinking (CT) The Role of Implicit Beliefs of Creativity on Creative Thinking Performance Currently little is known about how implicit beliefs of creativity influence peoples performance on creative thinking tasks. Research indicates that some people hold an entity view of creativity, believing that creativity requires genius abilities present among a select group of individuals ( Simonton, 2000; Sternberg & Lubart, 1996). Drawing from research on theories of intelligence (e.g., Dweck, 1999), as well as research on the role of motivation on creative performance (e.g., Amabile, 1996; Csizentmihalyi, 1999), it was hypothesized that holding an entity view of creativity would result in lower creative thinking performance due to reduced effort and motivation. In contrast, holding an incremental view of creativity was expected to result in greater creative thinking

PAGE 64

64 performance. To explore this proposition, t he following hypothesis was tested: P eople who report an enti ty belief of creativity will have significantly lower scores on associative fluency, divergent thinking, and convergent thinking tests compared to people who report an incremental belief of creativity.

PAGE 65

65 CHAPTER 5 METHOD Research Design The goal of this study was to identify the potential predictive power of intelligence, working memory, and processing speed on associative, divergent, and convergent components of creative thinking. Structural equation modeling (SEM) was used to explore t he relationships among the latent variables that represented the proposed constructs. For sophisticated statistical analyses such as SEM, a standard sample size of 200 is recommended (Hair et al. 1998; Kline, 2005; Tabachnick & Fidell, 1996). Simulation r esearch has shown that with a good model and multivariate normal data, sample sizes of approximately 200 cases have showed to produce r easonable results (Hox & Bechger, 1998). In this study, a convenience sample from the College of Education online participant pool was used, and data was collected from a total of 265 participants. Participants completed the entire study in one session, which lasted approximately one and a half hours. Scores from the measures included summative scores (e.g., single Sn apshot score to represent originality on divergent thinking tests) as well as aggregated scores (e.g., total of items correct on the Ravens Progressive Matrices). The SEM multivariate analysis technique is a powerful quantitative data analysis technique that estimates and tests the theoretical relationships among latent and/or observed variables (Kline, 2005). Estimates from SEM analyses include structural path coefficients that explain the relationships among variables in the SEM model. SEM combines information from various measures by estimating multiple regression equations simultaneously and also modeling the measurement errors that may be

PAGE 66

66 associated with the observed indicators (Kline, 2005; Loechlin, 2004). Benefits of applying SEM include analyses of commonalities in classes of tests, control for method variance that is specific to tests, and assessments of the fit of structural models (Kline, 2005; Silvia, 2008a, b). In addition, when examining a c omplex construct such as intelligence, structural equation models allow for the estimation of a higher order latent factor general intelligence or g which is composed of lower order latent factors (e.g., processing speed, working memory) ( Silvia, 2008a, b). In this study, SEM allowed for the test of relationships among components of intelligence and creative thinking, in order to explore on how complex sets of cognitive processes and abilities involved in both intelligence and creativity may be related. F inally s tudents self reported scores on a theory of creativity scale were used to compare the performance of students who hold entity belief to students who hold incremental belief about creativity, using independent samples t tests. Data Collection Pa rticipants A total of 265 participants were recruited through an online research participant pool from educational psychology classes at a large southeastern university. The sample consisted of 59 males and 206 females. The breakdown of participants ethni cities was as follows: 160 Caucasian, 42 Black, 39 Hispanic, 18 Asian, and 6 Other. The average age of the sample was 20.33 years (SD = 2.54) and students had completed an average of 2.53 years of school (SD = 1.13). Table 51 presents the demographic breakdown of the sample in percentages. Prior to conducting the study, all study procedures were approved by the universitys Institutional Review Board.

PAGE 67

67 Table 5 1. Demographic breakdown of participants Demographic Variable Fre quency Percentage Gender Male 59 22.26% Female 206 77.74% Ethnicity Caucasian 160 60.38% Black 42 15.85% Hispanic 39 14.72% Asian 18 6.79% Other 6 2.26% Year in School Freshman 63 23.77% Sophomore 64 24.15% Junior 72 27.17% Senior 66 24.91% Participants read and signed an informed consent form before completing a battery of paper andpencil and computer based intelligence, working memory, processing speed, and creativity tests, as well as a creativity beliefs and demographics questionnaire. Upon completion of the experiment, participants received one credit for completing a course requirement to participate in a research study. Materials Creative thinking t asks Associative fluency (AF) tasks Letter Fluency t ask (B orkowski et al., 1967). The letter fluency task was developed to assess phonetic fluency, and requires participants to generate a list of as many words as possible that begin with the letter F. Participants were given 2 minutes to complete the task. The t otal number of appropriate words generated for the letter F was used for the total score. Category Fluency t asks (Benton & Hamsher, 1978). The category fluency task requires participants to generate a list of as many different types of animals and jobs as

PAGE 68

68 possible. Participants were given 2 minutes for each of the tasks (total of 4 minutes). The total number of appropriate animals and jobs generated was used for the total score. Both letter and category fluency tasks were originally developed to diagnose phonetic and semantic category specific impairments due to neurological disorders such as aphasia (Benton, 1994). However these fluency tasks are believed to tap into peoples ability to fluency retrieve and effective organize verbal information have also been used in psychological studies to assess semantic memory (e.g., Collins & Quillian, 1972; Schelble, Therriault, & Miller, 2012), fluid intelligence (e.g., Silvia, 2008a, b) and associative processing (e.g., Benedek et al., 2012). In this study, the fluency tasks were used to assess associative fluency, operationalized as the ability to efficiently retrieve a broad range of a ssociations from a taxonomic category The score on the letter fluency task was moderately correlated with the category fluency scores for name animals and jobs, r = .42, .35, p < .001, respectively, providing evidence for convergent validity. Divergent t hinking (DT) tests Guilfords Unusual Uses Tests (Guilford, 1967). The Unusual Uses test requires participants to develop unusual uses for a common household item. The item for this task was a wire coat hanger. Participants were instructed to think of creative uses for the item, and were given 3 minutes to generate as ma ny unusual uses as possible. Participants responses on the Unusual Uses test was scored using the Snapshot scoring method, proposed by Silvia and colleagues (2008, 2009). The Snapshot scoring method gives a set of responses on a divergent thinking test a single holistic rating on a scale of 1 ( not at all creative) to 5 ( very creative ), producing one score for each persons

PAGE 69

69 ideational output (Appendix A). An obvious benefit of the Snapshot scoring method is the significant reduction in labor and time to sc ore the tasks. Studies employing scoring methods similar to the Snapshot method showed high inter rater reliabilities ranging from .92 to .98 (Runco & Mraz, 1992). There is evidence of good construct reliability, as H values representing maximal reliabilit y (the degree to which indicators capture information about the underlying factor) were over .80 for the Snapshot scoring method (Silvia et al., 2008, 2009). Snapshot scores related positively with openness to experience, a personality trait shown to be po sitively related to creativity (Feist, 1998) ( = .33), and negatively with conscientiousness, a personality trait shown to be negatively related to creativity ( = .29), providing evidence for the concurrent validity of Snapshot scores (Silvia et al., 20 09). The Abbreviated Torrance Test for Adults (ATTA; Goff & Torrance, 2002) The ATTA was adapted from The Torrance Tests of Creative Thinking (TTCT; Torrance, 1966) and is a widely used measure for assessing divergent thinking (Plucker & Renzulli, 1999). The TTCT has been reviewed and found to be a reliable and valid measure of creative thinking (interrater reliability greater than .90, and test retest reliability between .50.93) (Kim, 2006; Torrance, 2008). The ATTA contains three 3minute verbal and fig ural tasks from the TTCT. In the verbal task (Activity 1) participants are asked to identify the troubles they might encounter if they could walk on air or fly without being in an airplane or similar vehicle. In the picture completion task (Activity 2), th e participants are presented with two incomplete figures, and asked to draw pictures with the figures. Finally, in the picture construction task (Activity 3), participants are

PAGE 70

70 presented with nine identical isosceles triangles arranged in a 3 x 3 matrix and asked to draw pictures using the triangles (Appendix B). The ATTA is scored on four norm referenced measures and 15 criterionreferenced indicators (Goff & Torrance, 2002). The four norm referenced measures include fluency (i.e., number of ideas), originality (i.e., unconventionality or uniqueness of ideas), elaboration (i.e., details or embellishments of ideas), and flexibility (i.e., different types of ideas) (Goff & Torrance, 2002). For example, when scoring Activity 1, participants responses are comp ared to a list of common responses (e.g., air sickness, get cold) from the manual. Each response that is not present on the list receives one point for originality. Similarly, for Activities 2 and 3, a criterion for awarding elaboration points is provided in the manual (e.g., color, shading, decoration) (Goff & Torrance, 2002). The fluency, originality, elaboration, and flexibility ratings are summed across the three tasks and converted to a scale that was developed using the conventional stanine scale consisting of a 9 point normalized standard score from 11 ( low) to 19 ( high ), centered at 15 (Goff & Torrance, 2002). The normalized scaled scores are summed to produce a total scaled score. There are 15 criterion referenced creativity indicators (e.g., richn ess and colorfulness of imagery, expressions of feelings and emotions, abstractness of titles), each scored on a 3point scale of 0 ( absence) to 2 ( two or more present ) (Goff & Torrance, 2002). The composite of total scaled scores from the norm referenced measures plus criterion referenced indicators combine to yield a Creativity Index (CI) ranging from 44 to 106. The CI is rescaled and reported as a creativity level ranging from 1 ( minimal ) to 7 ( substantial ) (Goff & Torrance, 2002). Evidence for the predi ctive and discriminant validity of scores from the ATTA has been reported in recent

PAGE 71

71 studies (e.g., Althuizen, Wierenga, & Rossiter, 2010; Kharkhurin, Samadpour, & Motalleebi, 2008). The norms reported in the ATTA manual are based upon adults who had completed the D TTCT prior to the year 2000. The manual reports the Kuder Richardson (KR21) reliability coefficient of .84 for the total raw score for the four creative abilities, and .90 for the total raw score plus the creativity indicators score. Inter rater reliabilities range from .95 t o .99 (Goff & Torrance, 2002). In this study, t he inter rater reliability for the creativity indicators score was .96 for the first 100 tasks that were scored independently by two raters Because of the high inter rater reliability, the subsequent tasks were scored by one rater. Convergent t hinking (CT) t ests Remote Associates Test (RAT; Mednick, 1962; Mednick & Mednick, 1967). The RAT requires participants to identify a solution that is associated with three presented cue words either semantically or through formation of a compound word (e.g., birthday, light, stick, answer: candle) (Appendix C). The RAT was developed by Mednic k (1962) based on his associative theory of creativity. He proposed that more creative individuals have flatter associative hierarchies that support the activation and combination of more distally related elements in the mental network (Mednick, 1962). The refore, more creative individuals who are better able to find a mediating link between seemingly unrelated words are expected to show superior performance on the RAT. Mednick and his colleagues provide evidence of both predictive and construct validity of the RAT. Studies have shown that performance on the RAT correlated with faculty ratings of creativity for student architects ( r = .70) (Mednick, 1962) and graduate students in psychology ( r = .55) (Mednick, 1963), and achieving contracts for research proposals in

PAGE 72

72 science and engineering domains (Gordon, 1966). Studies also showed that high scores on the RAT were positively and significantly related to measures of associational fluency (Craig & Manis, 1962). The RAT scores were weakly but significantly correlated with scores on the three associative fluency tasks in this study ( r = .15 to .22, p < .001), and weakly and insignificantly correlated with performance on divergent thinking tasks ( r = .12 to .09, p < .001), providing evidence for discriminant validity. The RAT was completed on a computer and instructions were presented on the first screen. Participants completed four practice sets, each of which consist of three cue words presented on the screen, followed by a blank screen, where the participant typ ed in their response. Participants were given up to 15 seconds per set of 3 cue words before being prompted to generate the fourth word. For the four practice sets, the correct answer was shown following participants response. After the practice trial, participants completed a set of 30 cue words. The set of cue words were selected from Bowden and Jung Beemans (2003) normative data set of 144 compound remote associate problems. The problems were selected and programmed to increase in difficulty as the tas k progressed. Each correct answer was given a score of 1, for a total possible score of 30 points. Incorrect answers were given a score of 0. The total number of items correct was used. Insight problems (Dow & Mayer, 2004) For the insight problems task, participants w ere required to solve a set of 5 in sight problems (one spatial, three verbal, and one mathematical) (Appendix D). Each correct answer was given a score of 1, for a total possible score of 5 points. Incorrect answers were given a score of 0. The total score on the insight problems was significantly correlated with the RAT at r = .20, p <

PAGE 73

73 .001, providing some evidence for convergent validity, and was weakly related to scores on divergent thinking tests ( r = .05, n s, to .14, p < .01). Cronbachs alpha for the 5 items on the insight problems task was .27, resulting weak support for the internal consistency of the measure across the item scores. However, the reliability analysis indicated that deleting any of the five items would decrease overall reliability (deleting items 1, 2, 3, 4, or 5 results in C r onbachs alpha = .27, .15, .21, .26, and .25, respectively). Intelligence t ests Ravens Advanced Progressive Matrices (RAPM; Raven, Court, Raven, 1977; Raven, Raven, & Court, 1998). The short 12item version of the RAPM (Bors & Stokes, 1998) was used. The RAPM is a standardized intelligence test that consists of a series of matrices made up of geometric figures with one section of the matrix missing. Participants are r equired to identify the correct missing section from a set of eight possible answer choices. Participants completed two practice problems from the RAPM Set I and answers were provided after completion of the two practice problems. Participants were given up to 20 minutes to complete the short 12item version from the RAPM Set II (i.e., Items 3, 10, 12, 15, 16, 18, 21, 22, 28, 30, 31, and 34) (Bors & Stokes, 1998). Bors and Stokes (1998) showed that the short version of the RAPM showed satisfactory psychomet ric properties, including a Cronbachs alpha of .73, correlation with the full length RAPM set II of .92, test retest reliability of .82, and a moderately strong and statistically significant correlation of .42 with an information processing task (Bors & Stokes, 1998). In the present study, Cronbachs alpha for the 12 items of the RAPM was of .71 providing evidence for the internal consistency of the measure across the scaled activities.

PAGE 74

74 Weschler Adult Intelligence ScaleRevised, Vocabulary (WAIS R; Weschl er, 1981). The Weschler Adult Intelligence ScaleRevised is an a test consisting of 11 subtests (6 verbal and 5 performance) that are developed to collectively assess the global intelligence of adults between the ages of 16 to 74. This revised version of the original WAIS reflects modifications to test items to eliminate as many cultural, ethnic, and sex differences as possible. Several factor analytic studies of the WAIS R showed that evidence for separate verbal and nonverbal factors (Anastasi, 1982). Reliability studies showed high reliability coefficients for the verbal, performance, and full sc a le IQ, ranging from .52 for the Object Assembly subtest to .96 for the Vocabulary subtest (Spruill, 1984). For the purpose of the present study, the Vocabulary subtest of the WAIS R was used. The Vocabulary subtest correlates .85 with verbal IQ and .81 with total IQ score, providing evidence for its validity. The Vocabulary subtest consists of 35 successive words that increase in degree of unfamiliarity within a normal population. Participants were presented with the list of words, and given up to 15 minutes to provide as complete of a definition for each word as possible. The present researcher and a research assistant independently scored the Vocabulary subtest. The definition for each word was given a score ranging from 0 to 2. The items were scored according to the rubric in the WAIS R Tutorial Workbook (Swiercinsky, 1988). The inter rater agreement was r = .91. Any disagreements in scoring were discussed. Because inter rater agreement was high, the remainder of the tests was score by one person. In the present study, Cronbachs alpha for the 35 items of the WAISR was .88 providing evidence for the internal consistency of the measure across the item scores.

PAGE 75

75 Work ing memory test s Symmetry Span task (SymSpan, Unsworth, Redick, Heitz, Broadway & Engle, 2009). This task requires participants to determine the symmetry of a geometrical picture while simultaneously memorizing the position of a red square on a 4 by 4 square matrix. Participants were presented with a series of displays on a computer screen. Each display consisted of a geometrical picture, which the participant judged as yes it is symmetrical or no it is not symmetrical by clicking on the corresponding y es or no button. They were then presented with the 4 by 4 square matrix, in which one of the squares was highlighted in red. After a number of displays had been presented, a recall cue consisting of a blank 4 by 4 square matrix prompted the participants to recall all of the red squares in sequence from the series. The number of displays per series varied from two to six. As suggested by Conway et al. (2005), data was excluded from participants who incorrectly answered more than 15% of the symmetry problems, because below the 85% criterion level of processing accuracy indicates that the participant was not engaging in both the processing and storing components of the complex span task. Unsworth et al. (2009) examined the relationships among the processing accuracy, processing time, and storage accuracy components within complex span tasks (including the SymSpan) as well as each components unique and shared contribution in predicting fluid intelligence abilities in spatial, numerical, and verbal domains. Results showed that the SymSpan recall and processing accuracy components were significantly correlated to recall and processing components of the Operation Span ( r = .66, .24, p < .001, respectively) and the Reading Span ( r = .70, .34, respectively) (Unsw orth et al., 2009). In addition, a confirmatory factor analysis (CFA) of

PAGE 76

76 the complex task scores showed that a three factor model consisting of a processing accuracy, processing time, and storage accuracy structure fit best, providing evidence that the com plex span tasks are tapping into the intended three separate processes. Results from a communality analysis showed that 69% of the variance in fluid intelligence was accounted for by the three components of the complex task. Recall accuracy, processing tim e, and processing accuracy accounted for 42%, 45%, and 37% of the variance in fluid intelligence, respectively (Unsworth et al., 2009). Altogether, results suggest that the SymSpan is a complex span task that engages several mechanisms related to fluid int elligence. The total score on the SymSpan was used in the analysis, which represents the total number of squares that were correctly recalled in position. Backward Digit Span task (Wilde, Strauss, & Tulsky, 2004) The backward digit span task requires participants to recall a random sequence of digits in reverse (backwards). Participants were presented with a series of digits, one at a time, on a blank computer screen (e.g., 6, 3, 2, 5). After a series of digits have been presented, a recall cue consisting of a blank rectangular text box prompted the participants to recall of the digits in reverse (e.g., from previous example: 5, 2, 3, 6). The length of the series of digits began with three digits, and increased as the task progressed. This task is believed to be a complex span task which relies heavily on working memory processing due to the simultaneous storage and processing requirements involved for remembering and mentally reversing the information (Wilde et al., 2004). The mental manipulation of digit s is believed to engage the central executive component of working memory (Baddeley, 1986, 2000). The longest series of digits recalled in reverse was the score

PAGE 77

77 on the Backward Digit Span task. The total score on the backward digit span task was significan tly correlated with the total score on the Symmetry Span task ( r = .23, p < .001), providing some evidence for convergent validity. Processing s peed t asks Pattern C omparison T ask (Conway et al., 2002) This task consists of a two parts. Part I and Part II contain a series of pairs of patterns on a page with a horizontal line between the members of a pair. Participants were required to indicate whether the pair of patterns are the same (S) or different (D). Thirty seconds were given for each of the two parts, and participants were instructed to work rapidly to complete as many items as possible. The series of patterns become progressively more difficult in complexity and in length. The number of correctly identified patterns was used for the total score. C ronbachs alpha for scores from Part I and II was .74 providing evidence for the internal consistency of the measure across the scores. Letter C omparison T ask (Conway et al., 2002) This task consists of a two parts. Part I and Part II contain a series of pairs of letters on a page with a horizontal line between the members of a pair. Participants were required to indicate whether the pair of letters are the same (S) or different (D). Thirty seconds were given for each of the two parts, and participants were instructed to work rapidly to complete as many items as possible. The series of letters become progressively more difficult in complexity and in length. The number of correctly identified patterns was used for the total score. Cronbachs alpha for scores from Part I and II was .73 providing evidence for the internal consistency of the measure across the scores. The total score on the letter comparison task was moderately correlated with the total score on the pattern comparison tasks ( r = .42, p < .001) providing some evidence for convergent validity.

PAGE 78

78 Implicit theory of c reativity s cale Theories of Intelligence Scale (TIS; Dweck, 1999). An adapted version of the TIS scale (originally developed to measure theories of intelligence) was used to determine part icipants theory of creativity Scores on the TIS scale among student populations have shown to be normally distributed with a mean of 34 and standard deviation of 8 (Au, 2001). The scale consists of 16 items. The items of the TIS were altered to target beliefs regarding creativity, by switching intelligence and intelligent to creativity and creative, respectively. Dwecks (1999) TIS has been adapted in a similar fashion in studies examining the role of entity versus incremental beliefs in social stereotyping (Levy, Stroessner, & Dweck 1998; Plaks, Stroessner, Dweck, & Shermanm 2001) and in mathematics ability (Cury, Elliott, Da Fonseca, & Moeller, 2006). Nine items are statements that endorse the entity view of creativity (e.g., You have a cer tain amount of creativity, and you cant really do much to change it), and seven items are statements that endorse the incremental view of creativity (e.g., You can always substantially change how creative you are) (Appendix E). Each item is rated on a six point Likert scale 1 ( strongly disagree) to 6 ( strongly agree). Scores on the TIS ranged from 45 to 93 with a standard deviation of 9.19 for the 16 item version, and ranged from 3 to 18 with a standard deviation of 4.07 for the 3 item version (consisti ng of the first 3 items from the 16 item scale). Higher scores represent stronger agreement with an incremental view. The first 3 items of the TCS scale were adapted from Dwecks (1999) TIS scale (i.e., You have a certain amount of creativity and you cannot do much to change it., Your creativity is something about you that you cant change very much., and You can learn new things, but you cant really change your basic creativity.) (Dweck, 1999,

PAGE 79

79 p. 177). For the purpose of this study, the first three items that measured entity beliefs were used. These items were reverse scored. The additional three items in Dwecks (1999) TIS measured incremental beliefs (i.e., No matter who you are, you can change your intelligence a lot, You can always greatly change how intelligence you are, No matter how much intelligence you have, you can always change it quite a bit) (Dweck, 1999, p. 177). Most studies using Dwecks (1999) TIS scale have used the 3 entity item scale. The following reasons for using the 3 it em scale is provided by Dweck and colleagues (1995, 1999). First, the items are repetitive, and respondents may find the questionnaire tedious. Second, Dweck found that respondents tend to give higher ratings for items endorsing the incremental theory, sug gesting that because these items are more socially desirable, participants are more likely to report incremental beliefs of intelligence. Finally, the 3 item scale allowed for the sample to be split into two groups (Dweck, 1999, 2006; Dweck, Chiu, & Hong, 1995). In addition, Dwecks (1999) scale used a reversed Likert scale, ranging from 1 ( strongly agree) to 6 ( strongly disagree). Dweck, Chiu, and Hong (1995) used the average scores on the 3 entity items to form an overall implicit theory score ranging from 1 to 6. Higher scores indicated a stronger incremental theory. It was recommended that to ensure only participants with clear theories are included in the study, participants with an overall score of 3.0 or below are classified as entity theorists, and participants with an overall score of 4.0 or above are classified as incremental theorists (Dweck et al., 1995). According to Dweck and colleagues (1995) this method typically results in about 15% of the participants being excluded from the analysis, and th e remaining 85% evenly distributed between the entity and incremental groups

PAGE 80

80 (Dweck et al., 1995). In this study, the standard Likert scale was used (i.e., 1 ( strongly disagree) to 6 ( strongly agree) for the 3 entity items, and scores on these items were r everse scored) Therefore, higher scores indicated stronger incremental theories. Dweck et al. (1995) reported high internal reliabilities ranging from .94 to .98, and test retest reliabilities of .80 for the three items on the theories of intelligence scale Evidence for the validity of the scale was demonstrated by results from a factor analysi s that showed that items on the theories of intelligence scale loaded on a separate factor from items regarding implicit beliefs about people and morality (Dweck et al., 1995). In order gain a more comprehensive assessment of students implicit beliefs about creativity, an additional 13 items were adapted from the Implicit Theories of Intelligence Scale (ITIS, AbdEl Fattah & Yates, 2006) (Appendix E). In a study conducted by AbdEl Fattah & Yates (2006), the ITIS was administered to an Australian and Egyptian sample. An exploratory factor analysis with oblique rotation showed that a two factor model consisting of 7 items in each factor fit the data best. For the Eg yptian sample, the two factors were negatively correlated ( r = .35, p < .001), indicating the factors measured opposite traits. Cronbachs alpha for the entity factor was .83 and for the incremental factor was .75. The entity and incremental factors explained 35.5% and 15.3% of the total variance, respectively. For the Australian sample, the two factors were negatively correlated ( r = .33, p < .001). Cronbachs alpha for the entity factor was .78 and for the incremental factor was .76. The entity and incr emental factors explained 26.5% and 18% of the total variance, respectively. In addition, a two factor CFA using the Egyptian and Australian data showed good fit ( 2 = 83.6, 83.7, p < .05, RMSEA = .04, .05, CFI = .99, .98, and SRMR = .04, .05, respectively ).

PAGE 81

81 In our sample, Cronbachs alpha for the entity items of the TCS was .70 and for the incremental items, .72, providing evidence for the internal consistency of the m easure across the items for the two type s of implicit belief s. The total score on the entity items of the TCS were negatively correlated to the total score on the incremental items ( r = .07, ns ) providing evidence for discriminant validity. Procedure Upon arrival at the laboratory, participants were asked to read and sign the informed consent form. Participants then completed a series paper and pencil and computer based intelligence, working memory, processing speed, and creativity tasks. After the completion of the study, participants filled out a theory of creativity scale and a demog raphics questionnaire. Data Analysis Exam ining the Role of Intelligence and Executive Functions i n Creative Thinking The structural equation model of the relationships among intelligence, working memory, processing speed, associative fluency divergent thi nking and convergent thinking were estimated with MPLUS version 6. Maximum likelihood estimation with robust errors (MLR) was used. This estimation procedure is recommended for data that is normally distributed, but includes missing data and has some indi cators with more or less kurtosis than would otherwise be common in a normal distribution. Under this estimation procedure, model parameter estimates are identical to those under maximum likelihood (ML) estimation; however, MLR adjust the model chi square and model standard errors to take into account the missing data and nonnormality issues. A correlational analysis between obs erved variables was first calculated Prior to testing the proposed structural model, the fit of the measurement models were tested

PAGE 82

82 using confirmatory factor analysis (CFA). CFA incorporates the testing of unidimensionality of the constructs of interest, and evaluates a data set by confirming the underlying theoretical structure (Kline, 2005; Mueller, 1996) To assess model fit, the following absolute and comparative goodness of fit (GOF) indices were obtained: the root meansquare error of approximation (RMSEA), the standardized root mean squa re residual (SRMR), the comparative fit index (CF I), and the Tucker Lewis Index (TLI). T he chi square test statistic was also obtained. The following cut off values recommended The chi square ( 2) test statistic is a measure of the absolute discrepancy between the matrix of implied variances and covariances, to the matrix of empirical sample variances and covariances; therefore, this statistic tests whether the implied matrix is significantly different from the observed matrix. A probability was used to test the null hypothesis that there is no significant different between the model implied covariances and the observed covariances. However, 2 has been shown to be sensitive to sample sizes and complexity of the model, making it more likely to reject a goodfitting model with larger sample sizes and more complex models (Kline, 2005) Taking this into account, the other GOF indices were referred to in order to make conclusions regarding the fit of the measurement and structural models. The RMSEA has been cited as one of the most informative criteria in SEM (Byrne, 2001). It takes into account the error of approximation, therefore, smaller values indicated better fit. The RMSEA is not affected by sample size, and is referred to as a populationbased index (Kline, 2005). The TLI and CFI are widely used indices in SEM to assess the relative improvement in model fit. The proposed model is compared to a baseline model fit

PAGE 83

83 criteria to assess how much better the estimated model fits with the observed data (Kline, 2005) Smaller TLI and CFI values indicate better model fit. Model fitting to determine possible improvements to the model was conducted in cases that were statistically and substantively justifiable, by referring to modification indices as well as taking in account theoretical considerations. Post hoc model respecifications were therefore justified on statistical and c onceptual grounds ( Jackson, Gillaspy, & Purc Stephenson, 2009). The same set of GOF indices used in the CFA w as used to determine the fit of the revised model. After identifying the best fitting measurement model, t he structural equation model was embedded in the measurement model. A model comparison test was conducted to assess the fi t from the measurement model compared to the structural equation model. Path analyses were conducted to assess the direct and indirect effects specified in the structural model. All standardized parameter estimates (including both significant and nonsignificant) are reported. Examining the Role of Implicit Beliefs of Creativity in Creative T hinking On the basis of scores on the theory of creativity scale, participants were categorized into either an entity belief or an incremental belief group following the cutoff points specified by Dweck and colleagues (1995, 1999) Specifically, participants with an average score less than 3 on the theory of creativity scale were categori zed in the entity group, and participants with an average score greater than 4 were categorized in the incremental group. Participants with an average score that fell between the values of 3 and 4 were excluded from the analyses. Using SPSS version 19, independent samples t tests were conducted to compare performance on the associative fluency divergent thinking, and convergent thinking tasks between the entity belief and incremental belief groups.

PAGE 84

84 CHAPTER 6 PILOT STUDY Methods Participants For the pilot study, data were collected from 83 participants from a large southeastern university, through an online research participant pool. Participants were asked to read and sign an informed consent form before completing a series of creativity and intelligence measures, and a demographics questionnaire. Upon completion of the experiment, participants received credit towards a course research requirement. Measures Divergent T hinking (DT) T ests. DT measures include Guilford s Alternative Uses Task (Guilford, 1967) (6 mins) which consists of developing unusual uses for two common household items (e.g., wooden pencil and wire coat hanger), and Guilfords Consequences Test (Christensen, Merrifield, & Guilford, 1953) (3 mins) which consists of generating ideas to a hypothetical situation. All of the DT tests were scored using both the Top2 and Snapshot scoring methods (Silvia et al., 2008, 2009). Convergent T hinking (CT) T ests. CT measures include the Remote Associates Test (RAT; Mednick, 1962; Mednick & Mednick, 1967) (10 mins) which consists of identifying an associate word from three cue words, and five insight problems that include spatial, verbal, and mathematical problems (5 mins). Associative Fluency (AF) Tests. AF measures included two letter fluency task in which participants were asked to generate as many words as they could f or the letters F and M (4 mins).

PAGE 85

85 Fluid In telligence (gf) T asks. Fluid intelligence was assessed by a short version of Ravens Advanced P rogressive Matrices (Raven, et al., 1998) (20 mins). Working M emory (WM) T asks. The WM tasks include the Symmetry Span task (10 mins), and Backward Digit Span task (3 mins). All of the tasks were completed either on a computer or in a paper and penci l format. Results of Pilot Study Preliminary results from the pilot study, including correlations among observed variables and a path analysis are presented below. Correlations among Observed Variables Correlations among the fluency, Top2, and Snapshot ratings are presented in Table 6 1 Correlations among the ratings using the Top2 scoring method were strongly correlated with the ratings using the Snapshot scoring method for the Unusual Uses tests (i.e., wooden pencil, wire coat hanger) and the Consequences test [ r (83) = .82 .9 3 p < .001)]. Correlations among the associative fluency divergent thinking, convergent thinking, intelligence, and working memory tasks are presented in Table 6 2 Table 6 1 Correlations among fluency, t op 2 ratings, and snapshot ratings for wooden pencil, wire coat hanger, and consequences divergent thinking tests 1 2 3 4 5 6 7 8 9 1. WP Fluency 1.00 2. WP Top2 .03 1.00 3. WP Snapshot .15 .82 ** 1.00 4. WCH Fluency .72 ** .05 .17 1.00 5. WCH Top 2 .17 .38 ** .42 ** .16 1.00 6. WCH Snapshot .25 .45 ** .50 ** .31 ** .89 ** 1.00 7. Conseq Fluency .19 .09 .11 .29 .07 .01 1.00 8. Conseq Top2 .09 .10 .03 .04 .04 .00 .31 ** 1.00 9. Conseq Snapshot .20 .00 .06 .13 .15 .15 .26 .93 ** 1.00 p < .05, ** p < .001.

PAGE 86

86 Table 6 2 Correlations among working memory tasks, intelligence tests, divergent thinking and convergent thinking tests p < .05, ** p < .001. Fluid intelligence was signif icantly related to associative fluency (i.e., letter F fluency test score ) [ r (83) = .3 3 p < .001)], and working memory (i.e., scores on Backward Digit Span and Symmetry Span tasks) [ r(83) = .25, p < .05), r (83) = .56, p < .001, respectively]. Convergent thinking (i.e., total score on i nsight problems ) was also significantly related to associative fluency (i.e., letter F fluency test) [ r (83) = .30, p < .001)] and working memory (i.e., Symmetry Span task) [ r (83) = .26, p < .05)]. Finally, d ivergent thinking (i.e., wire coat hanger unusual uses task score ) was correlated with associative f luency [ r (83) = .24, p < .05). Results of Path Analysis Exploring the Role o f W orking Memory and Intelligence o n Associative Processing, Di vergent Thinking, a nd Convergent Thinking Mplus 6 was used to explore the relations hip between Working Memory (WM1: Backward Digit Span, WM2: Symmetry Span), Intelligence (IQ), Divergent Thinking (DT), Convergent Thinking (CT), and Associate Fluency (AF ) (Figure 6 1 ). 1 2 3 4 5 6 7 8 9 10 11 1. Letter F 1.00 2. Letter M .56 ** 1.00 3. RAP M .33 ** .19 1.00 4. Backward Digit .19 .16 .25 1.00 5. SymSpan .11 .08 .55 ** .13 1.00 6. SymSpan Acc .03 .01 .56 ** .10 .89 ** 1.00 7. WP Snapshot .11 .19 .09 .02 .13 .14 1.0 0 8. WCH Snapshot .24 .14 .03 .02 .07 .06 .50 ** 1.00 9. Cons Snap shot .05 .05 .13 .03 .05 .07 .06 .15 1.00 10. RAT .18 .07 .14 .16 .05 .11 .02 .12 .09 1.00 11. Insigh t .30 ** .05 .18 .06 .28 .26 .02 .09 .16 .16 1.00

PAGE 87

87 *p < .05, ** p < .001 Figure 61. Relat ionships among working m emory ( WM1, WM2), intelligence, a ssociative f luency divergent t hinking and convergent t hinking. The GOF indices showed that the model was an adequate fit to the data ( 2 = 6.803, df = 5, p = .236, RMSEA = .07, CFI = .97, TLI = .91, SRMR = .046). Results showed that WM1 and WM2 had a s ignificant direct effect on IQ ( = .35, .46 p < .001, respectively). IQ had significant direct effects on AF ( = .46 p < .001). The effects of IQ on DT and CT were nonsignificant ( = .06 p = .635, = .16, p = .196, respectively). AF had a signifi cant direct effect on DT ( = .259, p = .025) but not on CT ( = .12 p = .343). The indi rect effects of WM1 and WM2 on AF was significant through IQ ( = .54 p = .007, = .14 p = .002 respectively). WM1 and WM2 did not have significant indirect effects on DT through IQ ( = .01 p = .638, = .00 p = .637, respectively) or on CT ( = .17 p = .227, = .05 p = .216, respectively). The indirect effects of WM1 and WM 2 on DT and CT through IQ and AF were also not significant.

PAGE 88

88 Altogether, results indicated that the Top2 and Snapshot divergent thinking test scoring methods produced strongly related scores. Scores on associative fluency assessed by the LetterF fluency task showed to be significantly related to other observed variables, whether scores on the LetterM fluency task was not. Results from the path analyses provide preliminary supporting evidence for the direct effect of working memory on intelligence, as well as t he direct effect of intelligence on associative fluency. Associative fluency predicted divergent thinking but not convergent thinking, providing mixed evidence for the associative basis of creative thinking. The direct effect of intelligence on divergent t hinking and convergent thinking were not significant, indicating that the role of intelligence on these two creative thinking processes may be indirect (through associative fluency).

PAGE 89

89 CHAPTER 7 RESULTS The scores from the battery of processing speed, working memory, intelligence, associative fluency, divergent thinking, and convergent thinking tasks were analyzed using SPSS version 19 and Mplus 6 The descriptive statistics for each of the m easures are presented in Table 71 Table 7 1 Descriptive s tatistics of observed variables Observed Variable Mean Standard Deviation Minimum, Maximum Skewness Kurtosis 1. BackDigit 5.67 1.55 0, 9 .39 .77 2. SymSpan 27.13 8.73 4, 41 .59 .26 3. Pattern1 17.62 4.85 3, 30 .20 .15 4. Pattern2 19.93 4.01 6, 33 .41 .77 5. PatternT 37.40 8.03 9, 56 .37 .61 6. Letter1 11.06 2.96 2, 19 .05 .26 7. Letter 2 11.39 2.57 4, 20 .33 .38 8. LetterT 22.40 4.98 2, 38 .08 .97 9. RAPMT 6.74 2.67 0, 12 .187 .585 10. WAISRVT 57.27 8.93 23, 70 1.14 1.89 11. LetterF 23.10 5.50 9, 40 .004 .28 12. Animals 26.99 5.79 12, 45 .478 .70 13. Jobs 21.06 4.55 9, 32 .102 .30 14. ATTAFlu 15.86 2.24 10, 19 .310 .62 15. ATTAOri 15.21 2.35 8, 19 .144 .84 16. ATTAElab 15.90 2.35 6, 19 .660 .64 17. ATTAFlex 15.32 2.29 1, 19 1.73 8.42 18. CRCVerb 2.06 1.19 0, 8 .88 2.44 19. CRCFig 5.85 2.53 0, 13 .41 .30 20. CRCT 7.91 3.16 2, 16 .37 .25 21. ATTACI 4.07 1.33 1, 7 .02 .52 22. WCHFlu 8.49 3.77 1, 20 .63 .04 23. WCHOri 2.66 .69 1, 5 .37 .18 24. RAT 16.37 5.22 2, 28 .37 .09 25. InsightT 1.64 1.10 0, 5 .37 .24 26. OE 10.57 2.12 5, 14 .24 .69 27. TCS16 67.70 9.19 45, 93 .02 .26 28. TCS3 11.71 4.07 3, 18 .01 1.02 BackDigit = Backward Digit Span, SymSpan = Symmetry Span, Pattern1, 2 = Pattern Comparison Task Part I and II, PatternT = Pattern Comparison task total score, Letter1, 2 = Letter Comparison Task Part I and II, LetterT = Letter Comparison task total score, RAPM = Ravens Advanced Progressive Matrices, WAISR V = Weschler Adult Intelligence Scale Revised Vocabulary subset, Letter F = Letter fluency task, Animals = Category fluency task (animals), Jobs = Category fluency task (Jobs), ATTA = Abbreviated Torranc e Test for Adults (Flu = Fluency, Ori = Originality, Elab = Elaboration, Flex = Flexiblity, scaled scores), CRC = Criterion Referenced Creativity Indicators (Ver = Verbal, Fig = Figural, Tot = Tally of CRC indicators), CI = Creativity Index, WCH = Unusual Uses task (wire coat hanger, Flu = fluency, Ori =

PAGE 90

90 Originality), RAT = Remote Associations Task, Insight 15 = Dot problem, Letter Scramble1, Letter Scramble 2, Letter problem, Mathematical problem, Insight = Total score for 5 Insight Problems, OE = Openness to Experience, TCS16, 13 = Theories of Creativity Scale (16 item and 3 item scale scores). Table 7 2 Description of latent variables and indicators Higher order latent variable First order latent variable Indicators Intelligence ( g ) Working Memory (WM) Symmetry Span (SymSpan) total score Backwards Digit Span (BackDigit) total score Processing Speed (PS) Pattern Comparison Task (Pattern) scores include total score (PatternT) and Part I (Pattern1) and Part II (Pattern2) total scores Letter Comparison Task (Letter) scores include total score (LetterT) and Part I (Letter1) and Part II (Letter2) total scores Intelligence (IQ) Ravens Advanced Progressive Matrices (RAPM) short version (12 items) total score Weschsler Adult Intelligence ScaleRevised, Vocabulary (WAISRV) total score Fluid intelligence (Gf) Ravens Advanced Progressive Matrices (RAPM) short version (12 items) total score Crystallized intelligence (Gc) Weschsler Adult Intelligence ScaleRevised, Vocabula ry (WAISRV) total score Creativity Associative Fluency (AF) Letter fluency task (LetterF) total score Category fl uency task (Animals) total score Category fluency task ( Jobs) total score Divergent Thinking (DT) Abbreviated Torrance Test for Adults (ATTA) scores include : Creativity I ndex (CI) composite, normalized score; fluency (flu), originality (ori), elaboration (elab), and flexibility (flex) standardized scores; and criterion reference creativity indicators for verbal (CRCVerb) and figural (CRCF ig) total scores Unusual Uses task (wire coat hanger, includes fluency (WCH Flu) score and originality (WCHOri ) score Convergent Thinking (CT) Remote Associates Test (RAT) total score Ins ight Problems (Insight), includes total score (InsightT), and it em scores (Insight 1 to Insight5)

PAGE 91

91 Table 7 3 Correlat ions between observed variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 1. BackDigit 1.00 2. SymSpan .23** 1.00 3. Pattern1 .12 .20** 1.00 4. Pattern2 .15* .24** .59** 1.00 5. PatternT .15* .25** .89** .87** 1.00 6. Letter1 .27** .12 .20** .35** .31** 1.00 7. Letter 2 .16* .14* .32** .42** .42** .55** 1.00 8. Letter Tot .24** .12 .30** .44** .42** .90** .86** 1.00 9. RAPM .22** .29** .06 .11 .09 .03 .02 .02 1.00 10. WAISR V .28** .11 .05 .08 .10 .16** .18** .22** .22** 1.00 11. LetterF .23** .09 .14* .15* .15* .20** .15* .20** .19** .30** 1.00 12. Animals .19** .11 .08 .10 .11 .20** .20** .23** .15* .33** .42** 1.00 13. Jobs .18** .09 .04 .08 .08 .24** .17** .24** .07 .29** .35** .43** 1.00 14. ATTAFlu .02 .04 .15* .14* .16** .14* .12 .15* .03 .19** .19** .21** .18** 1.00 15. ATTAOri .02 .03 .14* .12* .13* .12 .13* .13* .08 .02 .19** .09 .13* .19** 1.00 16. ATTAElab .05 .01 .13* .12 .14* .06 .08 .07 .14* .17** .26** .20** .27** .34** .24** 1.00 17. ATTAFlex .07 .03 .09 .16** .15* .04 .08 .07 .23** .16** .12 .26** .05 .51** .02 .38** 1.00 18. CRCVerb .12 .02 .12 .05 .09 .09 .02 .06 .04 .24** .09 .13* .12 .39** .25** .22** .18** 1.00 19. CRCFig .12 .05 .09 .11 .11 .13* .14* .15* .06 .20** .25** .17** .31** .23** .42** .51** .10 .54** 1.00 20. CRCTot .14* .03 .11 .11 .12* .14* .12 .14 .06 .25** .24** .22** .29** .33** .43** .49** .15* .71** .36** 1.00 21. ATTACI .09 .004 .11 .17** .15* .18** .15* .19** .90 .25** .28** .24** .29** .65** .53** .65** .50** .77** .66** .94** 1.00 22. WCHFlu .01 .07 .04 .03 .04 .18** .09 .15* .01 .18** .17** .27** .23** .40** .26** .15* .02 .33** .23** .17** .23** 1.00 23. WCHOri .06 .12 .08 .18** .16* .14* .08 .12 .09 .18** .22** .19** .23** .11 .14* .21** .03 .15* .03 .16** .14* .30** 1.00 24. RAT .26** .13* .02 .02 .02 .16* .12 .16* .32** .40** .22** .19** .15* .03 .12 .08 .07 .09 .07 .15* .15* .05 .06 1.00 25. Insight1 .17** .16* .01 .02 .01 .09 .02 .08 .27** .25** .08 .16** .08 .10 .10 .14* .09 .15* .10 .14* .15* .15* .11 .13* 1.00 26. Insight2 .05 .13* .07 .02 .04 .14* .13* .15* .11 .13* .10 .01 .02 .01 .05 .05 .02 .05 .02 .03 .02 .01 .00 .17** .07 1.00 27. Insight3 .07 .08 .01 .07 .05 .13* .17** .17** .04 .16* .15* .09 .16* .07 .10 .04 .01 .00 .19** .15* .13* .11 .06 .07 .02 .23** 1.00 28, Insight4 .02 .11 .02 .05 .05 .09 .02 .07 .11 .14* .07 .16** .01 .19** .00 .10 .13* .03 .02 .02 .10 .07 .01 .00 .07 .07 .02 29. Insight5 .13* .11 .01 .06 .02 .12 .05 .10 .01 .09 .01 .09 .08 .02 .09 .04 .09 .06 .02 .04 .01 .06 .02 .08 .05 .07 .06 30. InsightT .17** .24** .05 .08 .07 .12 .12 .14* .21** .30** .15* .19** .13* .09 .05 .14* .11 .10 .13* .14* .14* .07 .07 .20** .50** .60** .49** 31. OE .02 .06 .03 .01 .02 .07 .07 .05 .05 .03 .06 .08 .18** .17** .11 .18** .09 .17** .19** .22** .22** .18** .06 .05 .09 .07 .02 32. TCS16 .07 .03 .05 .11 .07 .05 .04 .06 .06 .10 .16* .19** .04 .03 .04 .08 .01 .07 .04 .05 .04 .14* .06 .08 .18** .06 .06 33. TCS3 .13* .07 .05 .09 .06 .004 .03 .04 .07 .11 .11 .16* .06 .03 .05 .11 .03 .10 .07 .11 .10 .01 .11 .18** .05 .02 .06

PAGE 92

92 Table 7 3. Continued 28 29 30 31 32 33 1. BackDigit 2. SymSpan 3. Pattern1 4. Pattern2 5. PatternT 6. Letter1 7. Letter 2 8. Letter Tot 9. RAPM 10. WAISR V 11. LetterF 12. Animals 13. Jobs 14. ATTAFlu 15. ATTAOri 16. ATTAElab 17. ATTAFlex 18. CRCVerb 19. CRCFig 20. CRCTot 21. ATTACI 22. WCHFlu 23. WCHOri 24. RAT 25. Insight1 26. Insight2 27. Insight3 28, Insight4 1.00 29. Insight5 .08 1.00 30. InsightT .37** .54** 1.00 31. OE .0.05 .02 .08 1.00 32. TCS16 .10 .00 .15* .27** 1.00 33. TCS3 .00 .12* .25* .78** .12 1.00

PAGE 93

93 MPlus 6 was used to run the proposed SEM model. The labels of the higher order and first order latent variables, as well as the corresponding indicators are presented in Table 7 2 The correlations among all of the observed variables including working memo ry, processing speed, intelligence, associative fluency, divergent thinking, convergent thinking, theories of creativity, and demographics are presented i n Table 7 3. Original Models ( Using Parcel Indicators ) Model 1a: The Relationships a mong Wor king Memory, Processing Speed, a nd Intelligence Model 1a CFA Prior to testing the proposed model, the relationships among Working Memory (WM), Processing Spe ed (PS), and Intelligence (IQ) were explored. The indicators of WM included the total Symmetry Span score (SymSpan) and the score on the Backward Digits task (BD). Indicators of PS included the total score on the Pattern Comparison task (PatternT) and the total score on the Letter Comparison task (LetterT). Indicators of IQ included the total score on the R avens Advanced Progressive Matrices (RAPMT) and the total score on Weschler Adults Intelligence Scale RevisedVocabulary (WAISRT). WM and PS were both specified to predict IQ. A CFA of the measurement model was conducted. The goodness of fit (GOF) indice s showed that the model met some of the criteria for good fit ( 2 = 22.53, df = 6, p = .001, RMSEA = .10, CFI = .86, TLI = .64, SRMR = .04). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.51), SymSpan (.45 ), PatternT (.65), LetterT (.61), RAPMT (.50), and WAISTot (.44), p < .001. The

PAGE 94

94 correlations between WM and PS, WM and IQ, and PS and IQ were r = .61, 1.0, p < .001, r = .37, p < .05 respectively Modification indices showed that allowing the residuals between PatternT and SymSpan to correlate would improve model fit (MI = 7.95, EPC = 12.91). This respecification is justifiable, because both tasks require problem solving with respect to geometric shapes (i.e., determining if the geometric shapes are the sa me or different, symmetrical or nonsymmetrical), and therefore are likely to engage similar processes and share method variance. The GOF indices of the new model showed that modification improved the fit to the data ( 2 = 13.04, df = 5, p = .02, RMSEA = .08, CFI = .93, TLI = .79, SRMR = .04). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.58), SymSpan (.39), PatternT (.51), LetterT (.80), RAPMT (.45), and WAISRT (.48), p < .001. Th e residuals of Patter nT and SymS pan were correlated at r = .22, p < .01 Model 1a SEM The structural model was embedded into the m easurement model (Figure 7 1 ). The GOF indices of the structural model showed adequate fit to the data ( 2 = 13.04, df = 5, p = .02, RMSEA = .08, CFI = .93, TLI = .79, SRMR = .04). WM significantly predicted IQ ( = .99, p < .01), however PS did not significantly predict IQ ( = .08, p = .78). WM and PS correlated at r = .46, p < .001. The residuals of PatternT and SymSpan correlated at r = .22, p < .01. Additional respecifications to the model based on the modification indices could not be substantively justified; therefore, no further modifications to the model were made. Because PS did not significantly predict IQ, the path from PS to IQ was deleted from the model. An alternative model in which WM predicted both IQ and P S was

PAGE 95

95 tested, and PS predicted I Q. In other words, PS was specified as partial mediator of WM and IQ. A similar model was tested in a study conducted by Unsworth and colleagues ( 2009) that showed evidence to support a partial mediation model where the recall component of complex span tasks was specified to have a direct effect on the processing time component ( = .52) and fluid intelligence ( = .42), and the processing time com ponent was also specified to have a direct effect on fluid intelligence ( = .58), p < .05. Their results indicated that the processing time component of complex span tasks accounts for some but not all of the relationship between the recall component and fluid intelligence (Unsworth et al., 2009). The GOF indices of the structural model showed adequate fit to the data ( 2 = 13.04, df = 5, p = .02, RMSEA = .08, CFI = .93, TLI = .79, SRMR = .04). WM significantly predicted IQ and PS ( = .99, .46, p < .01 respectively ), however PS did not significantly predict IQ ( = .08, p = .78). WM and PS correlated at r = .46, p < .001. The residuals of PatternT and SymSpan correlated at r = .22, p < .01. Additional respecifications to the model based on the modific ation indices could not be substantively justified; therefore, no further modifications to the model were made. The insignificant path from PS to IQ was removed. The GOF indices of the final model showed good fit to the data ( 2 = 13.03, df = 5, p = .02, R MSEA = .08, CFI = .93, TLI = .79, SRMR = .04). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.58), SymSpan (.39), PatternT (.51), LetterT (.80), RAPMT (.45), and WAISTot (.48), p < .001. WM predicted both IQ and PS ( = .96, .46, p < .001 respectively ). Th e residuals of PatternT and SymS pan were correlated at r = .22, p <

PAGE 96

96 .001. The proportion of variance accounted for by BD, SymSpan, PatternT, LetterT, and RAPMT was 34%, 16%, 26%, 63%, 21%, 23%, respectively. Figure 7 1 Relationships among working memory, processing speed, and intelligence M odel 1 b : The Relationships a mong Associativ e Fluency, Divergent Thinking, a nd Convergent Thinking Model 1b CFA The relationships among associative fluency (AF), divergent thinking (DT), and convergent thinking (CT) were explored. The indicators of AF included the total score on the letter fluency (LetterF) and category fluency tasks (Animals, Jobs). Indicators of DT included the creativity index on the Abbreviated Torrance Test for Adults (ATTACI) and the originality score on the Unusual Uses task for the wire coat hanger (WCHO). Indicators of CT included the total score on the Remote Associates Test (RAT) and the total score on the five insight problems (I nsightT). AF was specified to predict DT and CT. A CFA of the measurement model was conducted. The goodness of fit (GOF) indices showed that the model was a good fit to the data ( 2 = 6.04, df = 11, p = .87,

PAGE 97

97 RMSEA < .001, CFI = 1.0, TLI = 1.0, SRMR = .02) The standardized factor loadings of the indicators for each latent variable was as follows: LetterF (.61), Animals (.68), Jobs (.61), ATTACI (.46), WCHOri (.35), RAT (.49), InsightT (.40), p < .001. The correlations between AF and DT, AF and CT, and DT a nd CT were r = .96, .63, p < .001, and .49, p < .05, respectively. Model 1b SEM The structural model was embedded into the measurement model (Figure 72 ). 2 = 6.04, df = 11, p = .87, RMSEA < .001, CFI = 1.0, TLI = 1.0, SRMR = .02). The standardized factor loadings of the indicators for each latent variable was as follows: LetterF (.61), Animals (.68), Jobs (.61), ATTACI (.46), WCHOri (.35), RAT (.49), InsightT (.40), p < .001. Figure 7 2 Relationships among associative fluency, divergent thinking, and convergent thinking Results showed that AF was a significant p < .001, respectively). DT and CT were negatively correlated at r = .54, p = .74. The

PAGE 98

98 proportion of variance accounted for by LetterF, Animals, Jobs, ATTACI, WCHO, RAT, and InsightT was 37%, 47%, 38%, 21%, 12%, 24%, and 16%, respectively. Model 1: The Relationships a mong Processing Speed, Working Memory, Intelligence, As sociativ e Fluency, Divergent Thinking, a nd Convergent Thinking Model 1 CFA The models from model 1a and model 1b were combined to explore the relationships among constructs related to intelligence (WM, PS, and IQ) and creative thi nk ing (AF, DT, and CT) T he residuals between PatternT and SymSpan were specified to correlate. A CFA of the measurement model was conducted. The goodness of fit (GOF) indices showed that the model was a good fit to the data ( 2 = 44.38 df = 49, p = .66, RMSEA < .001, CFI = 1.0, T LI = 1.0, SRMR = .04). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.61), SymSpan (.38), PatternT (.53), LetterT (.76), RAPMT (.39), WAISRT (.56), LetterF (.62), Animals (.68), Jobs (.61), ATTACI (.45), W CHOri (.35), RAT (.51), Insight (.38), p < .001. The residuals between PatternT and SymSpan correlated at r = .23, p < .01. Model 1 SEM The structural model was embedded into the measurement model (F igu re 7 3 ). WM was specified to predict IQ and PS. As hypothesized, IQ was specified to predict AF, DT, and CT. AF was specified to predict DT and CT. The GOF indices showed that the structural model was a good fit to the data ( 2 = 50.53, df = 54, p = .61, RMSEA < .001, CFI = 1.0, TLI = 1.0, SRMR = .04). A m odel comparison test was conducted to test the decline in fit. The chi square test statistic of the latent variable path model was

PAGE 99

99 subtracted by the chi square test statistic of the measurement model; 2 (54) = 50.532 (49) 44.38 = 2 (5) = 6.15 Beca use this value is less than the critical value 2 05, 5 =12.59, I accept the null hypothesis and conclude that the structural model fits the data better. The standardized factor loadings of the indicators were as follows: BD (.58), SymSpan (.37), PatternT (.53), LetterT (.76), RAPMT (.41), WAISRT (.63), LetterF (.62), Animals (.68), Jobs (.61), ATTACI (.45), WCHOri (.35), RAT (.51), InsightT (.38), p < .001. Results showed that WM significantly predicted IQ and PS ( = .71, .54, p < .001, respectively). IQ was a significant predictor of AF and CT ( = .72, 1.72, p < .001, respectively). IQ predicted DT but this relationship was not significant ( = .03, p = .91). AF was a significant predictor of DT ( = .93, p = .001) but not of CT ( = .62, p = .14). Thu s, the results suggest that working memory significantly predicts intelligence as well as processing speed. Intelligence predicted two of the three creative thinking processes (associative fluency and convergent thinking), but not divergent thinking. In co ntrast to model 1b, associative fluency predicted only divergent thinking and not convergent thinking in this model. The insignificant paths between IQ to DT and AF to CT were removed (Figure 74 ). The GOF indices showed that the final structural model was a good fit to the data ( 2 = 55.60, df = 56, p = .49, RMSEA < .001, CFI = 1.0, TLI = 1.0, SRMR = .04). The standardized factor loadings of the indicators were as follows: BD (.59), SymSpan (.37), PatternT (.54), LetterT (.75), RAPMT (.43), WAISRT (.66), LetterF (.62), Animals (.68), Jobs (.61), ATTACI (.44), WCHOri (.35), RAT (.51), InsightT (.38), p < .001.

PAGE 100

100 Figure 7 3. Model 1 SEM : IQ predicts associative fluency, divergent thinking, and convergent thinking Results showed that WM significantly predicte d IQ and PS ( = .77, .51, p < .001, respectively). IQ was a significant predictor of AF and CT ( = .64, 1.17, p < .001, respectively). AF was a significant predictor of DT ( = .96, p < .001). The residuals between PatternT and SymSpan correlated at r = .23, p < .01. Figure 7 4. Model 1 SEM (insignificant paths removed)

PAGE 101

101 However, there was a problem with the model, indicated by the standardized path coefficient with a value greater than 1 from IQ to CT. It is possible that the model was not properly identified due to bad start values or complexity, or that there are larger specification problems in the measurement and/or structural portions of the model. One weakness of this model was that there were only two indicators per latent variable, with the exception of AF. Problems that have been cited with two indicator factors include empirically underidentified models, nonconvergence or iterative estimation, and specification errors (Kline, 2011), such as the path coefficient greater than 1 which was the c ase in the present model. Model 2 : The Relationships a mong Processing Speed, Working Memory, Intelligence, Associative Fluency, Divergent Th inking, a nd Convergent Thinking An alternative model was tested, in which IQ was specified to predict only AF, and indirectly predict DT and CT. WM was specified to predict both PS and IQ, and AF was specified to predict DT and CT. The GOF indices showed that the model was a good fit to the data data ( 2 = 92.41, df = 56, p = .002, RMSEA = .05, CFI = .92, TLI = .88, SR MR = .05). The standardized factor loadings of the indicators were as follows: BD (.61), SymSpan (.35), PatternT (.54), LetterT (.74), RAPMT (.38), WAISRT (.65), LetterF (.59), Animals (.63), Jobs (.55), ATTACI (.45), WCHOri (.35), RAT (.51), InsightT (.38), p < .001. The structural model was embedded into the measurement model (Figure 7 5 ). The GOF indices showed that the final structural model was a good fit to the data ( 2 = 92.41, df = 56, p = .002, RMSEA = .005, CFI = .92, TLI = .88, SRMR = .05). The standardized factor loadings of the indicators were as follows: BD (.61), SymSpan (.35), PatternT (.54), LetterT (.74), RAPMT (.38), WAISRT (.65), LetterF (.59), Animals (.63),

PAGE 102

102 Jobs (.55), ATTACI (.45), WCHOri (.35), RAT (.51), InsightT (.38), p < .001. Res ults showed that WM significantly predicted IQ and PS ( = .73, .54, p < .001, respectively). IQ was a significant predictor of AF ( = .88, p < .001). AF significantly predicted DT and CT ( = .97, .91, p < .001, respectively). WM had an indirect effect o n DT and CT through IQ and AF ( = .62, .59, p < .001, respectively). WM also had an indirect effect on AF through IQ ( = .65, p < .001). IQ had an indirect effect on DT and CT through AF ( = .85, .81, p < .001, respectively). The residuals between PatternT and SymSpan correlated at r = .22, p < .01. The proportion variance explained by BD, SymSpan, PatternT, LetterT, RAPMT, WAISRT, LetterF, Animals, Jobs, ATTACI, WCHO, RAT, and InsightT was 37%, 12%, 30%, 55%, 14%, 42%, 35%, 39%, 31%, 20%, 13%, 26%, and 14%, respectively. Figure 7 5 Model 2: Relationships among working memory, processing speed, intelligence, associative fluency, divergent thinking, and convergent thinking

PAGE 103

103 Overall, results showed that working memory predicted processing speed and intelligence directly and indirectly predicted the three creative thinking processes (i.e., associative fluency, divergent thinking, and convergent thinking). Intelligence predicted associative fluency, and indirectly predicted divergent and convergent thi nking through associative fluency. Associative processing directly predicted divergent and convergent thinking. The models above used parcels (i.e., aggregatelevel indicators comprised of the sum or average of two or more items) (Little, Cunningham, Shahar, & Widaman, 2002) as indicators of the latent variables. In the SEM literature, this practice has been considered controversial for several reasons. Several theoretical reasons against parceling include (1) the data should be as close of the participant s responses as possible in order to avoid arbitrary manufacturing, (2) the lack of justification for constructs that are multidimensional and heterogeneous and (3) failing to represent all sources of variance in each item ( Little, Lindenberger, & Nesselr oade, 1999; Little et al., 2002). On the other hand, proponents of parceling argue that item level data have lower reliability, lower communality, smaller common to unique factor variance ratios, and lower likelihood of normal distributions (Kline, 2011; Little et al., 2002). Furthermore, the overall fit of the model when using parcels tend to be more parsimonious, have fewer correlated residuals or dual loadings, and have reduced sources of sampling error (Kline, 2011; MacCallum, Widaman, Zhang, & Hong, 19 99). Taking these considerations into account, the following competing models were also tested using item level indicators.

PAGE 104

104 Original Models ( Using Item Level Indicators) Model 3a: The Relationships a mong Processing Spe ed, Working Memory, a nd Intelligence Model 3 a CFA A model exploring the relationships among PS, WM, and IQ was tested using item level indicators. In addition, this model did not include the total score on the WAISR V (a measure of crystallized intelligence) in order to specify a more pure me asure of fluid intelligence assessed by scores on the RAPM items. A confirmatory factor analysis (CFA) of the m easurement model was conducted. The goodness of fit (GOF) indices showed that the model was a good fit to the data ( 2 = 204.52, df = 132, p < .001, RMSEA = .05, CFI = .88, TLI = .86, SRMR = .05). The standardized factor loadings of the indicators were as follows: BD (.43), SymSpan (.54), Pattern1 (.65), Pattern2 (.80), Letter1 (.52), Letter2 (.60), and RAPM Item 1 to 12 (.31, .45, .51, .41, .49, .39, .49, .48, .31, .38, .45, and .31, respectively). The modification indices showed that allowing the residuals between Part I and II of the Letter Comparison task to correlate would improve the fit of the model (MI = 59.85, EPC = 3.15). This change is substantively justifiable because Part I and II of this task consist of the same type of items, and therefore, the process of answering these items have a high likelihood of sharing method variance. Similarly, the residuals between Part I and II of the Pat tern Comparison were allowed to correlate (MI = 50.49, EPC = 10.75). The goodness of fit (GOF) indices showed that the new model was a good fit to the data ( 2 = 143.84, df = 130, p = .19, RMSEA = .02, CFI = .98, TLI = .97, SRMR = .05). The standardized factor loadings of the indicators were as follows: BD (.43),

PAGE 105

105 SymSpan (.54), Pattern1 (.55), Pattern2 (.75), Letter1 (.48), Letter2 (.57), and RAPM Item 1 to 12 (.31, .45, .51, .41, .49, .39, .49, .48, .31, .38, .45, and .31, respectively). The residuals between Part I and II of the Pattern and Letter Comparison indicators correlated at r = .32 and r = .41, p < .001, respectively. Model 3 a SEM Because the model reached the criterion for good fit, and the remaining modification indices were not theoretically justifiable, the structural model was embedded into the measurement model (Figure 76) WM and PS were specified to predict IQ. The GOF indices of the structural model showed good fit to the data ( 2 = 143.84, df = 130, p = .19, RMSEA = .02, CFI = .98, TLI = .97, SRMR = .05). WM significantly predicted IQ ( = .77, p < .001), however PS did not significantly predict IQ ( = .28 p = .22). WM and PS correlated at r = .52, p < .001. The residuals between Part I and II of the Patter and Letter Comparison indicators correlated at r = .32 and r = .41, p < .001, respectively. The remaining modification indices were not theoretically justifiable, therefore, the model was not respecified. Again an alternative model was tested in which PS was specified as partial mediator of WM and IQF, based on Unsworth and colleagues (2009) study (Figure 86 ). The GOF indices of the structural model showed good fit to the data ( 2 = 143.84, df = 130, p = .19, RMSEA = .02, CFI = .98, TLI = .97, SRMR = .05). WM significantly predicted IQ and PS ( = .77, .52, p < .001, respectively), however PS did not significantly predict IQ ( = .28 p = .22). The remaining modification indices were not theoretically justifiable; therefore, the model was not respecified. In addition, the insignificant path from PS to IQ was removed. The GOF indices of the final structural model showed good fit to the data ( 2 = 143.84, df = 130, p = .19, RMSEA = .02, CFI =

PAGE 106

106 .98, TLI = .97, SRM R = .05). The standardized factor loadings of the indicators were as follows: BD (.43), SymSpan (.54), Pattern1 (.55), Pattern2 (.75), Letter1 (.48), Letter2 (.57), and RAPM Item 1 to 12 (.31, .45, .51, .41, .49, .39, .49, .48, .31, .38, .45, and .31, resp p < .001, respectively). The residuals between Part I and II of the Pattern and Letter Comparison indicators correlated at r = .32 and r = .41, p < .001, respectively. Altogether, results showed that working memory predicted both processing speed and intelligence, and processing speed did not predict intelligence. Figure 7 6 Model 3a : Relationships among working memory, processing speed, and inte lligence (item level indicators)

PAGE 107

107 Model 3b: The Relationships a mong Associativ e Fluency, Divergent Thinking, a nd Convergent Thinking Model 3 b CFA A model exploring the relationships among AF, DT, and CT was tested using item level indicators. LetterF, Animals, and Jobs were specified as indicators of AF. The components of the norm referenced (ATTA Fluency, ATTA Originality, ATTA Elaboration, ATTA Flexibility) and criterionreferences (CRC Verbal, CRC Figural) scores on the ATTA, and the Fluency (WCHFluency) and Originality (WCH Originality) scores on the Unusual Uses task were specified as indicators of DT. The score on the RAT and the score on each of the five Insight problems (Insight1 to Insight5) were specified as indicators of CT. A confirmatory factor a nalysis (CFA) of the measurement model was conducted. The goodness of fit (GOF) indices showed that the model was not an adequate fit to the data ( 2 = 366.46, df = 116, p < .001, RMSEA = .09, CFI = .78, TLI = .74, SRMR = .08). The standardized factor loadings of the indicators for each latent variable was as follows: LetterF (.60), Animals (.69), Jobs (.62), ATTA (Flu = .32, Ori = .44, Elab = .51, Flex = .15), CRCVerb (.95), CRCFig (.98), WCHFlu (.23), WCHOri (.16), RAT (.45), Insight items 1 to 5 (.27, .2 9, .32, .13, and .18, respectively), p < .001. The factor loadings for ATTAFlex, Insight items 2 and 5 were significant at p < .05, and for Insight item 4, nonsignificant at p = .19. The correlations between AF and DT, AF and CT, and DT and CT were r = .4 2, .59, p < .001, and .34, p <.01 respectively. Modification indices showed that allowing the residuals between the criterion referenced creativity indicators on the ATTA (CRCVerb and CRCFig) would significantly improve the model (MI =

PAGE 108

108 60.92, ECP = 6.27). This modification is theoretically justifiable, as the criterion referenced indicators for the verbal and figural activities on the ATTA shared overlapping criteria (e.g., richness and colorfulness of imagery, expression of feelings and emotions) to receive additional points on the ATTA, and therefore are likely to engage similar underlying processes and share method variance. The GOF indices showed that the respecified modification improved the model ( 2 = 306.47, df = 115, p < .001, R MSEA = .08, CFI = .83, TLI = .80, SRMR = .07). The next largest modification index suggested allowing the residuals between ATTAFlex (Flexibility) and ATTAFlu (Fluency) correlate (MI = 48.26, ECP = 1.87). This is also a justifiable modification based on Be nedek et al.s (2012) study that showed a moderate to large correlation ( r = .76) between the flexibility and fluency components of associative processing, and also indicated that the fluent generation of disassociations and associative combinations draw on fluent and flexible association processes (p. 7), suggesting that fluency and flexibility share overlapping processes and share method variance. The GOF indices showed that this modification improved the model ( 2 = 255.40, df = 114, p < .001, RMSEA = .07, CFI = .87, TLI = .85, SRMR = .06). The last modification made to the model allowed the residuals between WCHF and ATTAFlu to correlate (MI = 30.92, ECP = 2.92) because both on these indicators are scored based on the tally of responses generated, and are likely to engage similar freeassociation processes. The GOF indices showed that this modification improved the model ( 2 = 221.43, df = 113, p < .001, RMSEA = .06, CFI = .90, TLI = .88, SRMR = .06). The standardized factor loadings of the indicators for each latent variable was as follows:

PAGE 109

109 LetterF (.61), Animals (.67), Jobs (.62), ATTA (Flu = .44, Ori = .48, Elab = .65, Flex = .27), CRCVerb (.72), CRCFig (.74), WCHFlu (.35), WCHOri (.30), RAT (.43), Insight items 1 to 5 (.28, .31, .32, .14, and .18, r espectively), p < .001. The factor loadings for ATTAFlex, Insight items 1, 2, 3 and 5 were significant at p < .05, and for Insight item 4, nonsignificant at p = .16. The proportion of variance in LetterF, Animals, and Jobs that was associated with AF was 37.6%, 45%, and 39%, respectively. The proportion of variance in ATTAFlu, ATTAOri, ATTAElab, ATTAFlex, CRCVerb CRCFig, WCHFlu, and WCHOri, that was associated with DT was 19.1%, 22.9%, 42.3%, 7.2%, 51.9%, and 55%, respectively. The proportion of variance in RAT, and Insight Items 1 to 5 that was associated with CT were 18.7%, 7.8%, 9.7%, 10%, 2%, and 3.2%, respectively. The correlations between AF and DT, AF and CT, and DT and CT were r = .59, .58, p < .001, and .36, p <.01respectively. The correlations b etween the residuals of CRCVerb and CRCFig, ATTAFlex and ATTAFlu, and ATTAFlu and WCHF were r = .86, .47, .33, p < .001, respectively. Because the model reached the criterion for satisfactory fit, and the next set of large modification indices were not the oretically justifiable, additional respecifications were not made. Model 3 b SEM The structural model was embedded into the measurement model (Figure 7 7 ). AF was specified to predict DT and CT. The GOF indices showed that the structural model met some of the criteria for good to the data ( 2 = 222.39, df = 115, p < .001, RMSEA = .06, CFI = .90, TLI = .89, SRMR = .06). Standardized factor loadings of the indicators were as follow: LetterF (.61), Animals (.68), Jobs (.62), ATTA (Flu = .45, Ori = .48, Elab = .65, Flex = .28), CRCVerb (.72), CRCFig (.75), WCHFlu (.35), WCHOri (.30), RAT (.37), Insight items 1 to 5 (.28, .25, .30, .15, and .16, respectively), p < .001. The

PAGE 110

110 factor loadings of ATTAFlex, Insight items 1, 2, and 3 were significant at ( p < .05) and o f Insight item 4 was non significant ( p = .10). Results showed that AF was a significant predictor of DT ( = .30, p < .001) and CT ( = .57, p < .001). The proportion of variance in LetterF, Animals, and Jobs that was associated with AF were 37.6%, 46.1%, and 38%, respectively. The proportion of variance in ATTAFlu, ATTAOri, ATTAElab, ATTAFlex, CRCVerb CRCFig, WCHFlu, and WCHOri, that was associated with DT were 20.0%, 23.0%, 42.3%, 7.5%, 52.2%, and 55.6%, respectively. The proportion of variance in RAT, and Insight Items 1 to 5 that was associated with CT were 13.9%, 7.8%, 6.3%, 8.9%, 2.3%, and 2.5%, respectively. DT and CT were correlated at r = .45, p < .001. The correlations between the residuals of CRCVerb and CRCFig, ATTAFlex and ATTAFlu, and ATTAFlu and WCHF were r = .86, .47, .33, p < .001, respectively. Possible respecifications to the models based on the modification indices could not be substanti vely justified, therefore no modifications to the structural model was made. Figure 7 7 Model 3b Relationships among associative fluency, divergent thinking, and convergent thinking (item level indicators)

PAGE 111

111 Model 3: The Relationships a mong Processing S peed, Working Memory, Fluid Intelligence, Associativ e Fluency, Divergent Thinking, a nd Convergent Thinking (Item Level Indicators) A model exploring the relationship among PS, WM, IQF (fluid intelligence), AF, DT, and CT was tested. In contrast to the parceled indicators in model 1, this model used item level indicators. Based on the previous two models presented above, WM was specifi ed to predict PS and IQF, and AF was specified to predict DT and CT. In addition, IQF was specified to predict AF. The modific ations from the previous model 3a (WM, PS, and IQF) and model 3b (AF, DT, and CT) were included in the present model. Model 3 CFA A confirmatory factor analysis (CFA) of the measurement model was conducted. The goodness of fit (GOF) indices showed that the model was an adequate fit to the data ( 2 = 727.02, df = 540, p < .001, RMSEA = .04, CFI = .87, TLI = .85, SRMR = .06). The stan dardized factor loadings for the indicators of each latent variable were as follows: BD (.51) p < .001, SymSpan (.45), p < .001, Pattern 1, 2 (.39, .56, p < .001, respectively), Letter 1, 2 (.70, .73, respectively), RAPM items 112 (.31, .44, .51, .41, .49, .39, .49, .48, .30, .40, .45, and .30, p < .001, respectively), LetterF (.62), Animals (.68), Jobs (.61) p < .001 ATTA Fluency, Originality, Elaboration, Flexi bility (.44, .48, .65, .27, p < .001, respectively), CRCVerb (.44), CRCFig (.72), WCH Flu (.36), WCHOri (.31), RAT (.46), p < .001 Insight items 1 5 (.36, p < .001; .23, p < .001; .20, p < .01 ; .15, p < .0 5; .12, p = .13, respectively).

PAGE 112

112 Figure 7 8 Model 3 : Relationships among working memory, processing speed, intelligence, associative fluency, divergent thinking, and convergent thinking (item level indicators)

PAGE 113

113 Model 3 SEM WM was specified to predict IQ F and PS IQ F was specified to predict AF, a nd AF was specified to predict DT and CT (Figure 7 8) The GOF indices showed that the model was an adequate fit to the data ( 2 = 783.15, df = 547, p < .001, RMSEA = .04, CFI = .83, TLI = .81, SRMR = .07). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.43), p < .001, SymSpan (.60), p < .001, Pattern1 (.69), p < .001, Pattern 2 (.91), p < .001, Letter1 (.39), p < .001, Letter 2 (.48), p < .001, RAPM items 1 to 12 (.33, .45, .52, .41, .48, .39, .47, .48, .30, .38, .44, and .28, p < .001, respectively), LetterF (.62), p < .001, Animals (.67), p < .001, Jobs (.59), p < .001, ATTA Flu, Ori, Elab, and Flex (.44, .47, .66, and .28, p < .001, respectively), CRCVerb (.43), p < .001, CRCFig (.71), p < .001, W CHF (.35), p < .001, WCHO (.31), p < .001, RAT (.46), p < .001, Insight items 1 to 5 (.31, p < .001; .27, p < .01; .27, p < .01; .15, p = .09; and .16, p = .07 p < .01, .58, p < .001, respectively). IQ signi p < .001). AF significantly predicted DT and CT ( = .57, .70, p < .001). The latent variables PS, IQF, AF, DT, and CT accounted for 16%, 33%, 14%, 33%, and 49% of the variance in the model, respectively. Altogether results from the models using item leveled indicators showed the same pattern of relationships among processing speed, working memory, fluid intelligence, associative fluency divergent thinking, and convergent thinking as the models using parcel indicators.

PAGE 114

114 Model 4: The Relationships among I ntelligence ( g ) as a Higher Order Factor Associative Fluency, Divergent Thinking, a nd Convergent Thinking Model 4 CFA A model with intelligence modeled as a higher order latent variable was tested. This is based on argum ents made by some researcher that because intelligence is an abstract construct that explains the relationships among a diverse set of cognitive tests rather than the specific factors measured by the tests (Carroll, 1993; Spearman, 1904), when testing the relationship between intelligence and other constructs, intelligence should be modeled as a higher order, general factor composed of lower order cognitive skills (Kaufman et al., 2009; Silvia, 2008a, b). In this model, the higher order intelligence latent variable ( g ) was specified by three of the following first order factors: WM, PS, and IQ. SympSpan and BD were specified as indicators of WM. PatternT and LetterT were specified as indicators of PS. RAPMT and WAISRT were specified as indicators of IQ. Lett erF, Animals, and Jobs were specified as indicators of AF. ATTACI and WCHOri were specified as indicators of DT. RAT and InsightT were specified as indicators of CT. A confirmatory factor analysis (CFA) of the measurement model was conducted. The GOF indi ces showed that the model was an adequate fit to the data ( 2 = 67.68, df = 56, p = .14, RMSEA = .03, CFI = .97, TLI = .96, SRMR = .05). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.59), SymSpan (.39), P atternT (.48), LetterT (.81), RAPMT (.39), WAISRT (.57), LetterF (.62), Animals (.68), Jobs (.61), ATTACI (.45), WCHO (.36), RAT (.51), InsightT (.38), p < .001. The loadings of the first order latent variables WM, PS, and IQ on the secondorder latent

PAGE 115

115 var iable, g were .75, .40, 1.17, p < .001, respectively. There was a problem with identification, indicated by the path coefficient from IQ to g having a value greater than 1 ( r = 1.17). Modification indices showed that allowing the residuals between Pattern T and SymSpan to correlate would improve model fit (MI = 12.87, EPC = 13.90). This respecification is justifiable, because both tasks require problem solving with respect to geometric shapes (i.e., determining if the geometric shapes are the same or differ ent, symmetrical or nonsymmetrical), and therefore are likely to engage similar processes and share method variance. The new model showed that improved fit: ( 2 = 54.88, df = 55, p = .48, RMSEA < .001, CFI = 1.0, TLI = 1.0, SRMR = .04). The standardized f actor loadings for the indicators of each latent variable were as follows: BD (.59), SymSpan (.38), PatternT (.46), LetterT (.86), RAPMT (.38), WAISRT (.57), LetterF (.62), Animals (.68), Jobs (.61), ATTACI (.45), WCHO (.36), RAT (.51), InsightT (.38), p < .001. The loadings of the first order latent variables WM, PS, and IQ on the secondorder latent variable, g was .76, .38, 1.17, p < .001, respectively. The problem with identification persisted, indicated by the path coefficient from IQ to g having a value greater than 1 ( r = 1.17). Model 4 SEM The structural model was embedded in the measurement model (Figure 79 ). The higher order g factor was specified by PS, WM, and IQ, and g was specified to predict AF. AF was specified to predict DT and CT. GOF indices showed that the final model was an adequate fit to the data ( 2 = 96.90, df = 57, p < .001, RMSEA = .05, CFI = .91, TLI = .87, SRMR = .05). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.62), S ymSpan (.36), PatternT (.49), LetterT (.81), RAPMT (.36), WAISRT (.61), LetterF (.59), Animals (.62), Jobs (.55), ATTACI

PAGE 116

116 (.45), WCHO (.36), RAT (.51), and InsightT (.38), p < .001. The factor loadings of WM, PS, and IQ on g were .73, .44, and 1.01, p < .00 1, respectively. AF was significantly predicted by g ( = .87, p < .001), and AF significantly predicted DT and CT ( = .98, .92, p < .001, respectively). The residuals of SymSPan and PatternT correlated at r = .23, p < .001. However, there was a problem w ith identification, indicated by the factor loading of IQ on g with a value greater than 1. Figure 7 9 Model 4 : Relationships among higher order g associative fluency, divergent thinking, and convergent thinking (item level indicators) Model 5: The R elationships among I ntelligence ( g ) as a Higher Order Factor Associative Fluency, Divergent Thinking, a nd Convergent Thinking (Using Item Level Indicators) Model 5 CFA The relationships among WM, PS, IQ, and g as well as the relationships among g and AF, DT, and CT were explored. The higher order intelligence latent variable ( g ) was specified by the three following first ord er factors: WM, PS, and IQ. Sym Span and

PAGE 117

117 BD were specified as indicators of WM. Part I and II of the Pattern Comparison and the Letter Comparison tasks were specified as indicators of PS. Items 1 to 12 on the RAPM, and the WAISRV score were specified as indicators of IQ. The higher order g factor was specified to predict the three creative thinking processes, AF, DT, and CT. LetterF, Ani mals, and Jobs were specified as indicators of AF. The components of the norm referenced ( ATTA Fluency, ATTA Originality, ATTA Elaboration, ATTA Flexibility) and criterion references scores (CRC Verbal, CRC Figural) on the ATTA, and the Fluency (WCHFluenc y) and Originality (WCH Originality) scores on the Unusual Uses task were specified as indicators of DT. The score on the RAT and the score on each of the five Insight problems (Insight1 to Insight5) were specified as indicators of CT. A confirmatory factor analysis (CFA) of the measurement model was conducted. The goodness of fit (GOF) indices showed that the model was an adequate fit to the data ( 2 = 1025.02, df = 585, p < .001, RMSEA = .05, CFI = .70, TLI = .68, SRMR = .07). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.51), SymSpan (.46), Pattern 1, 2 (.61, .75, respectively), Letter 1, 2 (.57, .66, respectively), RAPM items 112 (.34, .43, .51, .40, .47, .39, .47, .49, .29, .40, .42, and .26, respecti vely), WAISRV (.37), LetterF (.62), Animals (.68), Jobs (.60), ATTA Fluency, Originality, Elaboration, Flexibility (.59, .44, .64, .40, respectively), CRCVerb (.49), CRCFig (.63), WCH Flu (.42), WCHOri (.29), RAT (.46), Insight items 1 5 (.36, .23, .20, .16, .13, respectively), p < .001. The factor loadings of Pattern1, Pattern2, and Letter2 were significant at p < .05. The factor loading of Letter1 was nonsignificant at p =.10. The loadings of WM, PS, and IQ on g were .92, .34, p < .001, and .73, respect ively, p < .05. Modification indices showed that allowing the residuals between

PAGE 118

118 Part I and II of the Pattern Comparison to correlate would improve the model fit (MI = 71.27, EPC = 10.61). The goodness of fit (GOF) indices showed that the new model improved the fit to the data ( 2 = 940.45, df = 584, p < .001, RMSEA = .05, CFI = .76, TLI = .74, SRMR = .07). Modification indices showed that allowing the residuals between ATTAFlu and ATTAFlex to correlate would improve model fit (MI = 49.40, EPC = 1.87). This respecification improved the fit of the model: ( 2 = 888.54, df = 583, p < .001, RMSEA = .04, CFI = .79, TLI = .78, SRMR = .07). Modification indices showed that allowing the residuals between ATTAFlu and ATTAElab to correlate would improve model fit (MI = 30.17, EPC = 1.30). This respecification is substantively justifiable based on Benedek et al.s (2012) study that showed a moderate to large correlations ( r = .55 to .77) between various components of associative processing, suggesting that fluency and elaboration share overlapping processes. Results showed improved the fit of the model: ( 2 = 855.90, df = 583, p < .001, RMSEA = .04, CFI = .82, TLI = .80, SRMR = .07). The modification indices showed that allowing the residuals between WCHF and ATTAFlu to correlate would improve the model fit (MI = 23.86, EPC = 1.95). Because both of these indicators are scored based on the tally of responses generated, they are likely to engage similar free association processes and therefore can be explained by method var iance. The GOF indices showed that this modification improved the model ( 2 = 830.3, df = 581, p < .001, RMSEA = .04, CFI = .83, TLI = .82, SRMR = .07). The last modification allowed the residuals between Insight2 and Insight3 to correlate. This was justified because items 2 and 3 on the Insight task were both word scramble problems, and therefore can be explained by method variance.

PAGE 119

119 The GOF indices of the final CFA model showed that this modification improved the model ( 2 = 819.48, df = 580, p < .001, RMSEA = .04, CFI = .84, TLI = .82, SRMR = .07). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.53), SymSpan (.44), Pattern 1, 2 (.37, .53, respectively), Letter 1, 2 (.72, .79, respectively), RAPM items 112 (.34, .43, .51, .40, .47, .39, .47, .49, .29, .40, .41, and .25, respectively), WAISRV (.38), Let terF (.62), Animals (.67), Jobs (.61), ATTA Fluency, Originality, Elaboration, Flexibility (.45, .49, .63, .19, respectively), CRCVerb (.46), CRCFig (.74), WCH Flu (.36), WCHOri (.30), RAT (.44), Insight items 1 5 (.34, .20, .19, .14, .10, respectively), p < .001. The factor loadings of Insight items 2,3, and 4 were significant at p < .05. The factor loading of Insight item 5 was nonsignificant at p =.22. The loadings of WM, PS, and IQ on g were .88, .36, .71, p < .001, respectively. The residuals between Pattern1 and Pattern2, ATTAFlu and ATTAFlex, ATTAFlu and WCHF, and Insight2 and Insight 3 were correlated at r = .51, .47, .30, .29, and .20, p < .001, respectively. Model 5 SEM The structural model was embedded in the measurement model (Figure 710). WM was specified to predict PS and g and g was specified to predict AF. AF was specified to predict DT and CT. GOF indices showed that the final model was an adequate fit to the data ( 2 = 856.53, df = 582, p < .001, RMSEA = .04, CFI = .82, TLI = .80, SRMR = .07). The standardized factor loadings for the indicators of each latent variable were as follows: BD (.54), SymSpan (.42), Pattern 1, 2 (.37, .54, respectively), Letter 1, 2 (.73, .78, respectively), RAPM items 112 (.34, .45, .52, .40, .47, .39, .46, .4 9, .30, .39, .42, and .27, respectively), WAISRV (.34), LetterF (.61), Animals (.66), Jobs (.59), ATTA Fluency, Originality, Elaboration, Flexibility (.44, .49, .63, and .20,

PAGE 120

120 respectively), CRCVerb (.46), CRCFig (.74), WCH Flu (.36), WCHOri (.30), RAT (.45), Insight items 1 5 (.34, .20, .19, .15, .10, respectively), p < .001. The loadings of WM, PS, and IQ on g were .94, .47, and .57, p < .001, respectively. Intelligence (g) significantly predicted AF and ( = .70, p < .001). AF significantly predicted DT and CT ( = .56, .83, p < .001). The residuals between Pattern1 and Pattern2, ATTAFlu and ATTAFlex, ATTAFlu and WCHF, and Insight2 and Insight 3 were correlated at r = .50, .47, .30, .28, and .19, p < .001, respectively. Overall, results from the models us ing item level indicators showed the same patterns of relationships among processing speed, working memory, intelligence, associative processing, divergent thinking, and convergent thinking as the models using parcel indicators. However, in this case, the model with parcel indicators (model 4) had a problem with identification whereas the model with item level indicators (mo del 5) did not. Table 74 provides a summary of the goodness of fit indices for all of the models discussed above.

PAGE 121

121 Figure 7 10. Model 5 : Relationships among higher order g associative fluency, divergent thinking, and convergent thinking (item level indicators)

PAGE 122

122 Table 7 4. Summary of chi squ a re (2) test degrees of freedom, and goodness of fit indices of final models Model 2 df p RMSEA CFI TLI SRMR Original Models (parcel indicators) 1a CFA WM, PS. and IQ 13.04 5 .02 .08 .93 .79 .04 1a SEM WM, PS. and IQ (WM PS, WM IQ) 13.03 5 .02 .08 .93 .79 .04 1b CFA AF, DT, and CT 6.04 11 .87 < .001 1.0 1.0 .02 1b SEM AF, DT, and CT (AF DT, AF CT) 6.04 11 .87 < .001 1.0 1.0 .02 1 CFA WM, PS, IQ, AF, DT, and CT 44.38 49 .66 < .001 1.0 1.0 .04 1 SEM W M, PS, IQ, AF, DT, and CT (IQ AF, IQ CT, AF DT) 55.60 56 .49 < .001 1.0 1.0 .04 2 SEM WM PS, IQ, AF, DT, and CT (IQ AF, AF DT, AF CT) 92.41 56 .002 .005 .92 .88 .05 Original Models (item indicators) 3 a CFA WM, PS. and IQF 143.84 130 .19 .02 .98 .97 .05 3 a SEM WM, PS. and IQF (WM PS, WM IQ) 143.84 130 .19 .02 .98 .97 .05 3 b CFA AF, DT, and CT 221.43 113 < .001 .06 .90 .88 .06 3 b SEM AF, DT, and CT (AF DT, AF CT) 222.39 115 < .001 .06 .90 .89 .06 3 SEM WM, PS, IQF, AF, DT, and CT (IQF AF, AF DT, AF CT) 783.15 547 < .001 .04 .83 .81 .07 Intelligence ( g ) as a higher order factor (parcel indicators) 4 CFA WM, PS, IQ, g AF, DT, and CT 54.88 55 .48 < .001 1.0 1.0 .04 4 SEM WM, PS, IQ, g AF, DT, and CT ( g AF, AF DT, AF CT) 96.90 57 < .001 .05 .91 .87 .05 Intelligence as a higher order factor (item indicators) 5 CFA WM, PS, IQ, g, AF, DT, and CT 819.48 580 < .001 .04 .84 .82 .07 5 SEM WM, PS, IQ, g, AF, DT, and CT ( g AF, AF DT, AF CT) 856.53 582 < .001 .04 .82 .80 .07

PAGE 123

123 The Role of Entity versus Incremental Beliefs of Creativity in Associative Fluency, Divergent Thinking, and Convergent Thinking The self reported ratings on the Theory of Creativity Scale (TCS) were scored according to the procedures proposed by Dweck and colleagues (1995). Participants who received an average score of 3 or lower across the 16 items on the TCS were assigned to the entity beliefs group. Participants who received an average score of 4 or higher across the 16 items were assigned to the increm ental beliefs group. Finally, the participants who received an average score between 3 and 4 were excluded from the analysis. Results of this group assignment procedure showed that only 6 of the 265 participants fell in the entity group (2.26%), 168 partic ipants fell in the incremental group (63.40%), and 91 participants were excluded from the analyses (34.34%). The proportions of part icipants in the three groups were significantly different from the 15% exclusion and 85% inclusion approximations proposed i n past studies of incremental beliefs (Dweck et al., 1995). However, it has been reported in the literature that implicit theories scales that have more items tend to result in larger numbers of participants reporting incremental beliefs, possibly due to t he desirable nature of the incremental items, social desirability bias, or boredom due to the similarity of the wording across the items (Dweck, 1999; Dweck et al., 1995). Because the entity group based on the 16 item TCS scale was not large enough to con duct group comparison tests, new average scores on the TCS were calculated based on the first three items on the TCS scale that were adapted from Dwecks (1999) original three item implicit theories of intelligence scale.

PAGE 124

124 Table 7 5 Descriptive statistics of entity and incremental groups on creativity tasks Theory of Creativity Creativity Indicator Entity Group M SD Incremental Group M SD LetterF 22.36, 5.48 23.02, 5.44 Animals 25.77, 4.80 27.28, 5.57 Jobs 20.44, 4.58 20.84, 4.13 ATTA Fluency 15.76, 2.62 15.93, 1.98 ATTA Originality 15.02, 2.42 15.27, 2.35 ATTA Elaboration 15.59, 2.43 16.02, 1.95 ATTA Flexibility 15.20, 2.35 15.44, 2.04 CRC Verbal 1.85, 1.09 2.15, 1.24 CRC Figural 5.60, 2.66 6.09, 2.38 CRC Total 7.45, 3.22 8.25, 3.07 ATTA C reativity I ndex 4.00, 1.35 4.11, 1.25 WCH Fluency 7.74, 3.63 9.20, 4.07 WCH Originality 2.63, .73 2.63, .70 RAT 15.85, 5.65 17.14, 4.19 Insight 1 .20, .41 .43, .50 Insight 2 .59, .49 .64, .48 Insight 3 .18, .39 .19, .40 Insight 4 .10, .31 .10, .31 Insight 5 .38, .05 .39, .49 Insight Total 1.45, 1.12 1.76, 1.00 Participants were assigned to an entity or incremental group, or excluded from the analysis, based on the same criteria used for the average score on the 16 item TCS (i.e., average score of 3 or lower, average score of 4 or higher, and average score between 3 and 4 were the benchmarks to be assigned to the entity, incremental, and excluded groups, respectively). Based on the average score on the three item TCS scale, 88 parti cipants (33.21%) fell into the entity group ( M = 2.38, SD = .55), 135 participants (50.94%) fell into the incremental group ( M = 5.03, SD = .72), and 42 participants (15.85%) were excluded from the analyses ( M = 3.44, SD = .28). In order to meet the assump tion of homogeneity of variance, 88 participants were randomly

PAGE 125

125 selected from the incremental group for equal group sizes. A series of t tests were conducted to compare the differences between students holding entity beliefs and students holding incremental beliefs of creativity on the battery of creative thinking measures used in this study. Table 75 summarizes the descriptive statistics of the entity and incremental groups performance on the creative thinking tests. Contrary to expectations, results showed that the entity beliefs group did not differ from the incremental beliefs group on any of the associative fluency tasks; LetterF, t (174) = .80, p = .42, Animals, t (172) = 1.91, p = .06, and Jobs, t (174) = .61, p = .55. There were also no significant differences between the two groups on the divergent thinking indicators and total scores; ATTA Fluency, t (174) = .49, p = .63, ATTA Originality, t (174) = .70, p = .49, ATTA Elaboration, t (174) = 1.30, p = .20, ATTA Flexibility, t (174) = .72, p = .47, Criterion Referenced (CRC) Verbal, t (174) = 1.68, p =.09, CRC Figural, t (174) = .129, p = .20, CRC Total, t (174) = 1.68, p = .10, ATTA Creativity Index (CI), t (174) = .58, p = .56, and wire coat hanger (WCH) Originality, t (173) = .07, p = .95. The incremental group outperformed the entity group on the number of ideas generated on the Unusual Uses task, wire coat hanger (WCH) Fluency, t (173) = 2.52, p = .01. Finally, the incremental group outperformed the entity group on Insight problem 1 (four dot problem), t (174) = .3.32, p = .001. Differences on the remainder of the convergent thinking indicators and total scores were nonsignificant; RAT, t (172) = 1.71, p = .09 Insight items 2 to 5, t (174) = .62, p = .54, t (174) = .19, p = .85, t ( 174) = .00, p =1.00, t (174) = .15, p = .88, respectively, and Insight Total, t (174) = 1.91, p = .06.

PAGE 126

126 Table 76 summarizes the results from the group comparisons between the entity group and the incremental group on various indicators and total scores of associative fluency, divergent thinking, and convergent thinking tasks. Table 7 6 Results of t tests between the entity and incremental beliefs groups Creativity Indicator t value p value LetterF t (174) = .80, p = .42 Animals t (172) = 1.91, p = .06 Jobs t (174) = .61, p = .55 ATTA Fluency t (174) = .49, p = .63 ATTA Originality t (174) = .70, p = .49 ATTA Elaboration t (174) = 1.30, p = .20 ATTA Flexibility t (174) = .72, p = .47 CRC Verbal t (174) = 1.68, p =.09 CRC Figural t (174) = .129, p = .20 CRC Total t (174) = 1.68, p = .10 ATTA Creativity Index t (174) = .58, p = .56 WCH Fluency t (173) = 2.52, p = .01 WCH Originality t (173) = .07, p = .95 RAT t (172) = 1.71, p = .09 Insight 1 t (174) = .3.32, p = .001 Insight 2 t (174) = .62, p = .54 Insight 3 t (174) = .19, p = .85 Insight 4 t (174) = .00, p =1.00 Insight 5 t (174) = .15, p = .88 Insight Total t (174) = 1.91, p = .06

PAGE 127

127 CHAPTER 8 DISCUSSION Working M emory P redicts P rocessing S peed and F luid I ntelligence Since Baddeley and Hitchs (1974) conception of the working memory model, several researchers have shown evidence that in contrast to short term memory and processing speed, working memory is a system comprising multiple components, including storage, proc essing, and executive systems (Baddeley, 2002; Conway et al., 2002; Engle et al., 1999; Unsworth et al., 1999). Therefore, short term memory and processing speed are conceptualized as single processing systems, whereas working memory is a more complex, mul tidimensional construct that engages simultaneous processing, storage, and executive capacities. It has been argued that it is the executive functions aspect of working memory that accounts for the strong relationship between working memory and intelligenc e (Conway et al., 2002, 2003 ; Engle et al., 1999). The findings in this study are in line with previous studies that have shown the predictive power of multifaceted complex working memory span tasks for fluid intelligence (Cowan et al., 2003; Kane et al., 2004; Unsworth et al., 2009) as well results that showed that when the role working memory and short term memory in intelligence are taken into account, the role of processing speed in predicting fluid intelligence is negligible (Conway et al., 2002). Therefore, the strong relationship between working memory and fluid intelligence found in this study is supported on both theoretical (e.g., Baddeley 1992; Baddeley & Andrade, 2000 ; Baddeley & Logie, 1999; Ericsson & Kintsch, 1995) and empirical (e.g., Conway et al., 2002; Engle et al., 1999) grounds. Altogether, these results coincide with the large body of literature exploring the

PAGE 128

128 r elationships among working memory, processing speed, and intelligence that consistently show working memory to be a significant predictor of fluid intelligence. In this study, it was hypothesized that both working memory and processing speed would predict intelligence. Based on the mixed findings regarding the relative contributions of working memory and processing speed to intelligence (e.g., Colom et al., 2005; Conway et al, 2002; Jensen, 1998; Kail & Salthouse, 1994 ) processing speed was expected to ac count for some of the variance in intelligence. However, results showed that the relationship between processing speed and intelligence was small and statistically insignificant. Similar to previous findings (e.g., Conway et al., 2002; Engle et al., 1999; Kyllonen & Christal, 1990) the contribution of processing speed on intelligence disappeared when working memory was accounted for. This is somewhat surprising, because processing speed was only one of two predictors of intelligence. Furthermore, previous studies that explored the combined roles of working memory, short term memory, and/or processing speed in intelligence showed evidence that mental speed contributed to intelligence (e.g., Colom et al., 2005; Coyle et al., 2011; Fry & Kale, 1996; Kaufman et al., 2009), although these findings have been mixed (e.g., Conway et al., 2002). Therefore, although working memory was expected to predict intelligence more strongly than processing speed, the mental speed factor was still expected to account for some o f the variance in intelligence. However, r esults in this study did not show evidence for the role of processing speed in intelligence. In fact, results in this study showed that working memory predicts processing speed, suggesting that the abilities associ ated with working memory (e.g., controlled attention, strategy deployment, management of interference) support speed of

PAGE 129

129 processing. These results are in line with Unsworth et al.s (2009) study that identified processing speed as one of the three component s underlying complex working memory span tasks. However, in contrast to Unsworth et al.s (2009) finding that processing speed accounted for some of the variance in fluid intelligence, results from this study show that processing speed does not predict fluid intelligence, and is therefore in this regard more consistent with Conway et al.s (2002) suggestion that working memory is an essential aspect of fluid intelligence, and processing speed tasks that place minimal demands on memory and attention (such as the simple mental speed measures used in the present study) do not significantly predict fluid intelligence. It is possible that the variance accounted for by processing speed in Unsworth's (2009) study was accounted for by the speed component in the SymS pan task. An additional explanation for the discrepancy between results in this study and findings reported in the past studies is that the bulk of the processing speed literature examines the relationships among processing speed, working memory, and intel ligence within the context of developmental cognitive research. For example, a recent study showed that processing speed mediated the relationship between age and intelligence, concluding that increases in intelligence can be attributed to increases in mental speed (Coyle et al., 2011). A probable explanation for absence of a relationship between processing speed and intelligence here compared to previous studies is that this study was conducted among a sample of young adults, among which developments in co gnitive abilities including mental speed and intelligence have become relatively stable. The findings from this study suggest that processing speed may be an elementary cognitive process underlying working memory capacity. The working

PAGE 130

130 memory tasks used in this study involved storing a series of numerical digits or a series of patterns of blocks in memory, while simultaneously engaging in a secondary processing task. It is likely that the complex set of storing, processing, and inhibiting mechanisms that support optimal use of a persons working memory capacity will predict other lower order cognitive abilities such as processing speed. Correlations between measures of working memory and processing speed showed that they were weak but significantly related. A n area for future research aimed to clarify the relationships among processing speed, working memory, and intelligence, is to conduct a longitudinal developmental study in which the relationships among processing speed, working memory capacity, and intelli gence are examined at different points in a persons lifespan. The existing body of literature seems to indicate that processing speed plays a much more influential role in cognitive abilities including executive functions and intelligence during childhood and adolescent years. However, as indicated by studies primarily conducted among college aged students, results among individuals in adulthood show evidence that working memory is the prominently predictive cognitive mechanism of fluid intelligence. Futur e research in this area should also take into account short term memory, and how the role of this memory system may change over a persons developmental span in relation to the other cognitive abilities and processes. Associative F luency P redicts D ivergent and C onvergent T hinking The predominant psychometric approach to studying creativity is dominated by studies of divergent thinking ( Dietrich, 2004; Runco, 1991, 1999). This may be surprising, considering that the psychology of creativity literature consists of numerous theoretical propositions and rich philosophical perspectives attempting to explain the

PAGE 131

131 broad range of the cognitive processes that underlie creativity. From Gestalt philosophers holistic perspective of creativity as a process of identifying gaps and restructuring the field and Wallas (1926) concept of incubation and illumination, to psychometric approaches proposed by Guilford (1950) including his concept of divergent thinking and Mednicks (1962) theory of the associative basis of creativity, and finally to the emerging field of creative cognition (e.g., Finke et al., 1992; Ward, 1995), there are fruitful opportunities to pursue diverse lines of empirical research in the psychological research of creative cognition. From a review of the earlier and more modern literature, associative fluency, divergent thinking, and convergent thinking were identified as three major components of creative thought. Although the contributions of these three cogni tive processes to creativity have been acknowledged in the creativity literature (albeit with different degrees of emphasis), the relationships among these three distinct creative thinking processes have not been empirically explored to date. Therefore, on e aim of the present study was to contribute to contemporary creativity research that is taking concepts of cognitive science to increase our understanding of creative thinking. In this study, the relationships among associative fluency, divergent thinking and convergent thinking were investigated using SEM Recently, a study conducted by Benedek et al. (2012) showed that associative abilities, including associative fluency, flexibility, dissociation, and combination explained nearly half of the variance of divergent thinking ability. In addition, a think aloud study showed that memory searches in which associations are fluency retrieved precedes more complex processes that lead to novel ideas on divergent thinking tests

PAGE 132

132 (Gilhooly et al., 2007). This study makes a novel contribution by investigating the role of associative processes, specifically associative fluency, on divergent as well as convergent thinking processes; it is the first known study to empirically investigate the role that associative fluenc y has in divergent thinking and convergent thinking together. Results showed that scores on associative fluency tasks were significantly correlated with various indicators of divergent thinking. For example, scores on the animal naming category fluency task correlated with elaboration, flexibility, creative index, and verbal and figural criterion referenced scores on the ATTA at r = .20, .26, .13 ,.17, and .22, p < .01, respectively. This was also the case regarding the relationships among performance on t he animal naming category fluency task and scores on the RAT and Insight tasks ( r = .19 and .21, p < .01, respectively). Results from the SEM analyses showed that associative fluency significantly predicted two complex and distinct creative processes, namely divergent and convergent thinking. The significant relationships found between associative fluency and both divergent and convergent thinking is noteworthy particularly because, divergent thinking and convergent thinking emerged as distinct types of cr eative thinking. Findings showed that several indicators of divergent thinking were weakly and insignificantly correlated with scores on convergent thinking tasks. For example, the creativity index on the ATTA had a correlation of r = .09, ns with the sco re on the RAT. Similarly, fluency, originality, elaboration, and flexibility scores on the ATTA were correlated with the score on the RAT at r = .03, .12, .08, and .17, ns respectively. Therefore, the relationships among associative fluency, divergent t hinking, and convergent thinking suggest that associative fluency is a broader, elementary cognitive process that is involved in both the

PAGE 133

133 generation of novel and unusual ideas (divergent thinking) as well as the combination of distally related ideas to identify a correct solution (convergent thinking). These findings lend evidence to Mednicks (1962) theory of the associative basis of creativity, which proposed that associative abilities, including the ability to activate and fluently retrieve associations, are an important feature of creative thinking. Associative processing may be a broader cognitive mechanism that two seemingly opposite cognitive processes in creative thinking may share. As suggested by Benedek et al. (2012), the associative processes may play a role in scanning for both unrelated asso ciative elements during divergent thinking, as well as related associative elements during convergent thinking. Overall, results from this study showed that associative fluency was a strong predictor of divergent and convergent thinking abilities (two dist inct types of creative thinking), and points to associative processes as an important underlying component of creative thought. In this study, the fluency component of associative processing was investigated; however, recent studies show that several disti nct associative abilities contribute to creative thinking (Benedek et al., 2012). Therefore, future research is needed to further explore how specific associative processes (e.g., fluency, flexibility, disassociation, and combination, Benedek et al., 2012) contribute both divergent and convergent creative thinking processes. Parsing broad cognitive processes involved in creativity into specific, observable sub processes is an important step towards gaining a more nuanced understanding of creative thinking. The findings from this study have important implications for theoretical frameworks of creative cognition. In the literature, divergent thinking is often described

PAGE 134

134 as an inductive process related to idea generation and convergent thinking is commonly characterized as a deductive process related to searching for a single, correct solution (Brophy, 2000; Guilford, 1957). Results show empirical evidence that divergent thinking and convergent thinking consist of distinct and negatively related cognitive mechanisms. As discussed further below, conclusions drawn from creativity studies have largely ignored these differences. With the exception of recent studies, many researchers have made general conclusions about creative ability from both divergent and convergent tests of creativity. In this study, the lack of relationship among several indicators of divergent thinking and scores on convergent thinking tests provide evidence that performance on divergent and convergent thinking should be interpreted differently. In addition to exploring other types of associative processes on divergent and convergent thinking, a next step for future research is to explore how these distinct creative thinking processes contribute to real life creative behaviors and achievements. Most of the work examining the predictive validity of creativity tests has focused on the relationships between divergent thinking performance and self report surveys of creative behaviors and accomplishments (Hocevar, 1979; Runco, 1991; Silvia et al., 200 8). In contrast, although convergent thinking has been argued to be an important component of creative behaviors (Brophy, 2000; Mumford et al., 1991; Treffinger et al., 1993), the relationship between convergent thinking and real life creativity has not be en empirically investigated. Models of creative cognition (e.g., Finke et al.s 1992 Geneplore Model) and creative problem solving (e.g., Isaksen et al., 2000; Treffinger, 1993; Wallas, 1926) propose that creativity is facilitated by a divergent thinking phase when ideas are freely generated at the beginning of the creative process, followed by a

PAGE 135

135 convergent thinking phase in which ideas are carefully evaluated, chosen, and developed. Anecdotal evidence from case studies provides some evidence for these mode ls of creative thinking; however, empirical studies of creative thinking are needed to clarify and operationalize the stages proposed. Because real life creative behaviors are difficult to study in the traditional confines of an experiment, where constraints such as time and other resources are limited, alternative methodologies that allow creativity to be studied in the interdisciplinary contexts in which creativity is occurring using various methodologies is likely to be a more appropriate approach to gar nering a more in depth and nuanced understanding of the creative process. Intelligence Predicts Associative Fluency, and Indirectly Predicts Divergent and Convergent Thinking One of the aims of this study was to provide a more nuanced approach to studyi ng the relationship between intelligence and creative thinking processes. To this end, we first modeled the relationships between three types of creative thinking processes (i.e., associative fluency, divergent thinking, and convergent thinking), followed by an examination of how intelligence predicts these creative processes. Based on Mednicks (1962) theory of individual differences in associative processing, and work by Mendelsohn (1976) that suggests that attentional control predicts associative abiliti es, we proposed that intelligence, specified by scores on both a fluid and crystallized intelligence measure in the structural equation model predicts associative fluency. In addition, intelligence was specified to indirectly predict divergent and convergent thinking through associative fluency. Altogether, our study empirically examined the longstanding view that associative processes underlie creative thinking,

PAGE 136

136 and also explored the role of intelligence in three creative processes (i.e., associative fluency, divergent thinking, and convergent thinking). We found that intelligence directly predicts associative fluency, and indirectly predicts divergent and convergent thinking. Intelligence showed to be a strong predictor of associative fluency ( = .85), which was then a particularly strong predictor of convergent thinking ( = .92) (but also predicted divergent thinking as well). These finding s are in line with recent work by Benedek et al. (2012)s recent findings that show compelling evidence for the overarching role of associative processes in divergent think ing processes and intelligence tests The patterns of relationships found among inte lligence, associative fluency, divergent and convergent thinking may be expected given the significantly positive correlations between performance on the RAT (a widely used convergent test of creativity) and measures of intelligence (i.e., RAPM, WAISRV) as well as working memory tasks (i.e., SymSpan, backward digit span). Our results support earlier findings (e.g., Jacobson et al., 1968, Taft & Rossiter, 1966) that showed that the RAT correlates more highly with traditional tests of convergent thinking comp ared to measures of divergent thinking. Intelligence also indirectly predicted divergent thinking through associative fluency in this study, but the paths in these relationships were not a strong. It is possible that because intelligence, executive, and co nvergent creativity tests rely more heavily on analytic cognitive processes that support the ability to identify a correct solution in complex tasks, tests such as the RAT and insight problems are more closely related to intelligence, traditionally defined compar ed to divergent thinking tests.

PAGE 137

137 Important to note is that the relationships among associative fluency, intelligence, and divergent thinking in this study share some similarities but also differences from the model proposed by Benedek et al. (2012). In both studies, the role of associative fluency as a relevant elementary cognitive process in more complex processes of creative ideation (i.e., divergent thinking) was examined. In both the structural equation model of this study as well as the model in Bendeks et al.s (2012) study, associative fluency was specified to predict divergent thinking. However, in this study, intelligence was specified to predict associative fluency, whereas in Benedek et al.s study (2012), four facets of associative proces sing, including associative fluency, were specified to predict intelligence. The literature shows that intelligence and executive functions are related to associative retrieval and learning processes (Carroll, 1993; Mendelsohn, 1976; Kaufman et al., 2011); however, the direction of the relationship between associative processing and intelligence is not clear. In this study, the theoretical justification for specifying intelligence as a predictor of associative fluency was driven by the conceptualization of the associative fluency construct as a broad spreading activation process that underlies divergent and convergent thinking The theoretical basis for specifying intelligence as a predictor of associative f luency is further discussed below Letter and semantic category fluency tasks were originally develop to assess cognitive decline (e.g., Benton & Hansher, 1978; Borkowski, Benton, & Spreen, 1967). These tasks were developed as tests of frontal and temporal lobe functioning, where cognitive processes such as sustained attention, retrieval of information, and verbal fluency are housed (Benton & Hansher, 1978; Borkowski, Benton, & Spreen, 1967).

PAGE 138

138 However, semantic category and letter fluency tasks have also been used as to assess individuals semantic memory (C ollins & Quillian, 1972), based on a memory model that is organized according to networks of ordered categories that can be accessed by spreading activation (access to one semantic category leads to the activation of related, lower order category) (Lund & Burgess, 1996; Roelofs, 1982). Category and letter fluency tasks have also been used in studies of creative cognition as indicators of fluid intelligence (e.g., Silvia, 2008a, b) and as associative fluency (e.g., Benedek et al., 2012) on the similar theoretical basis of spreading activation. Generally, spreading activation is conceptualized as a basic, core cognitive process underlying all other cognitive processes. However, a recent study suggested that these fundamental processes are predicted differentially by working memory capacity, a factor that accounts for individual differences in cognitive abilities (Schelble, Therriault, & Miller, 2012). This lends evidence for the specification of intelligence, a construct closely related to working memory, as a predictor of associative fluency. Drawing from the theoretical propositions of Mednicks (1962) association theory, as well as other creativity researchers who proposed associative fluency as a core process underlying creative thinking, the construct associative processing specified by semantic category and letter fluency tasks was specified as a creative thinking construct in lieu of a subcomponent or predictor of intelligence. However, it is also possible, that associative processing predicts intelligence. Recent studies showed that tasks related to associative processing, such as learning words associative with stimulus response pairs (e.g., Kaufman et al., 2009), predicted intelligence. In addition, Silvia (2008) used

PAGE 139

139 the same category and letter fluency tasks that were in the present study as indicators of a verbal fluency latent variable, which was specified to predict intelligence. Therefore, an alternative structural equation model in which associative fluency was specified to predict intelligence (rather than intelligence predicting associative fluency) was also tested. The GOF indices of this model showed that the model was a good fit to the data; however, there was a problem with identification, indicated by a standardized path coefficient greater than 1 from intelligence to convergent thinking. Furthermore, in this model, associative processing predicted intelligence and divergent thinking, but not convergent thinking. From a theoretical standpoint, these patterns of relationships among intelligence and the three creative thinking processes are more difficult to support, because we would expect that associative fluency, the ability to fluency retrieve words (letter fluency task) and responses from a wide range of semantic categories (category fluency task), would predict performance on the convergent thinking tasks which included the Remote Associates T est a test that involves making associations between words and has shown to be highly related to verbal intelligence (Taft & Rossiter, 1966), as well as the insight problems that included word scrambles. For these reasons, the convergent thinking tasks in this study are clearly engaging associative processes and we should expect associative fluency to predict convergent thinking Altogether based o n both statistical and conceptual grounds, the r esults from this study show stronger support for intelligence as a predictor of associative fluency rather than the associative fluency as a predictor of intelligence. I would also like to note that in this study, a Ravens matrices task and a standardized vocabulary test was used to assess fluid and crystallized intelligence,

PAGE 140

140 respectively. However, recent study using the largest battery of cognitive tests to date has shown that the general intelligence factor is best represented by tasks that tap into a verbal, perceptual, and mental rotation domain of cognitive ability (Johnson & Bouchard, 2005). Future research using more differentiated assessments of associative processing tasks, such as associative flexibi lity, combination, and disassociative tasks, as well as a wider range of intelligence tasks is needed in order to provide a more detailed understanding of the relationships among components of intelligence and associative processing. The Role of Working M emory in Associative Fluency, Divergent Thinking, and Convergent Thinking In addition to exploring the role of intelligence in three different types of creative thinking, this study also looked at working memory as another cognitive factor that may contribute to the cognitive structures of creative thought. Several researchers have suggested that people high in working memory are likely to have an advantage on divergent thinking tests due to their ability t o resist cognitive interference ( Gilhooly et al., 2007; Kane & Engle, 2003; Nusbaum & Silvia, 2011; Silvia, 2008a, b ; Sub et al., 2002). Research exploring possible executive functions in creativity has focused almost entirely on performance on divergent thinking tasks, showing compelling evidence t hat creative ideation requires overcoming interference caused by common answers and prior responses (Gilhooly et al., 2007; Kane, Bleckley, Conway & Engle, 2001; Silvia, 2008). In this study, we directly examined the role of one executive function, working memory, on creative thinking and found empirical evidence that working memory plays a role in not only divergent thinking, but associative fluency and convergent thinking as

PAGE 141

141 well. Possible explanations for the role of working memory in these three creativ e processes are further elaborated upon below. In divergent thinking tasks, working memory is likely to provide an advantage in being able to generate and consider several different ideas while simultaneously selecting the most original and ignoring the more obvious responses. Similarly, on the convergent thinking tasks such as the RAT and the insight problems, working memory is likely to play an important role in peoples ability to break away from a mental set or an ineffective approach to the problem, an d consider alternative, typically less common pathways in order to identify the correct solution. In fact, Mendelsohn (1976) argued early on that high performance on the RAT is due to greater breadth of attention deployment with respect to external cues and the better ability to maintain several streams of cognitive activity simultaneously (p. 347), cognitive mechanisms that align closely with components of Baddeleys (1986) model of working memory developed a decade later. Finally, these set shifting advantages linked to greater working memory capacity are also likely to support performance on the letter and category fluency tasks, by allowing people to draw from a wider set of semantic and taxonomic categories. Altogether, the ability to switch between response categories has been shown to predict better performance on divergent thinking tests in previous studies (e.g., Benedek et al., 2012; Gilhooly et al., 2007; Nusbaum & Silvia, 2011). This study contributes to the existing findings, by showing evidence that working memory, an individual difference factor which is related to the ability to use strategies associated with switching sets and overcoming interference, is also likely to benefit performance on convergent thinking and associative processing t asks.

PAGE 142

142 The Role of Implicit Theories in Associative, Divergent, and Convergent Thinking P rocesses Decades of research conducted by Dweck and colleagues showed evidence that peoples approaches to learning were rooted in their implicit beliefs about their intelligence; whether or not they believed intelligence was fixed of changeable (Dweck, 1999; Dweck & Leggett, 1988; Heyman & Dweck, 1992; Hong et al., 1999). The findings from this study did not show the same effects from a survey that was adapted from D wecks Theory of Intelligence scale to assess students beliefs about creativity. Based on scores from a battery of associative fluency, divergent thinking, and convergent thinking tasks, results showed that there were no significant differences in perform ance between students who reported entity beliefs of creativity compared to those who reported incremental beliefs of creativity. The only exceptions were for the fluency score of the Unusual Uses task, and the correct solution on the first insight problem (four dot problem), for which the incremental group outperformed the entity group. Some limitations to this study should be mentioned. Firstly, this study was exploratory in nature, adopting an existing scale originally developed to assess beliefs of int elligence to tap into possible beliefs of creativity. Whether or not students hold dichotomous (entity vs. incremental) beliefs of creativity ability has not been explored to my knowledge, although the permeating belief that creativity is fixed or of genius quality that exist among students has been alluded to in the literature ( Dietrich, 2004; Simonton, 2000; Sternberg & Lubart, 1996). It is possible that students views on creativity may not be as polarized as their views on intelligence that is reported by Dwecks (1999) research. Furthermore, the average score from the original 16item scale used in this study resulted in almost all students categorized into the incremental

PAGE 143

143 group. This was a vastly greater proportion than the division of beliefs reported in studies of theories of intelligence (Heyman & Dweck, 1992). As suggested by Dweck, in order to overcome possible social desirability biases in the survey responses, the 3item scale was used which resulted in more evenly distributed groups. However, us ing a scale with only 3 items is subject to serious validity and reliability issues. A possible follow up to exploring implicit beliefs students hold about creativity is experimentally manipulate beliefs of creativity by priming students towards entity or incremental beliefs of creativity and examining the possible effects this may have on subsequent creative thinking performance. Studies in which a theory of intelligence was induced by showing students an article that provided a compelling argument for ei ther t he entity or incremental belief showed evidence that students beliefs of intel ligence were subject to change (Bergen, 1991; Hong et al., 1999). Furthermore, studies showed that participants in the entity condition reported significantly lower interest in taking remedial tutorial following unsatisfactory feedback compared to participants in the incremental condition. There was no difference between the entity and incremental groups when given satisfactory feedback, providing further indication that holding an entity belief is linked to task avoidance in the face of failure (Dweck, 1999; Hong et al., 1991). The implications of the anticipated findings have important implications for educational practices. If results show that students who hold entity be liefs of creativity do in fact show lower creative thinking performance, it may be important to first target students beliefs about their creative ability in order for teaching techniques and educational curricula to aimed to train skills associated with original and openended thinking to have effect. The beliefs that students bring to the classroom regarding their

PAGE 144

144 creative ability may impede or diminish the effectiveness of their creative development, due to lack of effort, decreased motivation, performance oriented goals, and negative attributions made when faced with challenging tasks that wrestling with ambiguity. Currently, little is known about how students perceive their creative ability, as well as how this belief may potentially affect their creat ive performance. Implications and Future Directions Through surveys, case studies, and personality inventories, psychologist have amassed a large database of factors related to the construct o f creativity (Dietrich, 2010). H owever, rigorous empirical studi es of creative thinking processes remain largely underdeveloped. The psychometric study of creativity following the experimental work of Gestalt psychologists is often credited to Guilfords 1950 farewell address as the president of the American Psychologi cal Association, and the popularization of standardized divergent thinking tests marked the beginning of a more scientific approach to studying creativity, including efforts towards the operationalization of the creativity. However, as divergent thinking t ests gained popularity, they were used as tests of overall creative ability, leading to the misrepresentation of creativity as a monolithic entity. A similar practice was also adopted with Mednicks (1962) RAT test. Mednick (1962) developed the RAT based on his theory that individual differences exist in associative abilities. His associational theory of creativity drew on the notion that creative people have flatter associative hierarchies, enabling them to make more remote or uncommon associations. He suggested that a number of individual difference variables facilitate performance on the RAT, including the number of associations, the organization and strength of the links among the associations, and the capacity to select

PAGE 145

145 creative combinations among the associations. Despite the noticeably dissimilar conceptualization of creative thinking between Guilfords (1950) description of divergent thinking and Mednicks (1962) theory of associative processes, as well as the stark differences in the processes and g oal states of divergent thinking tests and the RAT test, the creativity literature has largely failed to acknowledge and distinguish the different aspects of creativity assessed by the two tasks. Furthermore, although early studies showed convincing evidence that the RAT is strongly related to measures of intelligence and executive functions (e.g., Mendelsohn, 1976) and called for a closer examination of the underlying executive cognitive mechanisms assessed by the RAT, this line of questioning was for the most part abandoned and the RAT has largely been adopted as a general measure of creativity. Moving away from the practice of using divergent thinking tests and RAT scores as proxy for creativity, modern creativity researchers are acknowledging and specifying the roles that various associative processes (Benedek et al., 2012), divergent thinking processes (Plucker, 1999; Silvia, 2008a, b), and convergent processes (Brophy, 2000) have in creative thinking activities such as metaphor generation (Silvia & Beaty, in press ), evaluating and refining of ideas (Brophy, 2000; Finke et al., 1992), category combination (Ward, 1995); mental imagery (Finke, 1992), inhibition of unoriginal ideas (Eysenck, 1995), implementation of strategies (Gilhooly et al., 2007; Nusbaum & Silvia, 2011), and more. Furthermore, the relationship among the many distinct creative processes and abilities associated with intelligence and executive functions is being reinvestigated.

PAGE 146

146 T he relationship between intelligence and creativity has been a topic of contentious debate in the psychological literature. Recently, c ontemporary creativity researchers are using both traditional methods from cognitive psychology (e.g., think alouds) and modern statistical methods (e.g., latent variable analysis) to re investigate the relationships among creative processes and intelligence. Recent studies show empirical evidence that divergent thinking is a higher order cognitive activity, and as such, a strong relationship between intelligence and creative thinking exists (Benedek et al., 2012; Gilhooly et al., 2007, Nusbaum & Silvia, 2011; Silvia, 2008). Much of this work focusing on the link between intelligence and creativity primarily examined how intelligence predicted divergent thinking. Our results co ntribute to these studies by exploring the role of intelligence, specified by scores from both a fluid and crystallized intelligence measure, in not only divergent thinking, but associative fluency and convergent thinking as well. Recasting creative think ing as a higher order cognitive process has important implications for studying creativity from an individual difference perspective. By identifying distinct cognitive processes in creative thinking and operationalizing these creative cognitive constructs, the psychometric study of creativity will be able to move beyond testing of divergent thinking and towards a more comprehensive study of the various processes that are involved in creativity. An executive interpretation of creative thinking also opens new lines of research to investigate how the acquisition and implementation of higher order cognitive processes in thinking, learning, and problem solving, may explain individual differences in creativity. Further exploring the conditions for fostering creati vity, which recent research indicates is settled in our individual

PAGE 147

147 differences (e.g., Dietrich, 2010; Gilhooly et al., 2007; Nusbaum & Silvia, 2011), is a fruitful field of contemporary research in cognition. In addition to further exploring the role of working memory in creative thinking processes, much remains to be studied regarding the relative contributions of different types of intelligence (e.g., crystallized versus fluid), domain knowledge, strategy use, metacognitive processes, and other executive functions on creativity. As Dietrich (2010) states in a recent review, It is hard to believe that creative behavior, in all its manifestations, from carrying out exquisitely choreographed dance moves, to scientific discovery, constructing poems, and coming up with ingenious ideas of what to do with a brick, engages a common set of brain areas or depends on a limited set of mental processes (p. 845). By looking into individual cognitive processes using methods and tools of cognitive science, creativity res earch can begin to shed new light on the currently fragmented literature filled with contradictory discourse. The field of cognitive research including studies of expertise, problem solving, and memory provide strong theoretical frameworks and rigorous mod els of interdisciplinary research that exemplify an integration of pluralistic perspectives and methodological approaches to better understand complex phenomena. Efforts such as these are needed to promote scientific advancements aimed at understanding the phenomenon of creativity in its entirety.

PAGE 148

148 APPENDIX A DIVERGENT THI N KING TEST RUBRIC Instructions for Judging Responses on Divergent Thinking Test (Silvia et al., 2008) Creativity can be viewed as having three facets. Creative responses will generally be high on all three, although being low on one of them does not disqualify a response from getting a high rating. We will use a 1 ( not at all creative) to 5 ( highly creative) scale. 1. Uncommon Creative ideas are uncommon; they will occur infrequently in our sample. Any response that is given by a lot of people is common, by definition. Unique responses will tend to be creative responses, although a response given only once need not be judged as creative. For example, a random or inappropri ate response would be uncommon but not creative. 2. Remote Creative ideas are remotely linked to everyday objects and ideas. Responses that stray from obvious ideas will tend to be creative, whereas responses close to obvious ideas will tend to be uncreati ve. 3. Clever Creative ideas are often clever: they strike people as insightful, ironic, humorous, fitting, or smart. Responses that are clever will tend to be creative responses. Keep in mind that cleverness can compensate for the other facets. For exampl e, a common use cleverly expressed could receive a high score.

PAGE 149

149 APPENDIX B ABBREVIATED TORRANCE TEST FOR ADULTS (ATTA) (GOFF & TORRANCE, 2002) ACTIVITY 1 JUST SUPPOSE you could walk on air or fly without being in an airplane or similar vehicle. What proble ms might this create? List as many as you can. You have 3 minutes to respond to this activity. ACTIVITY 2 Use the incomplete figures below to make some pictures. Try to make your pictures unusu al Your pictures should communicate as interesting and as com plete a story as possible. Be sure to give each picture a title No credit will be given for this activity unless the two incomplete figures are used.You have 3 minutes to respond to this activity. ACTIVITY 3 See how many objects or pictures you can make from the triangles below, just as you did with the incomplete figures. Remember to create titles for your pictures. No credit will be given for this activity unless the triangle figures are used. You have 3 min utes to respond to this activity.

PAGE 150

150 APPENDIX C ITEMS ON THE REMOTE ASSOCIATES TEST (BOWDEN & JUNG BEEMAN, 2003) Practice set Cue word 1 Cue word 2 Cue word 3 Answer light wheel cross boot birthday hand rain summer stick shopping tie ground candle cart bow camp Test set Cue word 1 Cue word 2 Cue word 3 Answer cottage cream loser show duck night dew fountain preserve aid flake cracker safety cane dream fish political sense worm piece river print opera fur hound food basket nuclear main carpet brick skate throat life fold wrist comb baking ranger rubber mobile fly cushion daddy break mine surprise courtesy shelf mind note berry hand rack pressure forward eight feud sweeper alert cake water spot row dollar stop bee pop tropical wagon cone fighter point plum light rush line place end dating account bird dish tail shot break snow album light ink cheese ice sore boat bill watch honey soda forest band snow fire pin sugar day gold party common book game bank blue soap coat blood fast ball family street red

PAGE 151

151 APPENDIX D INSIGHT PROBLEMS 1. 4 dots: Without lifting your pencil from the paper, show how you could join all 4 dots with 2 straight lines 2. Figure: Show how you can divide this figure into four equal parts that are the same size and shape ( included in pilot study and later removed). 3. Word: Rearrange the following letters to make a familiar word: flymia 4. Word: Rearrange the following letters to make a familiar word: lendraca 5. Letter Z: Can you figure out where to put the letter Z, top or bottom line and Why? A EF HI KLMN T V WXY -------------------------BCD G J OPQRS U 6. Series: Identify the next term in the series: 88 ... 64 ... 24 ... ___ Solutions: 1. 2. 3. family 4. C alendar 5. Top, all of the letters on the top have straight lines, and all of the letters on the bottom line have curved lines 6. 40 (88 64 = 24, 64 24 = 40)

PAGE 152

152 APPENDIX E THEORY OF CREATIVITY SCALE Using the scale below, please indicate the extent to which you agree or disagree with each of the following statements by writing the number that corresponds to your opinion in the space next to each statement. (* = Reverse scored item) Strongly Disagree 1 2 3 4 5 6 Strongly Agree 1. _____ You have a certain amount of creativity and you cannot do much to change it.* 2. _____ Your creativity is something about you that you cant change very much.* 3. _____ You can learn new things, but you cant really change your basic creativity.* 4. _____ Difficulties and challenges prevent you from developing your creativity.* 5. _____ The effort you exert improves your creativity. 6. _____ If you fail in a task, you question your creativity.* 7. _____ Criticism from others can help develop your creativity. 8. _____ You can develop your creativity if you really try. 9. _____ Good performance in a task is a way of showing others that you are creative.* 10. _____ When you exert a lot of effort, you show that you are not creative.* 11. _____ When you learn new things, your basic creativity improves. 12. _____ If you fail in a task, you still trust your creativity. 13. _____ Performing a task successfully can help develop your creativity. 14. _____ Your abilities are determined by how creative you are.* 15. _____ Good preparation before performing a task is a way to develop your creativity. 16. _____ You are born with a fixed amount of creativity. Items 1 3 were adapted from Dweck (1999) and items 416 were adapted from AbdEl Fattah & Yates (2006).

PAGE 153

153 LIST OF REFERENCES Abd El Fattah, S. M., & Yates, G. C. R. (2006). Implicit Theory of Intelligence Scale: Testing for factorial invariance and mean structure. Paper presented at the Australian Association for Research in Education Conference, Adelaide, South Australia. Ackerman, P. L., Beier, M. E., & Boyle, M. O. (2002). Individual differences in working memory within a nomological network of cognitive and perceptual speed abilities. Jou rnal of Experimental Psychology, General 131(4), 567 589. Ackerman, P. L., Beier, M. E., & Boyle, M. O. (2005). Working memory and intelligence: The same or different constructs. Psychological Bulletin, 131, 306 0. Albert, R. S. (1975). Toward a behavioral definition of genius. American Psychologist, 30, 140151. Althuizen, N., Wierenga, B., & Rossiter, J. (2010). The validity of two brief measures of creative ability. Creativity Research Journal, 22(1), 53 61. Amabile, T. M. (1996). Creativity in context: Update to the social psychology of creativity Boulder, CO: Westview. Anastasi, A. (1982). Psychological testing (5th ed.). New York: Macmillan. Anderson, J. R. (1993). Problem solving and learning. American Ps ychologist, 48(1), 3544. Aronson, J., Fried, C. B., & Good, C. (2002). Reducing stereotype threat and boosting academic achievement of AfricanAmerican students: The role of conceptions of intelligence. Journal of Experimental Social Psychology, 38 113 125. Au, A. (2001). The impact of learned resourcefulness on university adjustment and student retention among international, Canadian minority and Canadian nonminority students. Unpublished thesis dissertation, Trent University, Peterborough, ON. Baddeley, A. D. (1986). Working memory Oxford: Clarendon Press. Baddeley, A. D. (1992). W orking memory Science, New Series, 255 (5044), 556559. Baddeley, A. D. (2000). The episodic buffer: a new component of working memory? Trends in Cognitive Sciences, 4(11), 417423. Baddeley, A. D., & Andrade, J. (2000). Working memory and the vividness of imagery. Journal of Experimental Psychology General 129 126 145.

PAGE 154

154 Baddeley, A. D., & Logie, R. H. (1999). Working memory: The multicomponent model. In A. Miyake & P. Shah (Eds.), Models of working memory: Mechanisms of active maintenance and executive control (pp. 28 61). New York: Cambridge University Press. Baddeley, A. D., & Wilson, B. A. (2002). Prose recall and amnesia: Implications for the structure of working memory Neuropsychologia, 40, 17371743. Barnes, G. M. (1978). Creativity: A complementary relationship information demand and success at problem solving. Intelligence, 2, 169 180. Barron, F. (1963). Creativity and psychological health. Princeton, KJ: Van Nostrand. Barron, F. (1988). Putting creativity to work. In R. J. Sternberg (Ed.), The nature of creativity: Contemporary psychological perspectives (pp.7698). New York, NY: Cambridge University Press. Barron, F., & Harrington, D. M. ( 1981). Creativity, intelligence, and personality. Annual Review of Psychology, 32, 439 476. Batey, M., & Furnham, A. (2006). Creativity, intelligence, and personality: A critical review of the scattered literature. Genetic, Social, and General Psychology M onographs, 132, 355429. Batey, M., Furnham, A., & Safiullina, X. (2010). Intelligence, general knowledge and personality as predictors of creativity. Learning and Individual Differences, 20, 532535. Baughman, W. A., & Mumford, M. D. (1995). Process analy tic models of creative capacities: Operations influencing the combinationand reorganization process. Creativity Research Journal, 8(1), 3762. Becker, M. (1995). Nineteenthcentury foundations of creativity research. Creativity Research Journal, 8, 219 2 29. Beghetto, R. A., & Kaufman, J. C. (2007). Toward a broader conception of creativity: A case for mini c creativity. Psychology of Aesthetics, Creativity, and the Arts, 1 (2), 73 79. Benedek, M., Konen, T., & Neubauer, A. C. (2012). Associative abilities underlying creativity. Psychology of Aesthetics, Creativity, and the Arts, 6 (3), 273 281. Benton, A. L. (1994). Neuropsychological assessment. Annual Review of Psychology, 45, 1 23. Benton, A. L., & Hamsher, K. (1978). Multilingual aphasia examination manual Iowa City, IA: University of Iowa.

PAGE 155

155 Bergen, G. K. (1991). Beliefs about intelligence and achievement related behaviors Unpublished doctoral dissertation, University of Illinois at UrbanaChampaign. Binet, A., & Simon, T. (1916). The development of intelligence in children. Balitmore, MD: Williams & Wilkens (E. S. Kite, Trans.). Blackwell, L. S., Trezsniewski, K. H., & Dweck, C. S. (2007). Implicit theories of intelligence predict achievement across an adolescent transition: A longitudinal study and an intervention. Child Development, 78(1) 246 263. Borkowski, J., Benton, A., & Spreen, O. (1967). Word fluency and brain damage. Neuropsychologia, 5, 135 140. Bors, D. A., & Stokes, T. L. (1998). Ravens Advanced Progressive Matrices: Norm s for First Year University Students and the Development of a Short Form. Educational and Psychological Measurement, 58(3), 382398. Bowers, E. M., & Jung Beeman, M. (2003). Normative data for 144 compound remote associates problems. Behavior Research Met h ods, Instruments, & Computers, 35(4), 634 639. Bowers, K. S., Farvolden, P., & Mermigis, L. (1995). Intuitive antecedents of insight. In S. Smith, T. B. Ward, & R. A. Finke (Eds.), The creative cognition approach (pp. 2752). Cambridge, MA. : MIT Press. B rophy, D. R. (1998). Understanding, measuring, and enhancing individual creative problem solving efforts. Creativity Research Journal, 11(2), 123 150. Brophy, D. R. (2000). Comparing the attributes, activities, and performance of divergent, convergent, and combination thinkers. Creativity Research Journal, 13(3&4), 439455. Cacciari, C., Levorato, M. C., & Cicogna, P. (1997). Imagination at work: Conceptual and linguistic creativity in children. In T. B. Ward, S. M. Smith, & J. Vaid (Eds.), Creative thought: An investigation of conceptual structures and processes (pp.145177). Washington, DC: American Psychological Association. Cantor, J., Engle, R. W., & Hamilton, G. (1991). Shortterm memory, working memory, and verbal abilities: How do they relate? Intel ligence, 15, 229246. Carroll, J. B. (1993). Human cognitive abilities: A survey of factor analytic studies. Cambridge,UK: Cambridge University Press. Cattell, R. (1949). Culture Free Intelligence Test Scale 1 Handbook, Champaign, I L : Institute of Personality and Ability. Cattell, R. B. (1971). Abilities: Their structure, growth, and action. Boston: Houghton Mifflin.

PAGE 156

156 Consequences form A 1 Beverly Hills, CA: Sheridan Supply. Collins, A M., & Quillian, M. R. (1972). How to make a language user. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 309351). New York: Academic Press. Colom, R., Rebollo, I., Palacios, A., JuanEspinosa, M., & Kyllonen, P. (2004). Working memory is (almost) perfectly predicted by g. Intelligence, 32, 277 296. Colom, R., Abad, F. J., Rebollo, I., & Shih, P. C. (2005). Memory span and general intelligence: A latent variable approach. Intelligence, 33 623 642. Colom, R., & Shih, P. (2004). Is work ing memory fraction ated onto different components of intelligence? Intelligence, 32 ,431 444. Conway, A. R. A., Cowan, N., & Bunting, M. F. (2001). The cocktail party phenomenon revisited: The importance of working memory capacity. Psychonomic Bulletin & Re view, 8 331335. Conway, A. R. A., Cowan, N., Bunting, M. F., Therriault, D. J., & Minkoff, S. R. B. (2002). A latent variable analysis of working memory capacity, short term memory capacity, processing speed, and general fluid intelligence. Intelligence, 30, 163183. Conway, A. R. A., Kane, M. J., & Engle, R. W. (2003). Working memory capacity and its relation to general intelligence. Trends in Cognitive Sciences, 7(12), 547552. Conway, A. R. A., Jarrold, C., Kane, M. J., Miyake, A., & Towse, J. N. (2007 ). Variation in working memory New York, NY: Oxford University Press. Coyle, T. R., Pillow, D. R., Snyder, P. A. C., & Kochunov, P. (2011). Processing speed mediates the development of general intelligence (g) in adolescence. Psychological Science, 22, 1 265 1269. Cox, C. M. (1926). The early mental traits of three hundred geniuses Stanford, CA: Stanford University Press. Cropley, A. J. (2000). Defining and measuring creativi ty: Are creativity tests worth using? Roeper Review, 23(2), 72 80. Craig, M., & Manis, M. (1962). Prediction of scores of on the Remote Associates Test by size of response repertoire. Unpublished manuscript, 1960, cited in S. A. Mednick, The associative basis of the creative process. Psychological Review, 69, 220232.

PAGE 157

157 Csikszentmihalyi, M. (1999). Implications of a systems perspective for the study of creativity. In R. J. Sternberg (Ed.), Creativity research handbook (pp.313 335). Cambridge, England: Cambridge University Press. Csikszentmihalyi, M., & Getzels, J. W. (1971). Discovery or iented behavior and the originality of creative products: a study with artists. Journal of Personality and Social Psychology, 19, 47 52. Csikszentmihalyi, M., & Getzels, J. W. (1973). The personality of young artists: An empirical and theoretical explorat ion. British Journal of Psychology, 64, 91 104. Cury, F., Elliot, A. J., Da Fonseca, D., & Moler, A. C. (2006). The so cial cognitive model of achievement motivation and the 2 x 2 achievement goal framework. Journal of Personality and Social Psychology, 90, 666679. Cushen, P. J., & Wiley, J. (2012). Cues to solution, restructuring patterns, and reports of insight in creative problem solving. Consciousness and Cognition, 23 (3), 11661175. Daneman, M. (1991). Working memory as a predictor of verbal fluency. Journal of Psycholinguistic Research, 20 (6), 445 464. Deary, I.J. (2001). Intelligence: A very short introduction. Oxford: Oxford University Press DeYoung, C. G., Flanders, J. L., & Peterson, J. B. (2008). Cognitive abilities involved in insight problem solving: An individual differences model. Creativity Research Journal, 20(3), 278 290. Dietrich, A. (2004). The cognitive neuroscience of creativity. Psychonomic Bulletin & Review, 11(6), 10111026. Dietrich, A., & Kanso, R. (2010). A review of EEF, ERP, and neuroimaging studies of creativity and insight. Psychological Bulletin, 136(5), 822848. Dominowski, R. L. (1995). Productive problem solving. In S. M. Smith, T. B. Ward and R. A. Finke. (Eds.), The creative cognition approach (pp. 7396). Cambridge, MA: MIT Press. Dorner, D., & Kreuzig, H. W. (1983). Problem solving and intelligence. Psychologie Rundschau, 34, 185 192. Dow, G. T., & Mayer, R. E. (2004). Teaching students to solve i nsight problems: Evidence for Domain Specificity in Creative Thinking. Creativity Research Journal, 16(4), 389 402.

PAGE 158

158 Dunker, K. (1926). A qualitative (experimental and theoretical) study of productive thinking (solving of comprehensible problems). Pedagogic al Seminary and Journal of Genetic Psychology, 33 642 708. Duncker, K. (1945). On problem solving. Psychological Monographs, 58(5), Whole No. 270. Dweck, C. S. (1999). Self theories: Their role in motivation, personality, and development Philade l phia, PA : Psychological Press. Dweck, C. S., Chiu, C. Y., & Hong, Y.Y. (1995). Implicit theori es and their role in judgments and reactions: A world from two perspectives. Psychological Inquiry, 6 (4), 267 285. Dweck, C. S., & Leggett, E. L. (1988). A social cognitive approach to motivation and personality. Psychological Review, 95(2), 256 273. Ellis, H. (1904). A study of British genius London: Hurst & Blackett. Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A R. A. (1999). Working memory, short term memory, and general fluid intelligence: A latent variable approach. Journal of Experimental Psychology: General, 125, 309 331. Ericsson, K. A., & Simon, H. A. (1998). How to study thinking in everyday life: Contrasting think aloud protocols with descriptions and explanations of thinking. Mind, Culture, and Activity, 5(3), 178 186. Eysenck, H. J. (1995). Creativity as a product of intelligence and personality. In D. H. Saklofske & M. Zeidner (Eds.), International handbook of personality and intelligence (pp.231 247), New York, London: Plenum P ress. Finke, R. A. (1996). Imagery, creativity, and emergent structure. Consciousness & Cognition, 5, 381 393. Finke, R. A., Ward, T. B., & Smith, S. M. (1992). Creative cognition: Theory, research, and applic ations Cambridge, MA: MIT Press. Fry, A. F., & Hale, S. (1996). Processing speed, working memory, and fluid intelligence: Evidence for a developmental cascade. Psychological Science, 7 237 241. Galton, F. (1962). Hereditary genius: An inquiry into its laws and consequences. Cleveland, OH: World. (Original work published 1882). Getzels (Eds.), Perspectives in creativity (pp. 3759). Chicago: Aldine. Getzels, J. W., & Csikszentmihalyi, M. (1975). From problem solving to problem finding. In I. A. Taylor & J W. Getzels (Eds.), Perspectives in creativity. (pp. 90116). Chicago: Aldine Publishing Company.

PAGE 159

159 Getzels, J. W., & Jackson, P. W. (1962). Creativity and intelligence: Explorations with gifted students New York: Wiley. Gianotti, L. R. R., Mohr, C., Pizza galli, D., Leh mann, D., & Brugger, P. (2001). Associative processing and paranormal belief. Psychiatry and Clinical Neuroscience, 55, 595 603. Gick, M. L., & Holyoak, K. J. (1980). Analogical problem solving. Cognitive Psychology, 12, 3063055. Gick, M. L., & Holyoak, K. J. (1983). Schema induction and analogical transfer. Cogniitve Psychology, 15, 1 38. Gilhooly, K. J., Fioratou, E., Anthony, S. H., & Wynn, V. (2007). Divergent thinking: Strategies and executive involvement in generating novel uses for f amiliar objects. British Journal of Psychology, 98, 611 625. Gianotti, L. R. R., Mohr, C., Pizzagalli, D., Lehmann, D., & Brugger, P. (2001). Associative processing and paranormal belief. Psychiatry and Clinical Neuroscience, 55, 595 603. Good, C., Aronson J., & Inzlicht, M. (2003). Improving adolescents standardized test performance: An intervention to reduce the effects of stereotype threat. Journal of Applied Developmental Psychology, 24, 645 662. Greenberg, E. M. (1966). Construct validation of the Re mote Associ ates Test (RAT) using mediated pairedassociate (PA) learning and IQ. Masters thesis, The Ohio State University. Guilford, J. P. (1950). Creativity. American Psychologist, 5, 444 454. Guilford J. P. (1957). Creative abilities in the arts. Ps ychological Review, 64, 110 118. Guilford J. P. (1966). Intelligence: 1965 model. American Psychologist, 21, 20 26. Guilford J. P. (1967). The nature of human intelligence. New York: McGraw Hill. Guilford J. P. (1970). Creativity: Retrospect and prospec t. Journal of Creative Behavior, 4 149 168. Guilford J. P. (1975). Creativity: A quarter century of pr ogress In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in creativity Chicago: Aldine Publishing Company. Guilford, J. P., Merrifield, P. R., & Wilson, R. C. (1958). Unusual uses test Orange, CA: Sheridan Psychological Services.

PAGE 160

160 Heyman, G.D., & Dweck, C.S. (1992). Achievement goals and intrinsic motivation: Their relation and their role in adaptive motivation. Motivation and Emotion, 16, 231 247. Holyoak, K. J., & Koh, K. (1987). Surface and structural similarity in analogical transfer. Memory and Cognition, 15, 332 340. Hong, E., & Milgram, R. M. (1991). Original Thinking in Preschool Children: A validation of ideational fluency measures. Creativ ity Research Journal, 4, 253260. Hong, E., Milgram, R. M., & Gorsky, H. (1995). Original thinking as a predictor of creative performance in young children. Roeper Review, 18, 147 149. Hong, E., Milgram, R. M., & Whiston, S.C. (1993). Leisure activities i n adolescents as a predictor of occupational choice in young adults. Journal of Career Development, 19, 221229. Hong, Y., Chiu, C., Lin, D. M., Wan, W., & Dweck, C. S. (1999). Implicit theories, attributions, and coping: A meaning system approach. Journal of Personality and Social Psychology, 77, 588599. Horn, J. L., & Cattell, R. B. (1966). Refinement and test of the theory of fluid and crystallized intelligence. Journal of Educational Psychology, 57, 253270. Hox, J. J., & Bechger,T. M. (1998). An intr oduction to structural equation modeling. Family Science Review,11, 354373. Hu, L., & Bentler, P. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1 55. Hunter, S. T., Bedell Avers, K. E., Hunsicker, C. M., Mumford, M. D., & Ligon, G. S. (2008). Applying multiple knowledge structures in creative thought: Effects on idea generation and problem solving. Creativity Research Journal, 20(2), 137 154. Ilk owska, M., & Engle, R.W. (2010). Working memory capacity and self regulation. In: R. Hoyle (Ed.), The Handbook Personality and Self regulation. Malden, MA: Wiley Blackwell. Isaksen, S. G., Dorval, K. B., & Treffinger, D. J. (2000). Creative approaches to problem solving Dubuque, IA: Kendall/Hunt. Jackson, D., Gillaspy, J., & PurcStephenson, R. (2009). Reporting practices in confirmatory factor analysis: An overview and some recommendations. Psychological Methods, 14, 6 23.

PAGE 161

161 Jacobson, L. L., Elenewski, J. J., Lordahl, D. S., & Liroff, J. H. (1968). Role of creativity and intelligence in conceptualization. Journal of Personality and Social Psychology, 10 431436. Jensen, A. R. (1998). The g factor: The science of mental ability Westport, CT: Praeger. Jense n, A. R. (2006). Clocking the mind: Mental chronometry and individual differences St. Louis, MO: Elsevier Science. Johnson, W. & Bouchard, T. J., Jr. (2005). The structure of hum an intelligence: It is verbal, perceptual, and image rotation (VPR), not flui d and crystallized. Intelligence, 33, 393 416. Kail, R. & Salthouse, T. A. (1994). Processing speed as a mental capacity. Acta Pscyhologica, 86, 199 225. Kane, M. J., Bleckley, M. K., Conway, A. R., & Engle, R. W. (2001). A controlled attention view of wo rking memory capacity. Journal of Experimental Psychology: General, 130, 169183. Kane, M. J., Hambrick, D. Z., Tuholski, S. W., Wilhelm, O. Payne, T. W., & Engle, R. W. (2004). The generality of working memory capacity: A latent variable approach to verbal and visuo spatial memory span and reasoning. Journal of Experimental Psychology General, 133 189 217. Kaufman, J. C. (2009). Creativity 101. New York: NY: Springer. Kaufman, S. B., DeYoung, C. G., Gray, J. R., Brown, J., & Mackintosh, N. (2009). Associative learning predicts intelligence above and beyond working memory and processing speed. Intelligence, 37 374 382. Kaufman, J. C., Kaufman, S. B., & Lichtenberger, E. O. (2011). Finding creative potential on intelligence tes ts via divergent production. Canadian Journal of School Psychology, 26 (2), 83 106. Kharkhurin, A. V., & Motalleebi, S. N. (2008). The impact of culture on the creative potential of American, Russian, and Iranian College Students. Creativity Research Journal, 20 (4), 404411. Kim, K. H. (2005). Can only intelligent people be creative? A metaanalysis. Journal of Secondary Gifted Education, 16 57 66. Kim, K. H. (2006). Can we trust creativity tests? A review of the Torrance Tests of Creative Thinking (TTCT). Creativity Research Journal, 18(1), 3 14. Kim, K. H. (2008). Meta analyses of the relationship of creative achievement to both IQ and divergent thinking test scores. Journal of Creative Behavior, 42(2), 106130.

PAGE 162

162 Kline, R. B. (2005). Principles and practice of structural equation modeling. N ew Y ork : Guilford Press Kohler, W. (1927). The mentality of apes New York: Harcourt, Brace. Kyllonen, P. C., & Christal, R. E. (1990). Reasoning ability is (little more than) working memory capacity?! Intelligence, 14 389 433. Kwandwalla, P. N. (2011). An exploratory investigation of divergent thinking through protocol analysis. Creativity Research Journal, 6(3), 241 259. Laughlin, P. R., Doherty, M. A., & Dunn, R. P. (1968). Intentional and incidental concept formatio n as a function of motivation, creativity, intelligence, and sex. Journal of Personality and Social Psychology, 8 401 409. Levy, S. R., Stroessner, S., & Dweck, C. S. ( 1998). Stereotype formation and endorsement: The role of implicit theories. Journal of Personality and Social Psychology, 74 14211436. Lincoln, Y. S. & Guba, E. G. (1985). Naturalistic Inquiry Newbury Park, CA: Sage Publications. Little, T. D., Cunningham, W. A., Sharhar, G., & Widaman, K. F. (2002). To parcel or not to parcel: Exploring the question, weighting the merits. Structural Equation Modeling, 9(2), 151173. Little, T. D., Lindenberger, U., & Nesselroade, J. R. (1999). On selecting indicators for multivariate measurement and modeling with latent variables: When good indicators are bad and bad indicators are good. Psychological Methods, 4, 192 211. MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psychological Methods, 4, 84 99. MacKinnon, D. W. (1961). Creativity in architects. In Proceedings of the Conference on The Creative Person. Berkeley: University of California, Extension Division, 1961. MacKinnon, D. W. (1962). The nature and nurture of creative talent. American Psychologist, 17 484495. MacKinnon, D. W. (1975). IPARs contribution to the conceptualization and study of creativity. In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in creativity (pp. 6089). Chicago: Aldine Publishing Company. MacKinnon, D. W. (1978). In search of human effectiveness Buffalo, NY: Creative Education Foundation.

PAGE 163

163 MacKinnon, D. W., & Hall, W. B. (1971). Inte lligence and creativity. In The Measurement of Creativity pp. 118388. Int. Congr. Appl. Pscyhol., 17th, Liege, Belgium, H. W. Peter, symp; ch,Brussels: EDITEST Mackw orth, N. H. (1964). Originality Harvard University Press. Martin, A., Wiggs, C. L., Lalonde, F., & Mack, C. (1994). Word retrieval to letter semantic cues: A double dissociation in normal subjects using interference tasks. Neuropsychologia, 28, 823829. M artindale, C. (1995). Creativity and connectionism. In S. M. Smith, T. B. Ward, & R. A., Finke (Eds.), The creative cognition approach (pp. 249268). Martindale, C. (1999). Biological basis for creativity. In R. J. Sternberg (Ed.), Handbook of creativity (pp. 137152). New York: Cambridge University Press. Mednick, S. A. (1962). The associative basis of the creative process. Psychological Review, 69(3), 220 232. Mednick, M. T. (1963). Research creativity in psychology graduate students. Journal of Consulti ng Psychology, 27(3), 265 266. Mednick, S. A., & Mednick, M. T. (1967). Examiners manual: Remote Associates Test Boston: Houghton Mifflin. Mendelsohn, G. A. (1976). Associative and attentional processes in creative performance. Journal of Personality, 44, 341 369. Milgram, R. M. (1983). A validation of ideational fluency measures of original thinking in children. Journal of Educational Psychology, 75, 619624. Milgram, R. M. (1988). Tel Aviv Activities Inventory: Primary Grades Tel Aviv University, School of Education, Ramat Aviv, Israel. Milgram, R. M. (1990). Tel Aviv Activities Inventory: Adolescent Form Tel Aviv University, School of Education, Ramat Aviv, Israel. Milgram, R. M, & Arad, R. (1981). Ideational fluency as a predictor of original p roblem solving. Journal of Educational Psychology, 73, 568 572. Milgram, R. M., Milgram, A., Rosenbloom, G., & Rabkin, L. (1978). Quantity and quality of creative thinking in children and adolescents. Child Development, 49, 385388. Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81 97. Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity of diversity of executive functions and their contributions to

PAGE 164

164 complex frontal lobe tasks: A latent variable analysis. Cognitive Psychology, 41 49100. Mohr, C., Graves, R. E. Gianotti, L. R. R., Pizzagalli, D., & Brugger, P. (2001). Loose but normal: A semantic association study. Journal of Psycholinguistic Research, 30, 475483. Moser, J. S., Schroder, H. S., Heeter, C., Moran, T. P., & Lee, Y H. (2011). Mind your errors: Evidence for a neural mechanism linking growth mindset to adaptive post error adjustments. Psychol ogical Science, 22, 14841489. Mumford, M. D., Mobley, M. I., Reiter Palmon, R., Uhlman, C. E., & Doares, L. M. (1991). Process analytic models of creative capacities. Creativity Research Journal, 4, 91 122. Mumford, M. D., Supinski, E. P., Threlfall, K. V ., & Baughman, W. A. (1996). Process based measures of creative problem solving skills: III. Category selection. Creativity Research Journal, 9 395 406. Newell, A., Shaw, J. C., & Simon, H. A. (1962). The process of creative thinking. In H. E. Gruber, G. Terrell, & M. Wertheimer (Eds.), Contemporary approaches to creative thinking (pp. 63119). New York: Atherton. Nielsen, B. D., Pickett, C. L., & Simonton, D. K. (2008). Conceptual versus experimental creativity: Which works best on convergent and diverg ent thinking tasks? Psychology of Aesthetics, Creativity, and the Arts, 2(3) 131138. Nusbaum, E. C., & Silvia, P. J. (2011). Are intelligence and creativity really so different? Fluid intelligence, executive processes, and strategy use in divergent think ing. Intelligence, 39, 3645. Osborn, A. F. (1953). Applied Imagination: Principles and Procedures of Creative Thinking. Charles Scribners Sons, New York. Phillips, L. H. (1997). Do frontal tests measure execut ive function? Issues of assessment and evidence from fluency tests. In P. M. A. Rabbitt (Ed.), Methodology of frontal and executive function (pp. 191213). Hove, England: Psychology Press. Plaks, J. E., Stroessner, S. J., Dweck, C. S., & Sherman, J. W. (2001). Person theories and attention allocation: Preferences for stereotypic versus counter stereotypic information. Journal of Personality and Social Psychology, 80, 876893. Plucker, J. A. (1999). Is the proof in the pudding? Reanalyses of Torrances (1958 to present) longitudinal data. Creativ ity Research Journal, 12, 103114.

PAGE 165

165 Plucker, J. A., Beghetto, R. A., Dow, G. T. (2004). Why isnt creativity more important to educational psychologists? Potentials, pitfalls, and future directions in creativity research. Educational Psychologist, 39(2), 8396. Plucker, J. A., & Renzulli, J. S. (1999). Psychometric approaches to the study of human creativity. In R. J. Sternberg (Ed.), Handbook of creativity (pp. 3561). New York: Cambridge University Press. Plucker, J., & Runco, M. (1998). The death of creativity measurement has been greatly exaggerated: Current issues, recent advances, and future directions in creativity assessment. Roeper Review, 21, 36 39. Rastogi, D. & Sharma, N. K. (2010). Creativity under concurrent and sequential tas k conditions. Creativity Research Journal, 22(2), 139 150. Redmond, M. R., Mumford, M. D., & Teach, R. (1993). Putting creativity to work: Effects of leader behavior on subordinate creativity. Organizational Behavior and Human Decision Processes, 55, 120 151. Rossman, E., & Fink, A. (2010). Do creative people use shorter associative pathways? Personality and Individual Differences, 49, 891 895. Rothenberg, A., & Greenberg, B. (1976). The index of scientific writings on creativity: General: 15661974 Hamden, CT: Archon. Runco, M. A. (1991). The evaluative, valuative, and divergent thinking of children. Journal of Creative Behavior, 25, 311319. Runco, M. A. (2004). Creativity. Annual Review of Psychology, 55, 657 687. Runco, M. A. (2003). Creativity, c ognition, and their educational implications. In J. C. Houtz (Ed.), The educational psychology of creativity (pp. 25 56). Cresskill, NJ: Hampton Press. Runco, M. A. (2007). Creativity: Theories and themes: Research, development, and practice. San Diego, CA : Academic Press. Runco, M. A. (2010). Divergent thinking, creativity, and ideation. In J. C. Kaufman & R. J. Sternberg (Eds.), Cambridge handbook of creativity (pp. 416446). New York : Cambridge University Press. Runco, M .A., & Chand, I. (1995). Cognition and creativity. Educational Psychology Review, 7 (3), 243 267. Runco, M. A., & Mraz, W. (1992). Scoring divergent thinking tests using total ideational output and a creativity index. Educational and Psychological Measurement, 52, 213221.

PAGE 166

166 Ruscio, J., Whit ney, D. M. & Amabile, T. M. (1998). Looking inside the fishbowl of creativity: Verbal and behavioral predictors of creative performance. Creativity Research Journal, 11(3), 243263. Ryan, R. M. (1982). Control and information in the intrapersonal sphere: An extension of cognitive evaluation theory. Journal of Personality and Social Psychology, 43 450461. Schelble, J. L., Therriault, D. J., & Miller, D. M. (2012). Class ifying retrieval strategies as a function of working memory. Memory and Cognition, 40 218 230. Sifonis, C. M. (1995). Scene schemas and creativi ty: Examining the influence of schema based knowledge on the creative process. Unpublished masters thesis. Texas A&M University, College Station, TX. Silvia, P. J. (2008a ). Creativity and intel ligence revisited: A latent variable analysis of Wallach and Kogan (1965). Creativity Research Journal, 20(1), 34 39. Silvia, P. J. ( 2008b). Another look at creativity and intelligence: Exploring higher order models and probable confounds. Personality and Individual Differences, 44, 10121021. Silvia, P. J., & Beaty, R. E. (in press ). Making creative metaphors: The importance of fluid intelligence for creative thought. Intelligence Silvia, P. J., Martin, C., & Nusbaum, E. C. (2009). A snapshot of creativi ty: Evaluating a quick and simple method for assessing divergent thinking. Thinking Skills and Creativity, 4 79 85. Silvia, P. J., Winterstein, B. P., Willse, J. T., Barona, C. M., Cram J. T., Hess, K. I., et al.,(2008). Assessing creativity with divergent thinking tasks: Exploring the reliability and validity of new subjective scoring methods. Psychology of Aesthetics, Creativity, and the Arts, 2 68 85. Simonton, D. K. (2000). Creativity: Cognitive, personal, developmental, and social aspects. American Psychologist, 55(1), 151 158. Sligh, A. C., Conners, F. A., & Roskos Ewoldsen, B. (2005). Relation of creativity to fluid and crystallized intelligence. Journal of Creative Behavior, 39, 123 136. Spearman, C. (1904). General intelligence, objectively det ermined and measured. American Journal of Psychology, 15 201 293. Spearman, C. (1930). Creative mind. London: Misbet. Spruill, J. (1984). Wechsler Adult Intelligence ScaleRevised. In D. J. Keyser & R. C. Sweetland (Eds.), Test critiques: Volume I (pp. 72 8 739). Austin, TX: PRO ED.

PAGE 167

167 Stein, B. S. (1989). Memory and Creativity. In R. A. Glover, R. R. Ronning, & C. R. C.R. Reynolds. Handbook of Creativity, Assessment, Research and Theory New York: Plenum Press. Sternberg, R. J. (1985). Beyond IQ: A triarchic theory of human intelligence. Cambridge University Press. Sternberg, R. J., & Dess, N. K. (2001). Creativity for the new millennium. American Psychologist, 5 444 454. Sternberg, R. J., & Lubart, T. I. (1996). Investing in creativity. American Psychologist 51(7), 677 688. Sternberg, R. J., & OHara, L. A. (1999). Creativ ity and intelligence. In R. J. Sternberg(Ed)., Handbook of creativity (pp. 251 272). New York: Cambridge University Press. Storm, B.C., Angello, G., & Bjork, E.L. (2011). Thinking can cause forgetting: Memory dynamics in creative problem solving. Journal of Experimental Psychology: Learning, Memory, and Cognition, 37, 12871293. Sub, H. M., Oberauer, K., Wittmann, W. W., Wilhelm, O., & Schulze, R. (2002). Working memory capacity explains reasoning ability and a little bit more. Intelligence, 30, 261288. Swiercinsky, D. (1988). WAIS R Tutorial Workbook Austinm TX: PRO ED. Tashakkori, A., & Teddlie, C. T. (Eds.). (1998). Handbook on mixed methods in the b ehavioral and social sciences. Thousand Oaks, CA: Sage. Taft, R., & Rossiter, J. (1966). The Remote Associates Test: Divergent or convergent thinking? Psychological Reports, 19, 13131314. Taylor, I. A. (1975). A retrospective view of creativity investigat ion. In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in creativity. (pp. 136). Chicago: Aldine Publishing Company. Torrance, E. P. (1968). A longitudinal examinati o n of the fourth grade slump in creativity. Gifted Child Quarterly, 12, 195199. Torran ce, E. P. (1969). Prediction of adult creative achievement among high school seniors. Gifted Child Quarterly, 16, 75 88. Torrance, E. P. (1972). Predictive validity of the Torrance Tests of Creative Thinking. Journal of Creative Behavior, 6 236 252. Torra nce, E. P. (1966). The Torrance Tests of C rea tive T hinking: Norms technical manual Lexington, MA: Personnel Press.

PAGE 168

168 Torrance, E. P. (1974). The Torrance Tests of Creative Thinking: Norms technical manual Princeton, NJ: Personnel Press. Torrance, E. P. (2008). The Torrance Tests of Creative Thinking: Norms technical manual Bensenville, IL: Scholastic Testing Services. Treffinger, D. J. (1995). Creative problem solving: Overview and educational implications. Educational Psychology Review, 7 (3), 301 312. Treffinger, D. J., Isaksen, S., & Dorval, B. (2000). Creative problem solving: An introduction (3rd ed.), Waco, Texas: Prufrock Press. Treffinger, D. J., Young, G., Shelby, E., & Shpardson, C. (2002). Assessing creativity: A guide for educators. Storrs, C onnecticut : The National Research Center on the Gifted and Talented. Unsworth, N., & Engle, R. W. (2005). Working memor y capacity and fluid abilities: Examining the correlat ion between operation span and R aven. Intelligence, 33, 6781. Unswo rth, N., Redick, T. S., Heitz, R. P., Broadway, J. M., & Engle, R. W. (2009). Complex working memory span tasks and higher order cognition: A latent variable analysis of the relationship between processing and storage. Memory, 17(6), 635 654. Vernon, P.E. (Ed.) (1970). Creativity: Selective readings Bungay, Suffolk: Richard Clay Ltd.Vincent, A. S., Decker, B. P., & Mumford, M. D. (2002). Divergent thinking, intelligence, and expertise: A test of alternative models. Creativity Research Journal, 14(2), 163 1 78. Vincent, A. S., Decker, B. P., & Mumford, M. D. (2002). Divergent thinking, intelligence, and expertise: A test of alternative models. Creativity Research Journal, 14(2), 163178. Wallach, M. A. (1971). The intelligence/creativity distinction New York: General Learning Press. Wallach, M. A., & Kogan, N. (1965). Modes of thinking in young children: A study of the creativity intelligence distinction. New York: Holt, Rinehart & Winston. Wallas, G. (1926). The art of thought N ew Y ork : Harcourt, B race, & World. Ward, T. B. (1994). Structured imagination: The role of category structure in exemplar generation. Cognitive Psychology, 27, 1 40. Ward, T. B. (1995). Whats old about new ideas? The creative cognition approach, (pp. 157178). Massachusetts Institute of Technology.

PAGE 169

169 Ward, T. B., Smith, S. M., & Vaid, J. (1997). Conceptual structures and processes in creative thought. In T. B. Ward, S. M. Smith, & J. Vaid (Eds.), 1997, Creative thought: An investigation of conceptual structures and processes (pp. 83 110). Washington, DC: American Psychological Association. Weiner, R. P. (2000). Creativity and Beyond. Cultures values, and change. New York: State University of New York Press. Weisberg, R. W. (1986). Creativity, genius, and other myths New York: Freeman. Weisberg, R. W. (1988). Problem solving and creativity. In Sternberg, R. J. (Ed.). The nature of creativity: Contemporary psychological perspectives. New York, NY: Cambridge University Press. Weisberg, R. W. (1995). Case studies of creative thinking. In Smith, S. M., Ward, T. B., & Finke, R. A. (Eds.) (1995). The creative cognition approach Cambridge, MA: MIT Press. Wertheimer, M. (1945 ). Productive thinking New York: Harper. Weschler, D. (1958). The measurement and appraisal of adult intel ligence, (4th ed.). Baltimore: Williams and Wilkins. Wechsler, D. (1974). Manual for the Wechsler Intelligence Scale for ChildrenRevised New York: Psychological Corporation. Wilde, N. J., Strauss, E., & Tulsky, D. S. (2004). Memory span on the Wechsler s cales. Journal of Clinical and Experimental Neuropsychological Society, 9 302. Wiley, J. (1998). Expertise as mental set: The effects of domain knowledge in creative problem solving. Memory & Cognition, 26(4), 716730. Williams, B. A., & Pearlberg, S. L. (2006). Learning of threeterm contingencies correlates with Raven scores, but not with measures of cognitive processing. Intelligence, 34, 177 191. Yahav, A. (1965). Associative processes, environmental stimuli, and the Remote Associates Test of creativit y. Doctoral dissertation, University of California, Berkeley.

PAGE 170

170 BIOGRAPHICAL SKETCH Christine Lee received her bachelors degree in Psychology at the University of Florida in 2007. Throughout her graduate career, she served as a research assistant on a NSF funded Engineering Education project and the lab manager of a Cognitive Laboratory Christine also taught undergraduate courses in Human Growth and Development and Child Development for Inclusive Education. After receiving a d octorate in Educational Psychology in 2012, she will be serving as a postdoctoral researcher at California State University East Bay, on a NSF Integrated Middle School Science Education project. Christine plans to pursue a career in academia, teaching and conduct ing research focused on creativity cognition, and learning in educational contexts