Title: What we believe and what we ask
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00103060/00001
 Material Information
Title: What we believe and what we ask message contradictions and news gathering
Physical Description: vii, 100 leaves : ; 29 cm.
Language: English
Creator: Horvath-Neimeyer, Paula S
Publication Date: 1992
 Subjects
Subjects / Keywords: Reporters and reporting   ( lcsh )
Journalism -- Objectivity   ( lcsh )
Stereotypes (Social psychology)   ( lcsh )
Genre: bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis (Ph. D.)--University of Florida, 1992.
Bibliography: Includes bibliographical references (leaves 90-99).
Statement of Responsibility: by Paula S. Horvath-Neimeyer.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00103060
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 001790459
notis - AJL4119
oclc - 29238709

Full Text











WHAT WE BELIEVE AND WHAT WE ASK:
MESSAGE CONTRADICTIONS AND NEWS GATHERING














By

PAULA S. HORVATH-NEIMEYER


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


1992















To

Greg and Tyler














ACKNOWLEDGMENTS

I owe a debt of gratitude to a great many people who

have coaxed, cajoled, and prodded me along the path toward

the competition of this dissertation. Without their help

and encouragement, my tenure as a doctoral student might

have been considerably longer and less fruitful.

First, I want to thank Dr. Michael Weigold for his

guidance in planning, conducting, and completing this work.

His humor, patience, and knowledge were indispensable during

this arduous process.

Next, I want to thank the remaining members of my

committee: Dr. Leonard Tipton, Dr. Kurt Kent, Dr. Jeff

Farrar, and Dr. James Algina. I am grateful for the support

and understanding each of these people displayed, as well as

their willingness to make themselves available during even

the most minute dissertation crisis.

Finally, I would like to thank my family. I am

grateful to my husband, Greg, for the patience, support, and

encouragement that he gave so willingly even during my

darkest doctoral days. And last, I want to thank my son,

Tyler, who was born during the start of this dissertation,

for providing the extra incentive I needed to complete my

doctoral work.


iii















TABLE OF CONTENTS



page

ACKNOWLEDGMENTS.................. ............... iii

ABSTRACT......................................... vi

CHAPTERS

1 INTRODUCTION.................................... 1

2 REVIEW OF THE LITERATURE.................... 11

Stereotyping of the Elderly.................. 12
Information-Seeking Biases................... 14
Stereotypes and Information Seeking......... 22
Factors That May Affect Biased Information
Seeking.................................. 25
Hypotheses.................................. 34

3 METHODS...................................... 38

Subjects..................................... 38
Measures and Procedures..................... 38
Manipulations .............................. 39
Dependent Measures............................ 42

4 RESULTS..................................... 46

Preliminary Analysis......................... 46
Primary Analyses............................. 49

5 DISCUSSION... ............................... 57

Limitations.................................. 59
Result Interpretation....................... 61
Further Research.............................. 66
Importance to the Discipline................ 68

APPENDICES

A STEREOTYPE-CONSISTENT MESSAGE............... 70

iv











B STEREOTYPE-INCONSISTENT MESSAGE............. 71

C NO ADDITIONAL INFORMATION CONDITION
INSTRUCTIONS ............................. 72

D "BE OBJECTIVE" CONDITION INSTRUCTIONS........ 73

E "CONSIDER THE OPPOSITE" CONDITION
INSTRUCTIONS................................ 74

F PRETEST..... ................................ 75

G EXPERIMENT QUESTIONNAIRE.................... 82

REFERENCES............................................. 90

BIOGRAPHICAL SKETCH......... ...................... 100














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


WHAT WE BELIEVE AND WHAT WE ASK:
MESSAGE CONTRADICTIONS AND NEWS GATHERING

By

Paula S. Horvath-Neimeyer

December 1992


Chairman: Dr. Michael Weigold
Major Department: Mass Communication


Research suggests that the appearance of news

distortion may be the result of cognitive and cultural

biases, such as stereotypes, that guide the news-selection

and interpretation process of reporters. This study

explored how reporters' stereotypes affected their

information seeking as well as factors that might influence

those stereotypes. Student reporters were asked to read a

scenario concerning a traffic accident that involved an

elderly driver that contained a cue that was either

consistent or inconsistent with an ageist stereotype. They

were then instructed to "be objective," to "consider the

opposite," or given no additional instructions. Two other

dependent variables measured the likelihood of various

causes of the accident and subjects' personal images of the








elderly. Results showed that subjects ranked

age-stereotypic questions (questions stereotypic of the

elderly) higher than they did nonstereotypic questions

(questions that pertained to a "youthful" stereotype)

regardless of the cue or instruction condition. Similar

results were exhibited in the remaining two dependent

measures, with support for a cue effect found only when

subjects rated the likelihood of accident causes. The

results suggest that cognitive biases such as stereotypes

may guide journalists' news-gathering strategies.


vii















CHAPTER 1

INTRODUCTION

Nearly a half a century ago, media magnate Henry R.

Luce suggested that a national inquiry be launched to

explore press freedom and responsibilities. The Commission

on the Freedom of the Press's specific mandate was to

examine the then-current state of the mass media and offer

suggestions for the future. It was a time when new

technology promised to extend the media's reach to even

greater audiences, and Commission members questioned whether

the media could be counted upon to behave responsibly in

wielding this mass power. In its findings in 1947, the

Commission expressed concern with the media's news-gathering

strategies and the resulting distortion. "The news is

twisted by the emphasis on firstness, on the novel and

sensational. . Too often the result is meaninglessness,

flatness, distortion, and the perpetuation of

misunderstanding" (p. 68).

Not much seems to have changed in the intervening

years. Today a majority of the public believes that the

press is biased (Gallup, cited in Stevenson & Greene, 1980).

Newspaper readers report that coverage is less than

impartial, with opinionated content from the editorial pages

1










often affecting news coverage (Bogart, 1984). For the

public, the media, which were once thought to offer their

audiences a window on the world, have become "clouded

mirrors" (Real, 1989, p. 261) that often distort and

disguise the reality they attempt to portray.

And the public is not the only group complaining that

news stories distort what they see as reality. Mass

communication scholars (Ettema & Glasser, 1988; Fishman,

1980; Real, 1989; Stocking & Gross, 1989a, 1989b) agree that

news stories are often, at best, a skewed rendering of

reality. In Tuchman's terms, news is a "selective reality"

with its own "internal validity" (1978, p. 97). In other

words, within the context of the story, the information

presented is accurate, perhaps in part because the reporter

chooses only the information that supports that view.

Stocking and Gross (1989b) suggest that the etiology of

news distortions lies within the process of news production.

At some point, as the reporter gathers information and

passes it to the editor and on to the public, disparate,

distorted constructions often emerge. It is the source of

the distortion that draws the attention of mass

communication researchers. "What reporters and editors do

with the many bits of information that come their way each

day is a matter of great interest" (Stocking & Gross, 1989b,

p. 85).










Early researchers exploring possible sources of news

distortion focused their attention on individual editors.

Gatekeeping studies (Buckalew, 1969; McCombs & Shaw, 1976;

Stempel, 1985; White 1950; Whitney & Becker, 1982) in

particular attempted to illuminate the role an editor plays

in the selection and placement of news items. More recent

studies have focused on environmental factors (Stocking &

Gross, 1989b) such as deadlines, newshole considerations,

institutional policies, the community, and culture as

contributing sources of news distortion.

Other researchers (Eason, 1988; Ettema & Glasser, 1988;

Fishman, 1980; Jensen, 1987) suggest that one of the major

sources of news distortion is the reporter. These studies

characterize the reporter as a powerful force in the

news-production process who actively directs the selection

and presentation of information that is imparted to the

audience in news stories. Both Fishman (1980) and Ettema

and Glasser (1988) outline how the behavior of reporters can

affect the information they collect and the news stories

they write. Differences in reporters' behaviors can result

in distorted versions of reality. Fishman concludes that

the reason is methodological:

Knowledge is not a passive record of
perceptions; it is the consequence of something
people do. News is the result of the methods
newsworkers employ. Were different methods used,
different forms of news would result and publics
would know the world outside their direct
experience in a very different way. (p. 14)










What Fishman suggests is that reporters shape the news

by employing various information-gathering strategies.

According to this notion of news gathering, the "reality"

portrayed in a news story is a reconstruction of events

based upon the questions reporters ask and the information

they select. In turn, the strategy a reporter selects

depends, at least in part, upon a host of cognitive biases

and errors that guide the reporter's information-collection

and information-selection strategies.

That such cognitive biases and errors may result in

many alternative constructions of reality has long been

accepted by researchers in social cognition and

communication (Applegate & Sypher, 1983; Higgins & Bargh,

1987; Nisbett, Borgida, Crandall, & Reed, 1976). One reason

for differences in reality construction is that people do

not generally weigh social information in the same ways

(Kahneman & Tversky, 1973); therefore they perceive and

think about the social world differently. These alternative

constructions may be understood as distorted renditions if

they differ either from socially accepted "reality" or the

constructions of others.

Stocking and Gross (1989b) suggest that the cognitive

paradigms developed by scholars in social cognition may

provide fruitful theoretical foundations for research on

reporters' contributions to news bias. "In fact, this

knowledge identifies a range of phenomena, extending well










beyond the much studied phenomenon of story selection, to

describe and explain. In so doing, it offers numerous

possibilities for new research on how journalists process

the news" (p. viii).

One manifestation of reporters' cognitive biases may be

the appearance of stereotypes in the mass media. Such

stereotypes have been differentially described by

researchers. Early researchers (Katz & Braly, 1933) defined

stereotyping as an unadaptive type of perception that relied

upon individual characteristics. Researchers wrote that

poor social and psychological adjustment was the root cause

of stereotyping (Perlmutter, 1956). More recent researchers

have defined stereotyping as a kind of categorizing behavior

(Cauthen, Robinson, & Krauss, 1971) that is often

operationalized by having people select from a group of

adjectives those that describe the group (Berninger &

DeSoto, 1985). Even more recently, Linville (1982)

described stereotypes by looking at how extremely people

rated others who were not part of their social or cultural

group.

Mass media studies of stereotyping have generally

defined stereotypes as broad generalizations about the most

common characteristics of a group. From racism to sexism,

mass-mediated portrayals of stereotypes have been a subject

of considerable interest among mass communication

researchers. Ethnic and racial minorities (Alexander, 1976;










Culley & Bennett, 1976; Greenberg & Kahn, 1970; Tan, Li, &

Simpson, 1986; Tan & Suarchavarat, 1988), women (Busby,

1975; Davis, 1982; Kalisch & Kalisch, 1984; Manstead &

McCulloch, 1981; VanSlyke Turk, 1987), men (Pollner, 1982),

the disabled (Bogdan, Biklen, Shapiro, & Spelkomman, 1982),

and the mentally ill (Gerbner & Tannenbaum, 1961; Nunnally,

1973; Steadman & Cocozza, 1977; Taylor, 1957; Wahl & Roth,

1982; Winnick, 1978) have all been subjected to stereotypic

portrayals within the media.

Another focus of mass media research on stereotypes has

been the way in which the media portray the elderly and the

effect those images have on media audiences. In general,

researchers report that senior citizens tend to be

underrepresented within the media and that the images that

do appear are stereotypic and negative (Broussard, Blackmon,

Blackwell, Smith, & Hunt, 1980; Gerbner, Gross, Signorielli,

& Morgan, 1980; Hess, 1974; Kvasnicka, Beymer, & Perloff,

1982; Northcott, 1975).

For example, Broussard, Blackmon, Blackwell, Smith, and

Hunt (1980) found that metropolitan daily newspapers largely

ignored the elderly. They surveyed story length, number of

items, and amount of space devoted to people over 60 and

found the amount of news mentioning the elderly was

disproportionate to the number of elderly readers. In

another study, Kvasnicka, Beymer, and Perloff (1982)

determined that advertisements depicting people over 60 in










magazines that did not target an elderly audience were

underrepresented. When people over 60 did appear in non-

specialized magazines, they were depicted in a less-than

favorable light, often for products dealing with physical

decline.

Neussel (1982) suggests that such depictions demean the

elderly either implicitly, through their exclusion or

subordination in the media, or explicitly, through the

distortion and degradation within the images portrayed

(Pingree & Hawkins, 1978; Stocking & Gross, 1989b).

While a sizeable body of research has explored the

presence of ageism in the mass media, less attention has

been paid to the genesis of those stereotypic images. A

single study (Horvath-Neimeyer & Kent, 1990) linked student

journalists' age-related stereotypes with a tendency to

collect stereotype-confirming information. Additional

research exploring the minds of journalists to ascertain how

cognitive factors may guide the gathering of information and

its presentation as news is lacking.

Stocking and Gross (1989a) lament this oversight and

propose that researchers use theories of cognition to

explore biases in reporting. "The need--to learn more about

such errors and biases, and to bring what we know and learn

into the educational setting--is great" (p. 10).

This study explores how cognitive factors, such as a

reporter's personal beliefs, may affect the collection of










information and thereby contribute to the presence of

stereotypes in news stories. More specifically, it examines

processes related to how reporters gather information

relevant to their personal beliefs about a stimulus--a

confirmatory bias. In this study, the stimulus is a

message concerning an elderly man involved in an automobile

accident.

Confirmatory bias turns on the notion that people

actively seek information by asking questions that elicit

information that confirms previously held beliefs or

hypotheses (Snyder & Campbell, 1980; Snyder & Cantor, 1979;

Snyder & Skrypnek, 1981; Snyder & Swann, 1978a; Snyder &

White, 1981). Such biased information gathering precludes

the collection of disconfirmatory information, resulting in

the kinds of cultural stereotypes that appear ineradicable

even in the face of contradictory evidence.

This study also examines whether a confirmatory bias

can be reversed or eliminated through either instructions

reporters receive or cues embedded within the message.

There is some literature that suggests the information-

gathering strategies people use can be affected by message

cues. Hewes, Doelger, and their colleagues (Doelger, Hewes,

& Graham, 1986; Hewes & Graham, 1989; Hewes, Graham,

Doelger, & Pavitt, 1985; Hewes, Graham, Monsour, & Doelger,

1989) have postulated that messages are not always taken at

face value, especially when they contain cues that








9

contradict the manifest information. This research suggests

that the presence or absence of message cues affects the

types of information-seeking strategies adopted.

A second factor that may affect information-seeking

strategies is information set. Lord, Lepper, and Preston

(1984) suggest that people tend to adopt confirmatory

strategies unless specifically instructed to consider the

opposite of their beliefs. Even when asked to be unbiased

in their information gathering, people were not able to

overcome a tendency to ask confirmatory questions.

The current study hypothesizes that both instruction

set and message cues can prompt reporters to employ

different information-gathering strategies. When a message

contains no contradictions, it is accepted at face value and

receivers do not attempt to "second guess" its meaning. It

is expected that in these cases, journalists would employ

information-gathering strategies that are confirmatory; the

message is believed to be valid so there's no reason to

gather disconfirmatory information. However, when the

message contains a contradictory cue, it is hypothesized

that reporters will engage in less-biased information

gathering. Because they're unsure about the validity of the

message, they gather both confirmatory and disconfirmatory

information in an attempt to discover the message's true

meaning.










In this study, the researcher examines how reporters'

selection of information-gathering strategies is affected by

reading a message about an elderly man's automobile accident

that contains message- and instruction-related differences.

Previous research (Carver & de la Garza, 1984; Horvath-

Neimeyer & Kent, 1990) shows that when researchers ask

subjects to rate questions they would like to ask to learn

more about the accident, subjects prefer stereotype-

confirming questions. More specifically, subjects told

that a driver is elderly prefer questions about his physical

and mental capabilities. Subjects told the driver is young

prefer questions about his alcohol consumption and the

condition of his car. In the previous study, subjects had

no reason to question their stereotypes. However, it is

expected that when the message contains a cue that

contradicts the traditional stereotype of the elderly,

subjects will adopt a less-biased information-gathering

strategy.














CHAPTER 2

REVIEW OF THE LITERATURE

Researchers exploring cognitive functioning have long

used the scientific process as a metaphor for the way people

gather information. They theorize that people develop and

test hypotheses about their personal lives in much the same

way as scientists develop and test hypotheses about their

research topics (Kelley, 1973; Kelly, 1955). "Might not the

individual man, each in his own personal way, assume more of

the stature of a scientist, ever seeking to predict and

control the course of events with which he is involved?

Would not he have his theories, test his hypotheses, and

weigh his experimental evidence?" (Kelly, 1955, pp. 5-6).

But people's information-processing capabilities are

often biased (Lord, Ross, & Lepper, 1979; Ross, 1977; Simon,

1957). Researchers (e.g., Cacioppo & Petty, 1979;

Festinger, 1957; Houston & Fazio, 1989; Lord, Lepper, &

Preston, 1984; Sears & Chaffee, 1979; Rosenberg & Elliott,

1987) report that subjects selectively process information

from their environments, behaving not according to the

scientific model but as "faulty computers" (Higgins & Bargh,

1987, p. 397). The factors hypothesized to guide the

information selection and interpretation process include










various cognitive and cultural biases such as social

stereotypes.

Stereotyping of the Elderly

When a person is said to hold a stereotype about a

specific social group, the term refers to a cognitive

concept that describes group membership (Hamilton, 1979).

Lippmann writes that stereotypes are some of the "pictures

within our heads" (1922, p. 1), pictures garnered, in large

part, from the mass media and distinguished from events in

the real world. He regards stereotypes as inherently unfair

representations that can lead to social injustice. Since

the 1922 publication of Public Opinion, social scientists

have come to believe that while stereotypes can prove useful

in the assimilation and interpretation of information, they

do often lead to biased representations of social group

members (Brewer, Dull, & Lui, 1981; McCauley, Stitt, &

Segal, 1980; Rosch, 1978).

Empirical study of stereotypes often focuses on the

shared cognitive representation of social groups. In this

method, researchers provide subjects with lists of traits

and ask them to select the ones that pertain to specific

social groups. Results show that subjects prefer the traits

thought to constitute the stereotype or concept of that

social group. This study uses the term stereotypes in the

sense of belief systems that contain the widely accepted

generalizations about the characteristics of a group of










people (Williams & Best, 1982; Smith, 1989)--in this case

the elderly.

Researchers exploring stereotypes of the elderly

identify a set of traits that seems to commonly describe

perceptions of this group. Words or phrases such as lonely,

weak, less satisfied with life, negative, conservative,

physically and mentally unstable, and insecure describe the

shared traits often said to be shared by the elderly

(Brewer, Dull, & Lui, 1981; Weinberger & Millham, 1975;

Axelrod & Eisendorfer, 1961). Such traits, usually

negative, reflect how this social group is construed by

society (Jackson & Sullivan, 1988).

The portrayal of the elderly in the mass media often

echoes the negative public stereotype. Gerbner, Gross,

Signorielli, and Morgan (1980) found that television

represented the elderly as less successful than younger

people, comedic, eccentric, and foolish. Other research

shows that Israeli newspapers support the stereotype by

portraying the elderly as outside the main stream of life

(Shinar, 1982). Magazine ads portray the elderly as less

educated, inactive, and valueless members of society

(Kvasnicka, Beymer, & Perloff, 1982; Bramlett-Solomon &

Wilson, 1989). Arnoff (1974) found the elderly in

prime-time television drama associated with evil,

unhappiness, and failure.










People use stereotypes not only to guide their

cognitive processing of information but also in their

evaluation of information (Hamilton, 1979; Bodenhausen,

1988; Bodenhausen & Lichtenstein, 1987). This biased

process may in turn help to maintain the stereotype (Miller

& Turnbull, 1986). People may also maintain their

stereotypes through selective attention to

stereotype-consistent information. For example, researchers

report that subjects consistently recall predominantly

stereotype-confirming information, findings supportive of

the idea that people use biased information-processing

strategies (Darley & Gross, 1983; Lord, Ross, & Lepper,

1979; Merton, 1948; Rothbart, Evans, & Fulero, 1979).

Although few studies have investigated the role of

stereotypes in information seeking (Aboud, 1975; Carver & de

la Garza, 1984; Horvath-Neimeyer & Kent, 1990), many studies

have explored the role played by cognitive biases.

Information-Seeking Biases

The strategies lay interviewers use when involved in

actively searching for information have been the focus of

considerable research in social psychology. This research

indicates that interviewers often formulate questions to

gather information that confirms previous hypotheses or

beliefs. The notion of confirmatory bias is the focus of a

body of research (Snyder & Campbell, 1980; Snyder & Cantor,

1979; Snyder & Skrypnek, 1981; Snyder & Swann, 1978a; Snyder








15

& White, 1981) that suggests subjects prefer to ask sources

questions that elicit hypothesis-consistent information.

The finding that individuals gather information in a

less-than-logical manner is not new. Merton (1948) noted

the presence of a "self-fullfilling prophecy" in social

interaction. Wason (1960) provided empirical evidence for

Merton's proposition when he discovered that subjects

presented with a rule-discovery task often failed due to a

reluctance to test alternate hypotheses. Mynatt, Doherty,

and Tweney (1977) also reported that subjects failed to

attempt to falsify hypotheses, making observations that

would only confirm their personal theories.

But it is the research program of Snyder, Swann, and

their colleagues (Snyder & Campbell, 1980; Snyder & Cantor,

1979; Snyder & Skrypnek, 1981; Snyder & Swann, 1978a; Snyder

& White, 1981), in particular, that is held to demonstrate

the existence of a confirmatory bias in hypothesis testing.

Their studies are an offshoot of other research (Snyder &

Swann, 1978b; Snyder, Tanke, & Berscheid, 1977; Swann &

Read, 1981) that showed that people's behavior is often

altered to conform to others' expectations (confirmatory

behavior). In their seminal work on confirmatory bias in

hypothesis testing, Snyder and Swann (1978a) reported that

subjects engaged in gathering personal information from

others selected questions that confirmed the subjects'

previously held hypotheses.










Snyder and Swann (1978a) presented subjects with

information about the personality (introverted or

extroverted) of individuals the subjects expected to

interview. Those subjects testing the hypothesis that the

individual was an extrovert received a list of attributes

typical of only extroverted personalities, while subjects

testing the introverted hypothesis received only a list of

attributes typical of introverts. The subjects were then

asked to select 12 questions from a list of 26 to ask during

the upcoming interview. Researchers constructed the

questions and prerated them according to whether they were

a) the types of questions typically asked of extroverts, b)

the types of questions typically asked of introverts, or c)

neutral.

The researchers found that subjects tended to seek

hypothesis-confirming information. Those subjects provided

with the extrovert hypotheses were more likely to choose

questions that people would ask of known extroverts ("In

what situations are you talkative?") than were subjects

given the introvert hypothesis. Subjects told their

interviewee was an introvert were more likely to choose

questions that supported their hypotheses ("What things do

you dislike about parties?") than were subjects given the

extrovert hypothesis.

When Snyder and Swann (1978b) allowed subjects to ask

targets the questions they had selected, they found that the










biased strategies elicited responses from targets that led

individual judges to rate them in accordance with a

hypothesis. Although the researchers randomly assigned the

targets to condition, those in the extrovert group were seen

by judges as possessing more extrovert-consistent traits,

while those in the introvert group were seen as possessing

more introvert-consistent traits. The authors concluded

that biased information-seeking strategies can elicit

hypothesis-confirming behavior.

Snyder and Swann proposed that people had a tendency to

use biased information-collection strategies when engaged in

hypothesis testing. In particular, people chose methods

that would elicit information that confirmed their

previously held hypotheses and avoided methods that would

elicit disconfirming information.

The confirmatory bias model provides the basis for a

series of studies on hypothesis testing by Snyder and his

colleagues. Snyder and Cantor (1979) examined people's

biased use of historical information to test hypotheses. In

this study, researchers presented subjects with personality

information that portrayed a fictional woman named Jane as

either an extrovert or an introvert. Later, they asked

subjects to test Jane's suitability for one of two jobs,

real-estate sales or library research, selected because of

the occupation's need for either extroverted (sales) or

introverted (research) employees. The researchers wrote


I










that subjects reported more historical information that

confirmed that Jane possessed the correct personality for

whichever job was under consideration.

Snyder and Campbell (1980) attempted to reproduce the

first author's 1978 findings as well as respond to criticism

(Semin & Strack, 1980) that the original findings were

merely the result of the researchers' biased methodology.

To counter criticism that their original study provided

subjects with only one-sided hypotheses, the researchers

added a condition in which they provided subjects with

information that described both extroverts and introverts.

As in the other conditions, subjects were asked to test

either the hypothesis that the individual in question was an

extrovert or the hypothesis that the individual was an

introvert. Despite the addition of the new condition, the

researchers found continued support for confirmatory bias,

even when subjects had been provided with a full set of

attributes on both personality types.

Yet another twist on the original procedural paradigm

was added in a later study (Snyder & White, 1981). Here,

researchers told subjects to select questions from the list

that would either verify or falsify the personality

hypothesis that had been provided. In this case, when

subjects were specifically instructed to falsify their

hypotheses, they did indeed select questions that would

elicit disconfirming information. Subjects told to verify








19

their hypotheses, on the other hand, selected predominantly

hypothesis-confirming questions.

Snyder and his colleagues also extended the notion of a

confirmatory bias to hypothesis-testing situations involving

self-knowledge (Snyder & Skrypnek, 1981). Swann, the

co-author of the original work on confirmatory bias, and

Read (1981) reported similar findings of confirmatory bias

when individuals test hypotheses about themselves. They

concluded that subjects regarded confirmatory evidence as

more compelling and diagnostic, a conclusion echoed by

Snyder. "The processes of human thought foster, promote,

and almost ensure the ready and willing adoption of

confirmatory strategies by hypothesis testers" (Snyder,

1981, p. 295).

Swann and Giuliano (1987) extended Snyder and Swann's

(1978a) original work on confirmatory bias by allowing

subjects to generate their own questions to determine

whether respondents possessed specific personality traits.

In general, they found that subjects tended to generate

questions more likely to elicit information that confirmed

their hypotheses. By manipulating the strength of these

hypotheses, the researchers found that the more certain

subjects were of their beliefs, the greater the tendency to

employ confirmatory information-seeking strategies.

Subjects less certain of their beliefs tended to employ less










constraining strategies, searching for a mixture of both

confirmatory and disconfirmatory information.

In contrast to the research supportive of a

confirmatory bias, other studies have failed to duplicate

Snyder and Swann's findings. Strohmer and Chiodo (1984)

failed to find confirmatory bias when counselors were asked

to develop and test their own hypotheses about clients'

personalities. The results of a study by Semin and Strack

(1984), in which subjects were asked to construct their own

hypothesis-testing questions, also failed to find biased

information seeking.

Other research finds that when subjects are given a

choice between hypothesis-confirming questions and

hypothesis-disconfirming questions that vary in

diagnosticity, subjects select questions based not on their

ability to confirm a hypothesis but on their diagnostic

value (Bassok & Trope, 1983; Trope & Bassok, 1982, 1983;

Trope, Bassok, & Alon, 1984; Trope & Mackie, 1987).

Individuals tend to prefer questions that will elicit the

most highly diagnostic information for differentiating

between alternative hypotheses.

Trope and Bassok (1982) presented subjects with

handwriting samples and told them that they could determine

various personality traits by analyzing handwriting

features. The researchers found that the subjects

constructed questions that would discriminate maximally










between categories. There was no evidence that subjects

attempted to confirm hypotheses provided by the researchers.

Consistent with that study, Trope, Bassok, and Alon

(1984) found that when subjects were asked to create their

own questions, rather than selecting from a list, they

almost never construct hypothesis-consistent questions.

Instead, subjects preferred open-ended diagnostic questions

(e.g. "How do you usually spend your Friday nights?").

Such findings lead some to conclude that confirmatory

bias may be only one of many information-gathering

strategies employed by people (Darley & Fazio, 1980; Devine,

Hirt, & Gehrke, 1990; Fischhoff & Beyth-Marom, 1983; Trope &

Bassok, 1983). In a review of the literature on

confirmatory bias, Higgins and Bargh (1987) wrote that there

is considerable evidence that people do utilize a

confirmatory strategy when gathering information, although

it is not the only information-collection strategy people

use. In particular, people seem to prefer confirmatory

questions when there is only one hypothesis under

consideration and when they believe that hypothesis to be

true. They concluded that more research is needed to

pinpoint where and when confirmatory bias occurs:

The literature does not permit any clear
conclusions concerning whether there are
cognitive limitations in processing evidence
or whether preferred questions tend to
confirm the hypothesis. In fact, the two
major contingency factors--beliefs concerning
the likelihood that the hypothesis is true,
and accessibility of alternative competing










hypotheses--reflect the availability and
accessibility of social knowledge rather than
the actual process of hypothesis testing.
(p. 402).

Stereotypes and Information Seeking

Although few studies have been conducted examining the

role of stereotyping in information-seeking strategies,

researchers note a possible link. In fact, Snyder and Swann

(1978a) hint that confirmatory bias as envisioned in their

research would lead to the over-collection of confirming and

under-collection of disconfirming information by people.

Such biased strategies could result in the kinds of cultural

stereotypes that appear ineradicable even in the face of

disconfirming evidence.

Three studies have specifically investigated the role

of stereotypes in information seeking (Aboud, 1975; Carver &

de la Garza, 1984; Horvath-Neimeyer & Kent, 1990). Aboud

(1975) looked at the amount and kind of information subjects

sought about members of ethnic groups. He reported that

subjects tend to seek more information from group members

who conform most closely to ethnic stereotypes than from

nonconforming members. Aboud interpreted the results of his

study to suggest "that even though we may encounter many

people who disconfirm our stereotypes, we use only those who

confirm it as a source of information upon which to base

these expectations" (p. 339).

His findings parallel other research on the importance

of physical traits in stereotyping. These studies have










shown that providing subjects with stereotypic photos

triggered cognitive stereotypes (Cash & Duncan, 1984;

Goldstein, Chance, & Gilbert, 1984; Saladin, Saper, & Breen,

1988; Sunnafrank & Fontes, 1983). One study (Freeman, 1987)

indicated that images were even more potent than trait

information in the construction of stereotypes.

In a second study that linked information seeking with

stereotypes, Carver and de la Garza (1984) used ageism as

the focus of their research. The researchers theorized that

people tend to seek information supportive of their

stereotypes. More specifically, they hypothesized that

subjects instructed to gather information about an elderly

target would preferentially seek out information that

confirmed their stereotypes of the elderly.

Researchers presented subjects with a traffic accident

scenario that described the driver of one car as either an

elderly or a young man. Then subjects were asked to either

construct a list or select from a prepared list the

information they would need to make a determination of blame

in the auto accident. The prepared list consisted of nine

informational categories that were based on a pilot study

that sought to elicit the most important factors used in

assigning blame.

Carver and de la Garza found that in both the

self-guided search and the search using the prepared list,

subjects sought different kinds of information depending on










the stereotype initially activated. Subjects who had been

told the driver was elderly, preferred information that

concerned his physical and mental adequacy. On the other

hand, subjects who had been told the driver was young,

preferentially sought information about whether he had been

drinking.

A third study (Horvath-Neimeyer & Kent, 1990)

translated the Carver and de la Garza paradigm into a study

of news-gathering strategies. Instead of the

blame-assignment task previously given subjects, student

reporters were asked to decide on the newsworthiness of

information and questions. The researchers found that

student reporters were also primarily guided by age-related

stereotypes. More specifically, subjects told that the

driver was elderly tended to rate as more newsworthy

information and questions regarding his mental and physical

health. Subjects told the driver was young tended to rate

as more newsworthy questions and information regarding his

alcohol consumption and the condition of his car.

This finding provides support for the Stocking and

Gross (1989) contention that processes such as confirmatory

bias occur just as frequently in newsroom settings as in the

laboratories of psychological researchers. "Although there

no doubt are differences between journalists and the college

student populations upon which much of this research is

based, it is not unreasonable to suppose that journalists,










too, systematically fall victim to these cognitive biases

and errors" (p. 5). It also provides a theory about the

etiology of the news distortion noted by other researchers

(Commission on the Freedom of the Press, 1947; Tuchman,

1976) and lends empirical support to the primarily anecdotal

evidence of reporter-based errors (Ettema & Glasser, 1988;

Fishman, 1980).

The social importance of the news media makes such

concerns about the biased collection of information even

more important. News gives the impression of presenting a

realistic portrayal of events, so it has the power to mold

public perception. Because they are filtered through the

biased cognition of reporters, news stories may yield

inaccurate images that, nonetheless, may come to represent

reality to a vast audience. Stocking and Gross (1989b)

conclude that a thorough understanding of how journalists

are effected by cognitive biases is crucial.

What (journalists) say on the air and write
in their columns can set private and public
agendas; under some circumstances, they can even
initiate behavioral change. Given their
importance for both individuals and society, it is
imperative that we use all the insights at our
disposal--including insights about cognitive
constraints and processes--to understand better
how they do their jobs. (p. 5)

Factors That May Affect Biased Information Seeking

In some instances individuals may be able to overcome

their bias for information that confirms either stereotypes

or provided hypotheses. Lord, Lepper, and Preston (1984)










presented subjects with the extrovert-introvert scenario

developed by Snyder and Swann (1978a). They told subjects

they were to explore the character of a person in another

room. The questions the subjects selected should ascertain

whether that person was an extrovert. To discover if

subjects' question preferences could be manipulated, the

researchers divided subjects into three groups that each

received different instructions prior to selecting their

questions.

The control group received simply a replication of the

instructions provided by Snyder and Swann (1978a) in their

original study. A second group was admonished to be

unbiased in selecting questions. Lord, Lepper, and

Preston's (1984) instructions reminded subjects to be as

accurate as possible in order to provide a "fair and

accurate test of the person's true character" (p. 1238). A

third group was told to consider the opposite when selecting

questions. In addition, instead of providing subjects' with

a list of extroverts' traits, the researchers provided a

list of the traits of introverts. Researchers then told

subjects that "introverts are the opposite of extroverts, so

reading this profile should be just as helpful to you" (p.

1238).

As hypothesized, subjects in the control group tended

to ask more extrovert questions, confirming their hypotheses

about the source. Subjects who were instructed to be








27

unbiased also tended to ask hypothesis-confirming questions,

although to a lesser extent than those in the control group.

In contrast, subjects who were admonished to consider the

opposite showed no preference for either extrovert or

introvert questions.

The authors concluded that people tend to ignore the

possibilities of alternative hypotheses unless specifically

instructed to consider the opposite. Simply asking them to

be unbiased in their social judgments was not enough to

force their consideration of alternatives. They suggested

that people's inability to consider the opposite without

being instructed to do so was the result of inadequate

cognitive strategies.

In addition, studies (Lord, Lepper, & Preston, 1984;

Trope & Bassok, 1982, 1983) show that when an alternative

hypothesis is explicitly provided, subjects do not have a

preference for hypothesis-consistent questions. Higgins and

Bargh (1987) propose that the provision of an explicit

alternative hypothesis may be the most important factor in

subjects' decisions to adopt a unbiased information-seeking

strategy.

A second factor that may affect subjects' selection of

an information-seeking strategy involves whether the subject

receives contradictory information that might generate

alternative hypotheses. Researchers (Doelger, Hewes, &

Graham, 1986; Hewes & Graham, 1989; Hewes, Graham, Doelger,










& Pavitt, 1985; Hewes, Graham, Monsour, & Doelger, 1989)

propose that people seeking information about one hypothesis

will adopt a very different strategy than will people

seeking information about several hypotheses. Their

research is concerned with the types of contradictions that

generate alternative hypotheses.

They suggest that the selection of an information-

seeking strategy is based upon a person's understanding of a

communication event. When a person receives a contradictory

message, they reinterpret the message by generating several

different hypotheses or constructions of the message's

meaning. When no such contradictory information is

received, people often accept the message's premise at face

value.

The contradictions the researchers have explored

consist of either contradictory cues embedded within the

message or contradictions within the context of the message

recipient. A within-message cue is any information provided

as part of the message that makes the receiver doubt the

veracity of the message's overt content. Contextual cues

are usually factors that the receiver already knows such as

background knowledge, relational history, the goals of the

message sender, and the unwillingness or inability of the

sender to provide accurate information.

The notion that people reinterpret or develop

alternative hypotheses when confronted with contradictory










information is predicated upon the idea that people have a

need for uncertainty reduction (Berger & Calabrese, 1975;

Kagan, 1975). Uncertainty is reduced when people's innate

curiosity and need for social control makes them strive to

obtain veridical information. The greater the need for

accurate information, as with reporters who are expected to

present truthful news accounts, the greater the tendency of

individuals to carefully monitor messages for

contradictions.

Hewes, Graham, and their colleagues (Doelger, Hewes, &

Graham, 1986; Hewes & Graham, 1989; Hewes, Graham, Doelger,

& Pavitt, 1985; Hewes, Graham, Monsour, & Doelger, 1989)

assume that one way individuals may reduce their uncertainty

about a contradictory message is to consider alternative

constructions of its meaning. They can then test the

veracity of these alternative constructions through social

information gathering. The expected result is a

reconstruction of the meaning of the message that seems to

coincide more closely with reality.

The process of message reinterpretation begins when the

recipient receives a message in a particular context (Hewes

& Graham, 1989). An initial interpretation is accorded the

message. That initial interpretation is accepted at face

value unless either contradictions embedded within the

message or contradictory contextual factors cause the

recipient to doubt.










If there is reason to doubt the veracity of the

message, and there is a sufficiently high need for accuracy,

the recipient will engage in reinterpretation and the

generation of alternative hypotheses. After devising

alternative constructions for the message's meaning, the

recipient may engage in information gathering in an attempt

to collect information that will lead her or him to the

correct alternative message construction.

The research on the process, called second-guessing by

the authors, has illuminated various aspects. In their

seminal study of message reinterpretation, Hewes, Graham,

Doelger, and Pavitt (1985) provided descriptive evidence for

the presence of the phenomenon. The three studies reported

relied mainly upon self-report data in which 71% of the

subjects reported that they engaged in reinterpretation.

The authors also found that when subjects were asked to

choose between operationalizations of three different

strategies used for coping with contradictory messages

(ignore, weight differentially, or reinterpret) the subjects

who self-reported that they engaged in reinterpretation

selected the operationalization of that strategy.

Hewes and Graham next extended their exploration of the

ways people reinterpret messages by examining the types of

contradictions, called message cues, that provoked the

process (Doelger, Hewes, & Graham, 1986). In the first of

two studies, they used an open-ended format to ask subjects








31

about reinterpretation situations and message cues. To code

the responses, the researchers developed a category scheme

that divided the cues into four types: a) Target Cues

discrepant with the individual's prior knowledge, b) Source

Cues reflecting characteristics of the source that reveal

potential bias, c) Receiver Cues reflecting characteristics

of the recipient that make him or her distrustful, and d)

Message Cues, in which either components of the message are

contradictory or the form of the message deviates from what

is expected. The results indicated general support for the

authors' categorization.

Hewes, Graham, Monsour, and Doelger (1989) next turned

their attention to exploring the relationship between

reinterpretation and social information-gathering

strategies. To accomplish this, subjects were asked to

recall reinterpretation situations then answer a series of

both open-ended and Likert-type items. The authors found

that in 70% of the cases, subjects reported that after

reinterpreting messages they were involved in social

information seeking.

Although the authors did not explore what types of

information-gathering strategies subjects employed, their

study suggest a link between message reinterpretation and

information-gathering behavior. In fact, Hewes and Graham

(1989) note that: "Social actors must be seen as potentially

thoughtful, naive scientists who use theories of bias, not








32

unlike scientific theories, to explore the social world" (p.

238). The question remains, just what type of information-

seeking strategies are prompted by the presence or absence

of message contradictions?

It could be argued that the choice of an information-

gathering strategy depends, at least in part, upon whether

alternative hypotheses are either provided or generated by

the subject. Individuals encountering a message and no

contradictory information tend to accept the message at face

value. They have no impetus to construct alternative

hypotheses (Doelger, Hewes, & Graham, 1986; Hewes & Graham,

1989; Hewes, Graham, Doelger, & Pavitt, 1985; Hewes, Graham,

Monsour, & Doelger, 1989). Under these conditions,

confirmatory information-seeking strategies should be

maximized since the search for alternative explanations of

hypotheses is minimized. On the other hand, subjects

testing a hypothesis and one or more alternatives, such as

those generated when a message is reinterpreted, may employ

less constraining strategies in order to discern which

hypothesis is correct.

Such an explanation is supported by the literature on

information-gathering strategies. When confirmatory bias is

found (Snyder & Campbell, 1980; Snyder & Cantor, 1979;

Snyder & Skrypnek, 1981; Snyder & Swann, 1978a; Snyder &

White, 1981), it occurs when researchers provide subjects

with a single testable hypothesis. Bassok and Trope (1983)








33

also found that subjects exhibit a stronger preference for a

confirmatory strategy when they present a single hypothesis.

However, when they present subjects with both a hypothesis

and its alternative, subjects prefer a less constrained

information-gathering strategy.

Some research suggests that another factor involved in

information-seeking strategy selection may revolve around

the relative certainty of subjects with respect to their

hypotheses. Studies by Swann and his colleagues (Swann &

Ely, 1984; Swann & Giuliano, 1987) show, for example, that a

confirmatory strategy is generally preferred by subjects who

expressed certainty about their hypothesis. In contrast,

subjects more uncertain about their hypotheses tended to

search for a mixture of confirmatory and disconfirmatory

information. The researchers conclude that subjects who

were more uncertain were more open to different types of

information.

Hewes, Graham, Monsour, and Doelger (1989) write that

individuals tend to express less certainty when multiple

hypotheses are generated under conditions of message

reinterpretation. They found that when message

reinterpretation yields many competing hypotheses, subjects

were less confident of the veracity of their hypotheses.

Conversely, when the message was accepted at face value,

subjects expressed greater confidence in their hypothesis.








34

Taken together, these findings suggest when individuals

accept a message at face value they (1) generate a single

hypothesis, (2) express more confidence in that hypothesis,

and (3) are more likely to engage in confirmatory

information seeking. As Swann and Giuliano (1989, p. 523)

have observed, people "who are certain of their beliefs

assume that it is pointless and inefficient to 'beat around

the bush.'. . Instead, they simply assume that the belief

is correct and probe for confirmatory evidence using

constraining search strategies." In contrast, when people

reinterpret a message they (1) generate multiple hypotheses,

(2) express less confidence in their hypotheses, and (3)

employ an open information-seeking strategy that collects

both confirmatory and nonconfirmatory information.

Hypotheses

Because of the importance of information collection to

the production of news (Donohew, Tipton, & Haney, 1980;

Stocking & Gross, 1989a; Stocking & Gross, 1989b), this

study explores information-collection strategy selection

within a reporting context. It is based upon the notion

that reporters are often provided with information about a

source or event prior to actively gathering additional

information. The initial information may trigger the

activation of a stereotype that can, in turn, direct

information-seeking activities.










The hypotheses are founded upon the assumption that

individuals hold cognitive stereotypes of the elderly. This

assumption comes from the literature supporting the

existence of an ageist stereotype (e.g.: Brewer, Dull, &

Lui, 1981; Weinberger & Millham, 1975; Axelrod &

Eisendorfer, 1961; Jackson & Sullivan, 1988). These

stereotypes take the form of cognitive hypotheses about a

group of people similar to the hypotheses supplied to

subjects in confirmatory biases studies.

It is hypothesized that when a message received by an

individual about a stereotyped social group contains

information that would confirm the stereotype, individuals

will engage in confirmatory information seeking. However,

when a message contains information that contradicts these

stereotype hypotheses, individuals engage in message

reinterpretation, generating hypotheses that offer

alternatives to the stereotype hypothesis. From this comes

the first two hypotheses, which state that:

H1) Consistent with the results of Horvath-Neimeyer and

Kent (1990), when given internally consistent messages and

when not instructed to do otherwise, subjects will show a

confirmatory bias in information seeking.

H2) A main effect is predicted for second guessing such

that individuals subjected to information, within a message,

that contradicts the stereotype will generate lower levels










of confirmatory bias as reflected in the questions they

select.

The third hypothesis predicts an interaction of

instruction set and message cues on confirmatory bias.

First, for subjects not provided second-guessing cues, when

no explicit instructions about how to gather information are

provided, subjects will exhibit a strong confirmatory bias.

A weaker confirmatory bias is expected when subjects not

exposed to second-guessing cues are presented with

instructions to be unbiased and objective. Finally,

subjects not exposed to second-guessing cues who are told to

consider the opposite will show no preference for

confirmatory strategies. Rather, these subjects will select

as many stereotype-disconfirming questions as they do

stereotype-confirming questions.

Conversely, it is expected that all subjects exposed to

second-guessing cues will use equal opportunity strategies,

irrespective of instruction set.

More specifically, this final hypothesis states that:

H3) In the presence of contradictory information,

subjects in all instruction sets will adopt equal

opportunity information-seeking strategies. In the absence

of such contradictions, instruction set will determine the

amount of confirmatory bias. Subjects provided no explicit

instructions about how to gather information will exhibit a

strong confirmatory bias, while subjects instructed to be


1








37

unbiased and objective will exhibit a weaker confirmatory

bias. However, subjects told to consider the opposite will

exhibit no confirmatory bias.















CHAPTER 3

METHODS

Subjects

The sample consisted of 136 students, 92 women and 44

men, enrolled in advanced newswriting courses at the

University of Florida. Subjects' ages ranged from 19 to 45

with a mean age of 22. Forty-three percent of the subjects

were journalism majors and 21% anticipated working as a

reporter following graduation. All subjects had taken at

least a beginning course on newswriting.

The instrument took approximately 30 minutes to

administer. Subjects completed it during the first portion

of their newswriting course. Participation in the

experiment was voluntary.

Measures and Procedures

Care was taken to ensure that the experiment closely

resembled typical assignments in advanced newswriting

courses. Assignments in these courses generally consist of

providing students with a set of facts about an event.

Students are expected to use the facts to craft a newspaper

article and are encouraged to ask the instructors for

additional information about the event.


j








39

In this study, students were provided with facts about

an automobile accident and then asked to imagine themselves

as reporters assigned to cover the story. The instructions

specified that subjects should select questions (described

below) in order to gather additional information prior to

writing their stories. They were not, however, instructed

to write a story about the accident.

Manipulations

The stimulus consisted of a five-sentence paragraph

describing an automobile accident involving an elderly (74

years old) driver. The stimulus was based upon a similar

stimulus (Carver & de la Garza, 1974; Horvath-Neimeyer &

Kent, 1990) used to assess a confirmatory bias in question

selection. The message was photocopied and was typed on an

official "News Release" form from the local Gainesville (FL)

Police Department. A typed statement below the Gainesville

Police insignia designated that the release was an

"Example--for research purposes only."

Subjects were randomly assigned in a 2 X 3 (second

guessing by instruction set) factorial design.

All subjects read a four-sentence paragraph containing

the basic information about the accident and participants in

the accident. It read:

On Feb. 18, 1991, while driving to the
grocery store in a suburban area, an elderly man
named James Smith was involved in an automobile
accident. The car that he was driving collided
with another moving vehicle in which there were
two people. The 74-year-old Smith was uninjured.








40

The people in the other car were injured and were
taken to Alachua General Hospital.


The second-guessing cue manipulation was presented in

the fifth sentence, immediately following "Alachua General

Hospital." Subjects either read information consistent with

ageist stereotypes of a 74-year-old individual or

information inconsistent with stereotypes of an individual

of that age.

For this study, 67 subjects received the message with

the stereotype-consistent information (Appendix A) while 67

subjects received the message with the stereotype-

inconsistent information (Appendix B). For second-guessing

cue subjects, the final sentence read: "Smith, who works

for a local insurance company, was driving a red 1991

Porsche sports car." For those not assigned cues, the final

sentence read: "Smith, who is retired, was driving a white

1984 Dodge."

Subjects were also randomly assigned across three

levels of the second variable, instruction set. This

variable consisted of alterations to a set of general

instructions advising subjects how to select questions to

gather more information on the accident. The general

instructions either provided subjects with (a) no additional

information, (b) additional information advising them to "be

objective," or (c) additional information advising them to

"consider the opposite" of their stereotypic conceptions.










Forty-five subjects received instructions with no

additional information (Appendix C), 42 subjects were

additionally told to "be objective" (Appendix D), while 46

subjects were additionally told to "consider the opposite"

(Appendix E). Subjects were directed to read all the

instructions very carefully. The general instructions given

to all groups were as follows:

You're a reporter who's been assigned to
cover this story. If you were trying to write a
story about the auto accident that you've just
read about, you'd probably want to have more
information than you were given. Below is a list
of questions that might be useful in generating
more information. Imagine the source you'll be
confronting has specific answers to each of the
questions. We are interested in which of these
questions you see as most important in gathering
the information necessary to write your story.
The questions you're choosing have all been
suggested by working journalists. Please remember
that you are acting as a reporter gathering more
information about this man's automobile accident.

Subjects in the "be objective" and "consider the

opposite" conditions were provided with additional

instructions contained in a second paragraph. Both are

based on additional instructions provided subjects in Lord,

Lepper, and Preston (1984). Subjects in the "be objective"

condition were told:

Also remember that you, as a reporter, should
be as accurate and objective as possible when
searching for information. What we want is for
you to be as unbiased as possible in providing a
fair and accurate description of this man's
automobile accident. That means collecting
information that highlights all the potential
sides of an issue. You would then sift through
the information to select and use information that
fairly portrays all those different sides. Most


I








42

journalists and journalism teachers refer to this
process as objectivity. They often stress that
all reporters should strive to be as objective as
possible. That's what we want you to do in this
assignment--be objective.


For subjects in the "consider the opposite" group, the

additional paragraph of instructions read:

Also remember that you, as a reporter, should
always try to take all viewpoints into
consideration, even viewpoints that appear in
conflict with stereotypes about certain groups of
people. However, considering all viewpoints is
often difficult because people tend to ignore
information that doesn't agree with their own
stereotypes. When completing this questionnaire,
keep in mind that elderly people often have the
same strengths and weaknesses as younger people.
Ask yourself when choosing each question whether
you would have made the same choice if Smith had
been a much younger person. In other words,
consider the opposite of your own cultural
stereotype of the elderly.

Dependent Measures

Three dependent measures were selected for this study.

The first, a list of 15 questions, was created to assess

subjects' information-gathering strategies. The second, 10

items designed to indicate subjects' personal constructions

about the cause of the accident, were included to explore

subjects' naive assumptions about what had occurred. These

measures could help illuminate whether subjects both thought

and behaved in identical ways. The final set of dependent

measures, eight items were selected to identify subjects'

personal beliefs about the elderly driver. It was

anticipated these global beliefs might help explain why








43

subjects thought and behaved as they did toward a particular

individual.

The 15 questions followed the presentation of the

message and instructions. Subjects examined the list and

ranked the questions from "1" to "15" in terms of their

perceived importance for gathering more information. For

example, a question ranked "1" would be considered the most

important in gathering additional information while a

question ranked "15" would be considered the least

important.

The 15 questions were selected from a list of 21

questions presented to a sample (n=21) of advanced

newswriting students. The questions had been derived from

similar ones used in previous studies investigating the age-

related paradigm (Carver & de la Garza, 1974; Horvath-

Neimeyer & Kent, 1990). During the pretest, subjects rated

the stereotypicality of the 21 questions using a 7-point

numerical scale ranging from "Stereotypic of Elderly" to

"Stereotypic of Young People." Five questions assessing

elderly stereotypes (stereotype consistent), five questions

assessing stereotypes of young people (stereotype

inconsistent), and five neutral questions (neither

consistent with elderly people or consistent with young

people) were selected based upon their pretest ratings. The

stereotype-consistent and stereotype-inconsistent questions

selected were those with both the most extreme










stereotypicality ratings and the smallest standard

deviations. Neutral questions were those with the ratings

closest to the midpoint that had the smallest standard

deviations (Appendix F).

Subjects next completed a series of demographic items,

including questions concerning their age, gender, major, and

interest in journalism.

A series of 10 items concerning the likelihood of

various causes of the accident followed. Subjects rated

each item using five-point scales ranging from "Very Likely"

to "Very Unlikely." The items were based on the 10

questions that were either stereotypic of the elderly or

stereotypic of young people, as assessed during the pretest.

The questions were modified to gauge subjects' assessments

of the cause of the accident. For example, the question

"How good was Smith's hearing" was modified as "To what

extent do you think it's likely that Smith might have had

poor hearing that contributed to the cause of the accident?"

Next subjects completed a series of manipulation

checks. These items assessed whether subjects could

correctly recall if they'd received a second-guessing cue

containing stereotype-inconsistent information ("Smith

worked for a local insurance company" and "Smith drove a red

1991 Porsche sportscar") or stereotype-consistent

information ("Smith was retired" and "Smith drove a white

1984 Dodge"). In addition, subjects were asked how certain










they were of receiving instructions that told them to "be

objective," "consider the opposite," or "neither."

The final section of the questionnaire assessed images

of the elderly driver. It consisted of a series of eight

semantic-differential type items that pertain to common

images of the elderly. These items were either based upon

the accident scenario or drawn from the work of Axelrod and

Eisendorfer (1961). The items were: typical 74-year-old

vs. unusual 74-year-old, frail vs. strong, bad driver vs.

good driver, confused vs. alert, incompetent vs. competent,

poor vision vs. good vision, physically unstable vs.

physically stable, and insecure vs. secure (see Appendix G

for entire questionnaire).

To eliminate the possibility of an order effect, six

versions of the questionnaire were prepared. Three

randomized presentations of items within the sections

containing the 15 questions, the 10 items concerning the

likely cause of the accident, and the eight semantic-

differential type items were prepared. In addition, the

presentation of the sections was also randomized.














CHAPTER 4

RESULTS

Preliminary Analysis

Prior to conducting the primary analyses for the study,

checks were performed in order to assure the successful

manipulation of independent variables, cue, and instruction

set. Two protocols were completed incorrectly and were

discarded, leaving a total of 134 subjects.

A series of manipulation checks assessed whether

subjects correctly recalled the instruction set and whether

they'd received a second-guessing cue. First, subjects

rated the three instructions ("be objective," "consider the

opposite," and no additional information) using a 5-point

scale ranging from "Very Sure Was Told" to "Very Sure Not

Told." Next subjects recalled what they had been told about

the elderly driver of the car. The stereotypic information

("Smith was retired," and "Smith drove a white 1984 Dodge")

as well as the second-guessing cues ("Smith worked for a

local insurance company," and "Smith drove a red 1991

Porsche") were rated using an identical 5-point scale.

A series of 2 (message type) X 3 instructionn set)

ANOVAs were conducted to test the effectiveness of the three

instructions. To ascertain whether subjects recalled the










instructions they had received, they were asked in three

separate items whether they had been told to "be objective,

"consider the opposite," or neither. Certainty ratings,

made following each of the three items, were made on a five-

point scale with "one" indicating the greatest degree of

certainty. Main effects for instructional set were revealed

for the "be objective" instruction, F(2,128)=19.8,p<.0001,

the "consider the opposite" instruction,

F(2,128)=18.08,p<.0001, and neither instruction,

F(l,128)=4.41,p<.02. No other main effects or interactions

reached significance.

Pairwise comparisons were conducted to test for

significant differences. The direction of the main effect

for subjects in the "consider the opposite" group was

consistent with expectations. Subjects in the "consider the

opposite" group indicated greater certainty that is what

they had been told (M=2.04) than either those in the "be

objective" group (M=3.37), F(1,86)=19.02,p<.0001, or the

neither instruction group (M=3.81), F(1,89)=33.95,p<.0001.

The direction of the main effect for subjects in the

"be objective" group was partly consistent with

expectations. Subjects in this group indicated

significantly greater certainty (M=1.69) that they had been

told to "be objective" than did subjects in the no

instruction group (M=3.44), F(1,85)=18.13,p<.0001. However,

subjects in the "consider the opposite" group were










significantly more certain they had also been told to "be

objective" (M=2.14) than were subjects in the neither

instruction group, F(l,89)=33.90,p<.0001. No significant

difference was found between subjects in the "consider the

opposite" and "be objective" groups.

The manipulation check for the no instruction group was

more problematic. Subjects in the no instruction group

indicated less certainty they had received neither

instruction (M=4.23) than did those in either the "consider

the opposite" group (M=3.43), F(l,52)=5.10,p<.03, or the "be

objective" group (M=3.19), F(l,66)=8.32,p<.005.

A second series of 2 (message type) X 3 (instruction

set) ANOVAs was conducted to test the effectiveness of the

cue manipulation. Two information bits ("Smith was retired"

and "Smith drove a white 1984 Dodge") were stereotype

consistent while two other information bits ("Smith worked

for a local insurance company" and "Smith drove a red 1991

Porsche sports car") were stereotype inconsistent. All

significant main effects were in the anticipated direction.

Main effects for cue were revealed for both of the

stereotype-consistent cues. Subjects in the "no cue"

condition were more certain they had been told that the

elderly subject was retired (M=2.57), F(l,132)=66.7,p<.0001,

and drove a white Dodge (M=1.34), F(l,132)=986.44,p<.0001,

than subjects in the "cue" condition (M=4.5 and M=4.9,

respectively). Likewise, subjects in the "cue" condition








49

were more certain they had been told that the subject worked

for a local insurance company (M=1.65),

F(l,132)=286.42,p<.0001, and drove a Porsche (M=1.12),

F(1,132)=4,136.96,p<.0001, than were subjects in the "no

cue" condition (M=4.63 and M=4.95, respectively).

Primary Analyses

Primary analyses were conducted with three sets of

items. The first set assessed information-seeking

strategies. It consisted of a list of 15 questions that

subjects ranked according to the importance of the

information each question would generate. The second set

assessed perceived causality and consisted of a list of 10

items that subjects rated according to how likely or

unlikely each factor (stereotype-consistent, stereotype-

inconsistent, neutral) was as a potential cause of the

accident (e.g., "To what extent do you think it's likely

that fast driving by Smith might have contributed to the

accident?"). The final set consisted of a list of eight

semantic-differential items (e.g., "Frail vs. Strong")

assessing the subjects' personal conceptions of the elderly

driver.

To analyze information-seeking strategies, each of the

15 questions was weighted according to its degree of age

stereotypy. These weights were derived from the pretest

data (Appendix F). The weights were standardized by

subtracting the mean for each variable from individual








50

observations and then dividing by the standard deviation for

that variable. They were then multiplied by subjects'

importance ratings (reverse coded so 15=most important,

l=least important) for each of the 15 items. To eliminate

the difficulty of dealing with negative numbers, the

absolute values of each variable were used in the analysis.

For each subject, the products of these standardized

weights and ranks for the five stereotype-consistent

questions were combined to form one dependent variable, and

the products for the five stereotype-inconsistent questions

were combined to form a second dependent variable. More

extreme scores indicated that subjects believed these

questions were more important in gathering additional

information. The neutral questions were not utilized in the

analysis as they contributed nothing to the hypotheses being

explored. They were included in the initial list as

distractor items.

Cronbach alphas were calculated for the stereotype-

consistent items and for the stereotype-inconsistent items.

The resulting alpha for the stereotype-consistent index was

0.54 while the alpha for the stereotype-inconsistent index

was 0.45.

A mixed ANOVA was conducted treating the two dependent

variables described above, stereotype-consistent and

stereotype-inconsistent indexes, as a within-subjects

variable and message type (cue vs. no cue) and instruction










set (no instruction, "be objective," and "consider the

opposite") as between-subjects variables. A main effect for

index, F(l,127)=84.34,p<.0001, was found (Table 1).

Subjects ranked stereotype-consistent questions more highly

(M=9.89) than they did stereotype-inconsistent questions

(M=7.14). In other words, subjects thought questions

regarding the driver's physical and mental condition more

important than questions dealing with the driver's possible

drinking and the speed of his car.

To ensure that the neutral items exerted no effect upon

the relationship between the stereotype-consistent and

stereotype-inconsistent items, a second mixed ANOVA was

conducted. In this case, three within-subjects variables

were created by calculating the mean rank for the five

neutral, the five stereotype-consistent, and the five

stereotype-inconsistent items. None of these means were

weighted. Message type and instruction set were treated as

between-subjects variables. Consistent with the previous

analysis, only a main effect of index,

F(2,254)=95.84,p<.0001, was found (Table 2). Stereotype-

consistent items were ranked more highly (M=10.34) than

either stereotype-inconsistent items (M=7.56) or neutral

items (M=6.21), showing that the neutral index had no

significant effect on the rankings of the other two indexes

(Table 3).











Table 1


Mixed ANOVA of Weighted Rankings for Each Index by Cue and
Instruction (Ins.)

Within-Subject Effects
-----------------------------------------------------------
SUM OF MEAN
SOURCE SQUARES DF SQUARE F-VALUE PROB > F
-----------------------------------------------------------
Index 502.518 1 502.518 84.34 .0001
Index X Cue 1.795 1 1.795 0.30 .5840
Index X Ins. 2.687 2 1.344 0.23 .7984
Index X Cue X Ins. 5.046 2 2.523 0.42 .6557
Error 756.708 127 5.958



Between-Subject Effects

SUM OF MEAN
SOURCE SQUARES DF SQUARE F-VALUE PROB > F

Cue 9.726 1 9.726 4.35 .0390
Ins. 5.987 2 2.993 1.34 .2657
Cue X Ins. 7.980 2 3.990 1.79 .1719
Error 283.843 127 2.235
















Table 2

Mixed ANOVA of Mean Rankings for Each Index by Cue and
Instruction (Ins.)

Within-Subject Effects
-----------------------------------------------------------
SUM OF MEAN
SOURCE SQUARES DF SQUARE F-VALUE PROB > F
-----------------------------------------------------------
Index 1,192.523 2 596.262 95.84 .0001
Index X Cue 22.494 2 11.247 1.81 .1661
Index X Ins. 44.272 4 11.068 1.78 .1335
Index X Cue X Ins. 22.760 4 5.690 0.91 .4560
Error 1,590.299 254 6.222
-----------------------------------------------------------


Between-Subject Effects
----------------------------------------------------------
SUM OF MEAN
SOURCE SQUARES DF SQUARE F-VALUE PROB > F
-----------------------------------------------------------
Cue 0.626 1 0.626 1.33 .2504
Ins. 1.213 2 5.607 1.29 .2783
Cue X Ins. 0.931 2 0.465 0.99 .3739
Error 59.625 127 0.469
-----------------------------------------------------------











Table 3


Means and Standard Deviations for Indexes by Cue and
Instruction
Young Old Neutral


M=7.74 M=10.57 M=6.35
Ins. 1 SD=1.89 SD=3.71 SD=1.58

No M=7.12 M=9.44 M=5.90
Cue Ins. 2 SD=2.12 SD=1.82 SD=2.29

M=7.52 M=10.63 M=5.90
Ins. 3 SD=1.77 SD=1.74 SD=1.66

-- -----------------------------------------------

M=8.00 M=10.22 M=5.72
Ins. 1 SD=2.14 SD=2.03 SD=2.17

M=7.81 M=10.67 M=5.45
Cue Ins. 2 SD=2.32 SD=2.31 SD=1.72

M=7.22 M=10.55 M=6.21
Ins. 3 SD=1.29 SD=2.04 SD=1.66

-- ------------------------------------------------

Note: Ins. = Instruction Set; 1 = Consider the Opposite,
2 = Be Objective, and 3 = No Additional Instructions.










The second set of primary analyses examined the 10

causality items. The 10 items were divided into two sets,

one set reflected stereotype-consistent causes of the

accident and the other reflected stereotype-inconsistent

causes of the accident. Cronbach coefficient alphas were

calculated for both the stereotype-consistent (0.68) and

stereotype-inconsistent (0.66) indexes.

A mixed ANOVA was conducted treating the two dependent

variables described above, stereotype-consistent and

stereotype-inconsistent indexes as a within-subjects

variable and message type (cue vs. no cue) and instruction

set (no instruction, "be objective," and "consider the

opposite") as between-subject variables. A main effect for

the index, F(1,128)=29.82,p<.0001, was found (Appendix J).

Subjects rated stereotype-consistent items as more likely

causes of the accident (M=2.74) than they did stereotype-

inconsistent items (M=3.18). As in their information-

seeking strategies, subjects responded that the driver's

physical and mental condition were more probable causes of

the accident than were the driver's possible drinking and

the speed of his car.

In addition, an interaction between message type and

the repeated factor was discovered, F(l,128)=4.35,p<.04. A

simple effects test indicated that there was a significant

difference along the stereotype-inconsistent index between

the cue condition (M=3.07) and the no-cue condition










(M=3.30), F(1,132)=4.04,p<.05. When rating items

consistent with a youthful stereotype according to how

probable they were as causes of the accident, subjects who

were told the driver was employed and driving a Porsche

rated these items as more probable causes. In contrast,

subjects told the driver was retired and driving a Dodge

rated these stereotype-inconsistent items as less probable

causes of the accident. There was no corresponding

difference along the stereotype-consistent index (cue

M=2.83; no cue M=2.64).

The third set of analyses examined the eight semantic-

differential-type items that measured personal constructions

of stereotypicality. Two of the items had been reverse

coded so their ratings were reversed. The eight items were

then compiled into a single index by adding them together,

then taking the mean. This number reflected subjects'

personal stereotypic constructions of the elderly driver.

Cronbach coefficient alphas were calculated for the index to

determine the reliability. The resulting alpha was 0.79.

An ANOVA (message type X instruction set) was conducted

using the index described above. A main effect for message

type was revealed, F(1,130)=58.79,p<.0001 (Appendix K).

Subjects rated stereotype-consistent items more highly

(M=2.77) in the no-cue condition than in the cue condition

(M=3.33).














CHAPTER 5

DISCUSSION

Subjects' choice of information-seeking strategies were

consistent with the results of Horvath-Neimeyer and Kent

(1990). When asked to rank questions they would like to ask

regarding the accident, subjects ranked age-stereotypic

questions, such as those regarding the driver's mental and

physical capabilities, higher than they did non-stereotypic

questions (items that pertained to a "youthful" stereotype).

This result offered support for the first hypothesis,

that subjects will show a confirmatory bias in information

seeking. However, that confirmatory bias was present

regardless of the presence or absence of a cue. In fact,

the study failed to show a main effect for second guessing

as had been hypothesized in the second hypothesis. Subjects

ranked the stereotype-consistent questions higher regardless

of their cue condition.

The third hypothesis predicted an interaction between

instruction set and message cues. It had been anticipated

that subjects exposed to second-guessing cues would exhibit

a strong confirmatory bias if given no explicit instructions

about how to gather information. In contrast, if they had

been told to be unbiased and objective, subjects should have








58

exhibited a weaker confirmatory bias, while subjects told to

consider the opposite would show no preference for a

confirmatory strategy. It was expected that all subjects

exposed to second-guessing cues would also show no

preference for confirmatory strategies, irrespective of

instruction set.

However, the hypothesis was not supported by the

results. Once again, subjects ranked stereotypic questions

higher regardless of the instructions they had received.

Similar results were exhibited in both the items

assessing causality for the accident and personal

constructions of the elderly. In both cases, subjects rated

the stereotype-consistent items more highly than they did

the stereotype-inconsistent items. Neither the presence of

a second-guessing cue nor instruction set had a significant

effect on confirmatory bias or personal constructions of the

elderly.

Support for an effect of cue was found only when

subjects were asked to assess the probable cause of the

accident. When rating items not consistent with an ageist

stereotype according to how probable they were as causes of

the accident, subjects who were told the driver was employed

and driving a Porsche rated these items as more probable

causes. In contrast, subjects told the driver was retired

and driving a Dodge rated these stereotype-consistent items

as less probable causes of the accident.










Limitations

The results of this study offer support for the

reliability of the information-seeking strategy instrument.

Similar results were obtained by Horvath-Neimeyer and Kent

(1990) and Carver and de la Garza (1984) using similar

instruments. In particular, the instrument used by Horvath-

Neimeyer and Kent differed only on a few stereotype-

inconsistent questions. These questions were added to the

current study to ensure that there were equal numbers of

stereotype-consistent and stereotype-inconsistent items.

There are, however, several possible limitations of

this study. One potential limitation may occur in the use

of a single stimulus in which the manipulations for second-

guessing cue and instruction were embedded. The failure to

detect the expected significant results due to the presence

of cue or various instructions may be attributable to

problems inherent within that stimulus. Other stimuli may

have not yielded the same results.

A similar limitation may occur due to the study's

exclusive focus on ageism. Since ageism was used as the

single instance of stereotyping, it is unclear how the

results would translate into other cases of stereotyping.

Moreover, because the sample was so homogeneously young,

consisting predominantly of college students in their late

teens and early twenties, the external validity of the










results in relation to a larger population remains an

unanswered question.

External validity may also be affected if the results

of this study, which utilized student reporters, are

extended to professional journalists. For example, Dimmick

(1974) found that professional journalists judged the

newsworthiness of stories very differently from

nonprofessional subjects. Although his work would suggest

the inadvisability of extending results to a professional

population, there is an important difference between these

two studies. As opposed to this study, Dimmick's work

utilized undergraduates who were not involved in journalism

classes. In this study, subjects were all members of

undergraduate journalism classes, a factor that should have

given them greater insight into professional practices and

make the extension of results more feasible.

Although care was taken to replicate actual classroom

reporting experiences, the use of provided questions does

not correspond well to what occurs in an authentic news-

gathering situation. In the real world, reporters engage in

interactive communication with their sources, altering their

questioning strategies according to responses they receive.

The absence of interpersonal interaction in this study may

have affected subjects' choice of stereotypic questions,

raising questions again of external validity.










Result Interpretation

The results of this study indicate that cognitive

biases such as stereotypes may guide journalists' news-

gathering strategies, at least in certain cases. In

particular, the findings are supportive of the notion that

cognitive representations of social age groups have a

bearing on the information and questions student journalists

regard as most important when considering writing a story.

In other words, people believe age-stereotypic questions

will generate more important information than will questions

that revolve around factors not typically associated with

the elderly. These results generally replicate those of

Carver and de la Garza (1984) and Horvath-Neimeyer and Kent

(1990).

The results also indicate subjects utilized the same

types of shared stereotypic traits of the elderly reported

by Axelrod and Eisendorfer (1961). Namely, older targets

are considered more physically and mentally impaired than

are younger people. These stereotypic traits were noted

both in the items that measured perceived casualty of the

accident and personal constructions of the elderly. In

particular, the latter consisted of eight semantic-

differential type items that pertained to common images of

the elderly as enumerated by Weinberger and Millham (1975).

The results of this study offered support for the

universality of these age-related traits.










But those age-related traits had greater cognitive

implications than simply providing a means by which subjects

could create a mental picture of an individual. Subjects

also utilized those traits when assessing the probability of

causes for the automobile accident described in the

stimulus. More specifically, subjects rated as more

probable causes of the accident items that focused on

stereotypic age-related traits. For example, subjects rated

it more likely that poor hearing and poor eyesight were the

causes of the accident than were reckless driving and the

drinking of alcoholic beverages.

Cognitive stereotypes were even further extended when

subjects were asked to select questions they'd like to ask

to gather more information about the accident. Once again,

questions that focused on stereotypic age-related traits

(e.g., What was Smith's general physical condition?) were

rated more highly than were questions that revolved around

traits not considered stereotypic of the elderly (e.g., Was

Smith driving too fast?).

These results lend support to the notion that

stereotypes may be considered cognitive schemas that, once

activated, guide cognitive processing. In this study, when

the subjects' age-related schema were activated by

mentioning the driver's age, the stereotypes guided both the

ratings of perceived causality and rankings of questions.

This finding is consistent with research that shows that








63

cognitive representations such as stereotypes are likely to

influence subsequent behavior (Hamilton 1979; McCauley,

Stitt, & Segal, 1980).

More problematic is the failure in this study to detect

significant effects of the second-guessing cue. The lone

effect of cue appeared when subjects provided a second-

guessing cue were asked to assess the probable cause of the

accident and only when rating items not consistent with an

ageist stereotype. These subjects rated items not

consistent with an ageist stereotype as more probable causes

of the accident. In contrast, the subjects who were not

provided a cue rated these stereotype-inconsistent items as

less probable causes of the accident. This discovery of a

cue effect was consistent with expectations. The question

that arises is why wasn't such an effect discovered

elsewhere?

Just as puzzling as the inability to find a cue effect

is the failure of this study to show an effect of

instruction. It was expected that subjects told to

"consider the opposite" or "be objective" would show lower

levels of confirmatory bias. Such was not the case.

Instructions seemed to have little bearing on subjects'

ranking of questions, ratings of perceived causality, or

personal constructions of the elderly.

Several possible explanations of these lack of findings

can be advanced. The simplest explanation is that the cue










and instruction manipulations did not operate as planned.

It may be, for example, that the cue used did not create the

type of second guessing reported by other researchers

(Doelger, Hewes, & Graham, 1986; Hewes & Graham, 1989;

Hewes, Graham, Doelger, & Pavitt, 1985; Hewes, Graham,

Monsour, & Doelger, 1989). The same may have been true of

instruction set. Although care was taken to transform the

instructions utilized by Lord, Lepper, and Preston (1984),

the wording was altered to conform to the type of

instructions generally provided in a reporting course. The

power of the instructions may have been lost in the

translations.

Although the explanation of a loss of power in both the

instructions and cue may be attractive, it is unlikely

considering the results of the study's manipulation checks.

All significant main effects for the cue manipulation were

in the anticipated direction. In addition, main effects for

subjects in the "consider the opposite" and the "be

objective" groups were consistent with expections. While

subjects in the no instruction group were less certain of

the instructions they had received, their uncertainty may

have been due, at least in part, to the potentially

confusing wording of the check (they were asked if they had

received "neither of the above" instructions). If it can be

assumed that the manipulation generally worked, then what

could have caused the lack of results?










A second explanation might focus upon the power of

cognitive schemata. Bartlett (1932) proposed that new

information is organized by interacting with old information

already within a person's schema. Neisser (1976) wrote that

schemataa are anticipations, they are the medium by which

the past affects the future" (p. 22). People use these

schemas, or past memories, to help define their responses to

new situations. Once a schema has been "primed" or

activated it is more likely to be used to encode new

information and guide an individual's responses (Higgins,

Rholes, & Jones, 1977; Wyer & Srull, 1981). Brewer, Dull,

and Lui (1981) found that people's schema of the elderly

guided their behavior on tasks such as statement and picture

sorting.

If an elderly schema were primed in this study in the

the accident scenario by mentioning the driver was "an

elderly man" and "74 years old," the schema could have been

used to guide individuals in question selection as well as

selection of probable accident causes and personal

constructions of the elderly. That schema may have been

strong enough to enable the subjects to filter out other

factors such as second-guessing cues and instruction set.

The notion that journalists may possess their own norm

for information seeking is the basis for yet another

possible explanation. This norm, which may be unique to the

news profession, compels both professional and student










journalists to seek the most newsworthy information and to

disregard outside forces that may attempt to alter their

information-seeking strategies. For example, reporters

usually ignore requests from the general public as well as

sources to withhold information or to slant their stories in

a certain manner.

It is not unreasonable to think the subjects in this

study responded similarly. As student journalists, the

subjects had presumably already been exposed to the norms of

reporting in earlier classes. Perhaps this information-

seeking norm prevented the subjects in this study from

reacting as expected to the two manipulations. Instead,

they operated within the boundaries of the norm by ignoring

instructions or message cues that attempted to make them

revise their normal searches.

Further Research

This study has prepared fertile ground for future

research. The importance of such research has been pointed

out by various researchers. McCauley, Stitt, and Segal

(1980), although not specifically referring to mass

communication research, underscored the importance of

learning the origins of such stereotypes. "The extent to

which social reality testing becomes social reality

construction (Snyder & Swann, 1978) must be an important

research question for the future" (p. 205).








67

More specifically, while this study explored what types

of questions student reporters ask sources, other studies

might focus on later stages of the reporting process. For

example, further studies could focus on how stereotypes

affect what information journalists select to use in their

news stories. It is expected that such stereotypes would

also guide reporters to exclude less stereotypic information

in favor of more stereotypic items.

Further research could also examine how other

stereotypes effect both news gathering and information

selection. Researchers have shown that racial, cultural,

and other stereotypes influence how people process

information. Do reporters also rely upon these other

stereotypes when involved in gathering and relaying

information?

To avoid questions of external validity, researchers

should extend this research to professional journalists. By

virtue of their relative inexperience, the news gathering

and information selection behavior displayed by students may

differ from that displayed by professionals.

Finally, further research might also focus on factors

that could alter subjects' stereotype-consistent responses.

For example, empathy with the stereotype source, previous

experience with a particular stereotype, and deadline

pressure may all effect the responses of journalists.










Importance to the Discipline

The findings of this study are important because they

show the effects of cognitive biases on student reporters

and, if generalization is possible, working professional

journalists. The implications are especially serious if

such biased information seeking results in skewed versions

of events portrayed in news reports. The concern here is

not whether journalists' stereotypes are valid, and in fact

some people may argue that stereotypes aid journalists by

guiding them toward questions that have the greatest chances

of obtaining information regarding people in a certain

social group.

The concern focuses more on how heavily such

stereotypes or cognitive representations are utilized. If

reporters rely upon their personal constructions of social

groups to selectively ask to stereotype-consistent

questions, they may fail to collect important information

that is inconsistent with their stereotypes. For example,

when confronted with a automobile accident involving an

elderly driver, reporters may ask numerous questions about

the driver's mental and physical health. Unless they

transcend their stereotype, however, they may ignore

questions regarding the driver's drug and alcohol use, and

may then fail to collect important information. Such subtle

stereotyping may be difficult to detect within news reports








69

themselves but may help perpetuate the public stereotype of

a particular group.

The social importance of the news media makes such

concerns about the biased gathering of information even more

critical. News gives the impression of presenting a

realistic portrayal of events, so it has the power to mold

public perception. Because they are filtered through the

possibly biased cognitions of reporters, news stories may

yield stereotypic images that, nonetheless, may come to

represent reality to a vast audience.

The importance of such research on the psychology of

journalists is underscored by the appearance of the new

media technology, which extends the mass media's reach into

a worldwide market. It mirrors the fears of the press's

increasing technical might that sparked the formation of the

Commission on the Freedom of the Press nearly 50 years ago.

The question then focused on the conscious ability of the

journalists to act responsibly; the questions today involve

more subtle unconscious mechanisms. And while research on

cognitive stereotypes won't ultimately resolve all the

questions of how journalists think and conduct their

work, it will enrich the ongoing discussion of these issues.














APPENDIX A
STEREOTYPE-CONSISTENT MESSAGE












On Feb. 18, 1991, while driving to the grocery store in a suburban area, an

elderly man named James Smith was involved in an automobile accident. The car

that he was driving collided with another moving vehicle in which there were two

people. The 74-year-old Smith was uninjured. The people in the other car were

injured and were taken to Alachua General Hospital. Smith, who is retired, was

driving a white 1984 Dodge.










APPENDIX B
STEREOTYPE-INCONSISTENT MESSAGE








On Feb. 18, 1991, while driving to the grocery store in a suburban area, an

elderly man named James Smith was involved in an automobile accident. The car

that he was driving collided with another moving vehicle in which there were two

people. The 74-year-old Smith was uninjured. The people in the other car were

injured and were taken to Alachua General Hospital. Smith, who works for a local

insurance company, was driving a red 1991 Porsche sportscar.













APPENDIX C
NO ADDITIONAL INFORMATION CONDITION INSTRUCTIONS

Please feel free at any point during this questionnaire to return to the previous page in
order to re-read the press release. As you go along remember to read all the
instructions very carefully.




ASSIGNMENT



You're a reporter who's been assigned to cover this story. If you were trying
to write a story about the auto accident that you've just read about, you'd probably
want to have more information than you were given. Below is a list of questions that
might be useful in generating more information. Imagine the source you'll be
confronting has specific answers to each of the questions. We are interested in which
of these questions you see as most important in gathering the information necessary to
write your story. The questions you're choosing have all been suggested by working
journalists. Please remember that you are acting as a reporter gathering more
information about this man's automobile accident.





Go ahead and turn the page.













APPENDIX D
"BE OBJECTIVE" CONDITION INSTRUCTIONS

Please feel free at any point during this questionnaire to return to the previous page in
order to re-read the press release. As you go along remember to read all the
instructions very carefully.




ASSIGNMENT



You're a reporter who's been assigned to cover this story. If you were trying
to write a story about the auto accident that you've just read about, you'd probably
want to have more information than you were given. Below is a list of questions that
might be useful in generating more information. Imagine the source you'll be
confronting has specific answers to each of the questions. We are interested in which
of these questions you see as most important in gathering the information necessary to
write your story. The questions you're choosing have all been suggested by working
journalists. Please remember that you are acting as a reporter gathering more
information about this man's automobile accident.

Also remember that you, as a reporter, should be as accurate and objective as
possible when searching for more information. What we want is for you to be as
unbiased as possible in providing a fair and accurate description of this man's
automobile accident. That means collecting information that highlights all the
potential sides of an issue. You would then sift through the information to select and
use information that fairly portrays all those different sides. Most journalists and
journalism teachers refer to this process as objectivity. They often stress that all
reporters should strive to be as objective as possible. That's what we want you to do
in this assignment be objective.



Go ahead and turn the page.













APPENDIX E
"CONSIDER THE OPPOSITE" CONDITION INSTRUCTIONS

Please feel free at any point during this questionnaire to return to the previous page in
order to re-read the press release. As you go along remember to read all the
instructions very carefully.





ASSIGNMENT


You're a reporter who's been assigned to cover this story. If you were trying
to write a story about the auto accident that you've just read about, you'd probably
want to have more information than you were given. Below is a list of questions that
might be useful in generating more information. Imagine the source you'll be
confronting has specific answers to each of the questions. We are interested in which
of these questions you see as most important in gathering the information necessary to
write your story. The questions you're choosing have all been suggested by working
journalists. Please remember that you are acting as a reporter gathering more
information about this man's automobile accident.

Also remember that you, as a reporter, should always try to take all
viewpoints into consideration, even viewpoints that appear in conflict with stereotypes
about certain groups of people. However, considering all viewpoints is often difficult
because people tend to ignore information that doesn't agree with their own
stereotypes. When completing this questionnaire, keep in mind that elderly people
often have the same strengths and weaknesses as younger people. Ask yourself when
choosing each question whether you would have made the same choice if Smith had
been a much younger person. In other words, consider the opposite of your own
cultural stereotype of the elderly.


Go ahead and turn the page.













APPENDIX F
PRETEST





PRESS RELEASE



Gainesville Police Department
February 19, 1991



On Feb. 18, 1991, while driving to the grocery store in a suburban area, a

man named James Smith was involved in an automobile accident. The car that he

was driving collided with another moving vehicle in which there were two people.

Smith was uninjured. The people in the other car were injured and were taken to

Alachua General Hospital.










Please feel free to return to the previous page at any point to re-read the press
release.













We are interested in looking at age-related stereotypes. More specifically, we want to
find out if the age of Smith, the driver described in the press release, affects the kinds
of questions people will ask. To do that, we plan to tell participants in our study that
Smith is either a young man (19 years old) or an elderly man (84 years old). We'll
then ask our participants to look at a series of questions to ask about Smith.

However, before we can conduct our study, we need to pre-rate the questions
according to how stereotypic they are. That's where we need your help. Please rate
each of the 21 questions according to whether it concerns a topic that is usually
associated with stereotypes of the elderly, stereotypes of young people, or not
stereotypical of either age group (neutral). There should be some of all three types of
questions.

To indicate your rating of the question, circle the "X" on the line that corresponds
with your assessment of the question. The "X" directly below the words "Stereotypic
of Elderly" would indicate the question was very stereotypic of elderly people, the
"X" to its right indicate the question is moderately stereotypic of the elderly, while
the "X" to its right would indicate the question was only a little stereotypic of the
elderly. Use the same kind of logic when rating a question that is stereotypic of
young people. Circling the "X" directly below the word "Neutral" indicates the
question is neither stereotypic of the elderly or young people.

For example, if you were rating a question that read:

1. Was Smith adjusting his bifocals?

...you would probably rate that question as moderately stereotypic of the elderly
because we think of older people as wearing bifocals. Your answer then, might look
something like this...

Stereotypic Neutral Stereotypic of
of Elderly Young People
X------X----- -------X----- --- ---X











If, however, the question read:

1. Was Smith wearing a seatbelt?

...you would probably rate that question as neutral, wearing a seatbelt isn't
stereotypic of either elderly or young people...

Stereotypic Neutral Stereotypic of
of Elderly Young People
X------X-------X -------X----- --- ---X


In contrast, if your next question read:

1. Was Smith listening to loud rock music?

...you would might rate that question as very stereotypic of young people; most
elderly people don't listen to loud music, much less rock music...

Stereotypic Neutral Stereotypic of
of Elderly Young People
X------X-------X -------X----- --- ---X


Now, go ahead and rate each of the questions on the next three pages by circling the
appropriate rating.












LIST OF QUESTIONS

Pretest Pretest
Mean St. Dev.
1. Was Smith obeying traffic laws at the time?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X -----X-----X---X--- ------X 4.86 0.65

2. What was Smith's general physical condition?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X --X-------------------X ---X 2.71 0.84


3. What was Smith's general mental condition?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------------X---------X--- --- ------X 2.61 1.32

4. Was Smith driving under the influence of alcohol?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X------X--- X-----X--- -------X 5.95 1.02

5. Was Smith's eyesight impaired?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X --------X ------X---- ----------X 2.05 1.07

6. Was Smith's car in good repair?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X---------X-------X------ -----------X 4.38 0.97










7. Was Smith driving recklessly?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X---------X-----X-----X--- -------X 5.86 1.01

8. Was Smith driving too fast?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X--------X ------X -----X---X---------X -----X 6.43 0.60

9. Was Smith driving under the influence of illegal drugs?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X--- X--X-X--------------- -------- ----X 6.14 1.15

10. How quick are Smith's reactions?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X------ -----X-----X-----X----- ----X----X 1.67 0.73

11. Who had the right of way?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X -----X ----X--- -------X 3.95 0.59

12. What were the road conditions at the time?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X---------X------X------- ---X---X 3.90 0.44

13. Was Smith forced to swerve by something in the road?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X--------X------X-------X -------- -X 4.00 0.32










14. What was the spatial layout of the accident area?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X----------- -X--- --- ---XX------X 3.90 0.30

15. Was Smith driving too slowly?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X---------X-----X--- --X--- X 1.62 1.20

16. Was there a stop sign at the corer?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X---------X------X------- ---X---X 4.09 0.70

17. Was Smith combing his hair in the mirror at the time
of the accident?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X---------X------X-------- -X 6.05 0.67

18. Was the road slippery at the time of the accident?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X----- X---------X-----X------------- 3.95 0.80

19. How good was Smith's hearing?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X----- X---------X-----X------------- 1.57 0.75

20. Was Smith adjusting his radio at the time of the accident?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X--- --- -X-------------- --- ------ X 6.14 0.79








81

21. Was Smith confused about where he was?

Stereotypic Neutral Stereotypic of
of Elderly Young People
X---------X ------X-----X--- X----- -X 2.04 0.97













APPENDIX G
EXPERIMENT QUESTIONNAIRE

PLEASE DO NOT TURN THIS PAGE UNTIL YOU HAVE READ THESE
INSTRUCTIONS. DO NOT PUT YOUR NAME ANYWHERE ON THIS FORM.
YOUR RESPONSES ARE ANONYMOUS.



INFORMED CONSENT


You are being asked to participate in research that explores how people seek
information. Your participation is particularly important in helping us understand
your behavior as student journalists. To help us examine how student journalists
behave, we will be asking you to first read a press release then to fill out a brief
questionnaire. There are no correct answers to any of the questions. We are simply
interested in your best judgements.

If at any time while you are involved in this study you wish to stop, you may do so.
You are under no obligation to continue. In addition, you do not have to answer any
questions you do not want to answer.

To ensure the anonymity of your responses, please refrain from writing your name
anywhere on this questionnaire.

If you would like any further information about this study, please contact:




Paula S. Horvath-Neimeyer
2000 Weimer Hall




Thank you for your help. Please remove this page from the questionnaire and
continue on to the next page.














News Release



Police Care
Example-for research purposes only




On Feb. 18, 1991, while driving to the grocery store in a suburban area, an

elderly man named James Smith was involved in an automobile accident. The car

that he was driving collided with another moving vehicle in which there were two

people. The 74-year-old Smith was uninjured. The people in the other car were

injured and were taken to Alachua General Hospital. Smith, who is retired, was

driving a white 1984 Dodge.














Please feel free at any point during this questionnaire to return to the previous page in
order to re-read the press release. As you go along remember to read all the
instructions very carefully.




ASSIGNMENT



You're a reporter who's been assigned to cover this story. If you were trying
to write a story about the auto accident that you've just read about, you'd probably
want to have more information than you were given. Below is a list of questions that
might be useful in generating more information. Imagine the source you'll be
confronting has specific answers to each of the questions. We are interested in which
of these questions you see as most important in gathering the information necessary to
write your story. The questions you're choosing have all been suggested by working
journalists. Please remember that you are acting as a reporter gathering more
information about this man's automobile accident.


Go ahead and turn the page.












Please rank the questions below to indicate how important you think each is. That is,
if you had to write the story, which questions would you most want to ask. Place the
number "1" in front of the question from this group that you think would generate the
most important information for your story; place a number "2" in front of the
question that you think would generate the second most important information; and so
on.

IMPORTANT: Before making your decisions, read through the entire list of
questions carefully. Then begin ranking them from "1" for the most important
question to "15" for the least important. Please rank every question in the list.

LIST OF QUESTIONS

Was Smith obeying traffic laws at the time?

What was Smith's general physical condition?

Was Smith driving under the influence of alcohol?

Was Smith's eyesight impaired?

Was Smith driving under the influence of illegal drugs?

Was Smith driving recklessly?

Was Smith driving too fast?

How quick are Smith' reactions?

Who had the right of way?

What were the road conditions at the time?

Was Smith forced to swerve by something in the road?

What was the spatial layout of the accident area?

How good was Smith's hearing?

Was Smith adjusting his radio at the time of the accident?

Was Smith confused about where he was?









86

FROM THIS PAGE ON, PLEASE DO NOT RETURN TO RE-READ THE PRESS
RELEASE.

Finally, we'd like to know just a few things about you.

1. What is your age?

(Please write clearly)

Answer the questions below by placing an "X" in the correct space.

2. Indicate your gender.

Female

Male

3. What is your major?

Journalism

Public Relations

Advertising

Broadcasting

Other (specify)

4. Do you plan to work as a reporter after graduation?

Yes

No

Don't Know

5. When seeking information as a student reporter, do you use different information-gathering
techniques than those you might use in every-day life when you're not acting as a reporter?

Yes

No









87

When answering the questions below, place an "X" on the line beneath the phrase that most closely
approximates your belief.
Very Somewhat Neither Somewhat Very
Likely Likely Unlikely Unlikely

6. To what extent do you think it's
likely that Smith might have had poor
eyesight that contributed to the
cause of the accident?

7. To what extent do you think it's
likely that drinking of alcohol by
Smith might have contributed to the
cause of the accident?

8. To what extent do you think it's
likely that a cause of the accident
might be that Smith was adjusting his
radio?

9. To what extent do you think it's
likely that poor or slow reactions by
Smith might have contributed to the
cause of the accident?

10. To what extent do you think that
taking of illegal drugs by Smith
might have contributed to the cause
of the accident?

11. To what extent do you think it's
likely that Smith might have had poor
hearing that contributed to the cause
of the accident?

12. To what extent do you think it's
likely that fast driving by Smith
might have contributed to the cause of
the accident?

13. To what extent do you think it's
likely that a cause of the automobile
accident might be the physical frailty
of Smith?

14. To what extent do you think it's
likely that a cause of the accident
might be that Smith was confused by
where he was?









88

15. To what extent to you think that
reckless driving by Smith might have
contributed to the cause of the
accident?


16. Have you ever been told that "reporters should try to be objective when covering a story?"

Yes

No


17. You may have been instructed to do something special on the second page of this questionnaire.
Read the three statements below then place an "X" in the space beneath the phrase that most closely
describes what you remember. The phrases range from "Very Sure I Was Told" to do this to "Very
Sure I Was Not Told" to do this.

Very Sure Somewhat Unsure Somewhat Very Sure
Was Told Sure Told Sure Not Not Told

a) To "be objective"

b) To "consider the opposite"

c) Neither of the above


18. The press release you read gave you certain information about Smith. Using the same rating scale
as that used in the question above, put an "X" in the blank beneath the phrase that most closely
describes what information you remember being told and what information you were not told.

a) Smith was retired

b) Smith worked for a local
insurance company

c) Smith drove a white
1984 Dodge

d) Smith drove a red
1991 Porsche sportscar


19. If you had to hazard a guess, what do you think was the cause of the auto accident described in
this press release? Please write clearly.









89


20. How certain or uncertain are you about the cause of the accident that you just wrote down? Put an
"X" in the space opposite the answer that most closely approximates how you feel.

Very Certain

Somewhat Certain

Neither Certain Nor Uncertain

Somewhat Uncertain

Very Uncertain


21. How certain or uncertain are you that the cause of the accident was related to Smith's age? Put
an "X" in the space opposite the answer that most closely approximates how you feel.


Very Certain

Somewhat Certain

Neither Certain Nor Uncertain

Somewhat Uncertain

Very Uncertain

22. My impression of Smith was that he is (has): (Please put "X" on blank that corresponds to how
closely your feelings match the phrases on either side.)

Typical 74-year-old Unusual 74-year-old

Frail Strong

Good Driver Bad Driver

Confused Alert

Incompetent Competent

Good Vision Poor Vision

Physically Unstable Physically Stable

Insecure Secure












REFERENCES


Aboud, F. E. (1975). Seeking information about different
ethnic groups: The role of motivation and
confirmation. Journal of Applied Social Psychology,
5, 331-341.

Alexander, F. S. (1976). Stereotyping as a method of
exploitation in film. Black Scholar, 7, 26-29.

Applegate, J. L., & Sypher, H. E. (1983). A constructivist
outline. In W. B. Gudykunst (ed.), Intercultural
communication theory (pp. 63-78). Beverly Hills:
SAGE.

Aronoff, C. (1974). Old age in prime time. Journal of
Communication, 24, 56-60.

Axelrod, S., & Eisendorfer, C. (1961). Attitudes toward
old people: An empirical analysis of the stimulus-
group validity of the Tuckman-Lorge Questionnaire.
Journal of Gerontology, 16, 75-80.

Bartlett, F. C. (1932). Remembering: A study in
experimental and social psychology. Cambridge:
Cambridge University Press.

Bassok, M., & Trope, Y. (1983). People's strategies for
testing hypotheses about another's personality:
Confirmatory or diagnostic? Social Cognition, 2,
199-216.

Berger, C. R., & Calabrese, R. (1975). Some explorations
in initial interaction and beyond: Toward a
developmental theory of interpersonal communication.
Human Communication Research, 1, 99-112.

Berninger, V. W., & DeSoto, C. (1985). Cognitive
representation of personal stereotypes. European
Journal of Social Psychology, 15, 189-211.

Bodenhausen, G. V. (1988). Stereotypic biases in social
decision making and memory: Testing process models of
stereotype use. Journal of Personality and Social
Psychology, 55, 726-737.











Bodenhausen, G. V., & Lichtenstein, M. (1987). Social
stereotypes and information-processing strategies.
Journal of Personality and Social Psychology, 52,
871-880.

Bogart, L. (1984). The public's use and perception of
newspapers. Public Opinion Quarterly, 48, 709-719.

Bogdan, R. Biklen, D., Shapiro, A., & Spelkomman, D.
(1982). The disabled: Media's monster. Social
Policy, 13, 32-35.

Bramlett-Solomon, S., & Wilson, V. (1989). Images of the
elderly in Life and Ebony. Journalism Quarterly, 66,
185-188.

Brewer, M. B., Dull, V., & Lui, L. (1981). Perceptions of
the elderly: Stereotypes as prototypes. Journal of
Personality and Social Psychology, 41, 656-670.

Broussard, E. J., Blackmon, C. R., Blackwell, D. L., Smith,
D. W., & Hunt, S. (1980). News of aged and aging
in 10 metropolitan dailies. Journalism Quarterly,
57, 324-327.

Buckalew, J. K. (1969). A Q-analysis of television news
editors' decisions. Journalism Quarterly, 46,
135-137.

Busby, L. J. (1975). Sex-role research on the mass media.
Journal of Communication, 25, 107-131.

Cacioppo, J. T., & Petty, R. E. (1979). Effects of message
repetition and position on cognitive responses, recall,
and persuasion. Journal of Personality and Social
Psychology, 37, 97-109.

Cash, T. F., & Duncan, N. C. (1984). Physical
attractiveness stereotyping among Black American
college students. Journal of Social Psychology, 122,
71-77.

Carver, C. S., & de la Garza, N. H. (1984). Schema-guided
information search in stereotyping of the elderly.
Journal of Applied Social Psychology, 14, 69-81.

Cauthen, N. R., Robinson, I. E., & Kraus, H. H. (1971).
Stereotypes: A review of the literature. Journal
of Social Psychology, 84, 103-125.










Commission on Freedom of the Press. (1947). A free and
responsible press. Chicago: University of Chicago
Press.

Culley, J. D., & Bennett, R. (1976). Selling women,
selling blacks. Journal of Communication, 26, 160-174.

Darley, J. M., & Fazio, R. H. (1980). Expectancy
confirmation processes arising in the social
interaction sequence. American Psychologist,
35, 867-881.

Darley, J. M., & Gross, P. H. (1983). A hypothesis-
confirming bias in labeling effects. Journal of
Personality and Social Psychology, 44, 20-33.

Davis, J. (1982). Sex bias in eight newspapers.
Journalism Quarterly, 59, 456-460.

Devine, P. G., Hirt, E. R., & Gehrke, E. M. (1990).
Diagnostic and confirmation strategies in trait
hypothesis testing. Journal of Personality and Social
Psychology, 58, 952-963.

Dimmick, J. (1974). The gate-keeper: An uncertainty
theory. Journalism Monographs, 37.

Doelger, J. A., Hewes, D. E., & Graham, M. L. (1986).
Knowing when to "second guess": The mindful
analysis of messages. Human Communication Research,
12, 301-338.

Donohew, L., Tipton, L., & Haney, R. (1980). Analysis
of information-seeking strategies. Mass Communication
Review Yearbook, 1, 383-389.

Eason, D. L. (1988). On journalistic authority: The
Janet Cooke Scandal. In J.W. Carey (Ed.), Media,
myths, and narratives: Television and the press (pp.
205-227). Newbury Park, CA: Sage.

Ettema, J. S., & Glasser, T. L. (1988). Narrative form
and moral force: The realization of innocence
and guilt through investigative journalism.
Journal of Communication, 38(3), 8-27.

Festinger, L. (1957). A theory of cognitive dissonance.
Evanston, IL: Row, Peterson.

Fischhoff, B., & Beyth-Marom, R. (1983). Hypothesis
evaluation from a Bayesian perspective. Psychological
Review, 90, 239-260.










Fishman, M. (1980). Manufacturing the news. Austin:
University of Texas Press.

Freeman, H. R. (1987). Structure and content of gender
stereotypes: Effects of somatic appearance and trait
information. Psychology of Women Quarterly, 11,
59-68.

Gerbner, G., Gross, L., Signorielli, N., & Morgan, M.
(1980). Aging with television: Images on television
drama and conceptions of social reality. Journal of
Communication, 30, 37-47.

Gerbner, G., & Tannenbaum, P. H. (1961). Regulation of
mental illness content in motion pictures and
television. Gazette, 6, 366-385.

Goldstein, A. G., Chance, J. E., & Gilbert, B. (1984).
Facial stereotypes of good guys and bad guys: A
replication and extention. Bulletin of the
Psychonomic Society, 22, 549-552.

Greenberg, B., & Kahn, S. (1970). Blacks in Playboy
cartoons. Journalism Quarterly, 47, 557-560.

Hamilton, D. L. (1979). A cognitive-attributional analysis
of stereotyping. Advances in experimental psychology,
12, 53-85.

Hess, B. (1974). Stereotypes of the aged. Journal of
Communication, 24, 76-85.

Hewes, D. E., & Graham, M. L. (1989). Second-guessing
theory: Review and extension. Communication Yearbook,
12, 213-248.

Hewes, D. E., Graham, M. L., Doelger, J., & Pavitt, C.
(1985). "Second guessing": Message interpretation
in social networks. Human Communication Research,
11, 299-334.

Hewes, D. E., Graham, M. L., Monsour, M., & Doelger, J. A.
(1989). Cognition and social information-gathering
strategies: Reinterpretation assessment in second-
guessing. Human Communication Research, 16, 297-321.

Higgins, E. T., & Bargh, J. A. (1987). Social cognition
and social perception. Annual Review of
Psychology, 38, 369-425.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.8 - mvs