IMPLICIT AND EXPLICIT RETRIEVAL OF SPELLING IN YOUNG AND OLDER ADULTS By SARA J. MARGOLIN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005
Copyright 2005 by Sara J. Margolin
This is dedicated to my husband, Erik Margolin.
iv ACKNOWLEDGMENTS I thank my advisor, Dr. Lise Abrams, for her advice, guidance, and encouragement, the Cognition and Aging Lab for its help in scheduling participants and collecting data, and I thank my husband, Erik, for hi s love, support, and patience.
v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ix CHAPTER 1 INTRODUCTION........................................................................................................1 Spelling and Aging.......................................................................................................5 Spelling Theories..........................................................................................................9 Hypotheses..................................................................................................................14 2 METHOD...................................................................................................................17 Participants.................................................................................................................17 Materials.....................................................................................................................1 8 Design......................................................................................................................... 19 Procedure....................................................................................................................20 3 RESULTS...................................................................................................................24 Spelling Accuracy.......................................................................................................24 Accuracy for Recognition Tasks................................................................................27 Response Times for Recognition Tasks.....................................................................31 4 DISCUSSION.............................................................................................................35 APPENDIX A TARGET WORDS, FREQUENCIES, MISSPELLINGS AND SENTENCES........43 B FILLER WORDS AND FREQUENCIES..................................................................50 C PSEUDOWORDS......................................................................................................51
vi D POST EXPERIMENT QUESTIONNAIRE...............................................................52 LIST OF REFERENCES...................................................................................................53 BIOGRAPHICAL SKETCH.............................................................................................58
vii LIST OF TABLES Table page 2-1 Means and Standard Deviations of ParticipantsÂ’ Background Characteristics........23 3-1 Means and Standard Deviati ons for Spelling Accuracy (%)....................................25 3-2 Means and Standard Deviations for Recognition Task Accuracy (%).....................30 3-3 Means and Standard Deviations for Recognition Task Response Times.................32 A-1 Target Words, Frequencies, Misspellings, and Sentences.......................................43 B-1 Filler Words and Frequencies..................................................................................50 C-1 Pseudowords.............................................................................................................51
viii LIST OF FIGURES Figure page 1-1 Representation of Node Structure Theory................................................................16 3-1 Frequency x Age Interaction for Accuracy..............................................................26 3-2 Method of Response x Freque ncy Interaction for Accuracy....................................26 3-3 Method of Response x Type of Test Interaction for Young Adults.........................27 3-4 Method of Response x Type of Te st Interaction for Older Adults...........................27 3-5 Type of Stimulus x Age Inte raction for Recognition Accuracy...............................29 3-6 Frequency x Type of Stimulus Interaction for Recognition Accuracy....................31 3-7 Type of Test x Type of Stimulus In teraction for Young AdultsÂ’ Response Times.33 3-8 Type of Test x Type of Stimulus In teraction for Older AdultsÂ’ Response Times...34
ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science IMPLICIT AND EXPLICIT RETRIEVAL OF SPELLING IN YOUNG AND OLDER ADULTS By Sara J. Margolin May 2005 Chair: Lise Abrams Major Department: Psychology Prior research on memory and aging has documented a dissociation in retrieval from memory. Explicit retrieval of words, wh ich requires deliberate recollection, declines with age, but implicit retrieva l, memory without conscious ef fort, shows little or no age decline. The present experiment examined whether this dichotomy in older adultsÂ’ memory retrieval can also apply to the area of spelling retrieval, in which older adults have also exhibited declines when asked to explicitly retrieve co rrect spellings. Young and older adultsÂ’ knowledge a bout the spellings of highand low-frequency words was tested via two direct and two i ndirect spelling tests, where one test of each type involved recognition of spelling and one involved produc tion of spelling. Using explicit retrieval, the two tests were spelling recognition and spelling producti on. The spelling recognition task asked participants to decide whether visually presented words were correctly or incorrectly spelled. The spelli ng production task auditorily presented words and asked participants to write down the correct spelli ng of each word. Using implicit retrieval, the two tests were lexical decision and senten ce generation. For the lexical decision task,
x participants saw visually pres ented strings of letters and j udged whether or not the string was a real word. In the sentence generation task, participants heard an auditorily presented sentence and wrote down the sentence as quickly as possible. Results revealed that retrieval of spelling diffe red from retrieval of words in that age differences did not occur differentially for implicit versus explicit retrieval. Both age groups showed similar patterns, in which explicit retrieval of spe lling was greater than implicit retrieval, and recognition of spelling was greater than produc tion, suggesting that retrieval of spelling was less affected by age than retrieval of words from memory. The only age differences resulted from word frequency: Older adults were marginally less accurate in identifying or spelling high-frequency words relative to young adults, whereas no age differences occurred for low-frequency words. These re sults are consistent with Node Structure Theory, which proposes greater age declin es when new connections are required. Knowledge of spelling was learned many year s ago and therefore does not require new connections to be formed, allowing older a dults to recognize and produce spelling to the same degree as young adults. The relevance of word frequency to age differences also supports the theory's view that the frequency with which a word is accessed influences the strength of the connecti ons to spelling. The results indicate the importance of continuing to use language as we age to facilitate both sp elling and memory.
1 CHAPTER 1 INTRODUCTION Prior research on memory and aging has ex amined retrieval from memory with two methodologies: implicit and explicit retrieval. Te sts of implicit retrieval (indirect tests) do not use conscious effort to recall words from memory, whereas tests of explicit retrieval (direct tests) demand conscious recollec tion (Howard & Wiggs, 1993; Roediger, 1990). Both of these types of retrieval have been ex amined with respect to older adultsÂ’ memory (e.g., Fleischman & Gabrieli, 1998; Jelicic, Craik, & Moscovitch, 1996; LaVoie & Light, 1994), suggesting that explicit re trieval declines with age, whereas tests of implicit retrieval show much less age d ecline, if any. For example, af ter rating a list of words for pleasantness, Light and Singh (1987, Experime nt 2) found that older adults had poorer cued recall for these words; i.e., given a thre e-letter word stem, ol der adults were less able to use these stems to recall words they ha d rated previously. In contrast, they showed equivalent priming effects on word-stem comple tion; i.e., when asked to complete a word stem with the first word that came to mind, they produced a word from the prior list as often as young adults. The purpose of the presen t research was to examine whether this dichotomy in older adults' memory retrieva l can also apply to the area of spelling retrieval, where older adults have also exhi bited declines. However, spelling retrieval in older adults has been assessed only by dire ct tests to date (Abrams & Stanley, 2004; MacKay & Abrams, 1998; MacKay , Abra ms, & Pedroza, 1999; Stuart-Hamilton & Rabbitt, 1997).
2 It is important to note that older adultsÂ’ implicit retrie val of words is not always intact (e.g., Chiarello & Hoyer, 1988; Davis et al., 1990; Hultsch, Masson, & Small, 1991; Jelicic et al., 1996; Small, Hultsch, & Masson, 1995). For example, Jelicic et al. (1996) asked young and older adults to comple te several retrieval tasks after studying a list of words. Two indirect memory tasks, word fragment completion and category production, were used. In the word fragment co mpletion task, participants were asked to complete each word fragment (e.g., _ ia _ _ nd) with the first word that came to mind. In the category production task, participants were asked to generate eight items for each given semantic category (e.g., animals). Older adults were just as likely as young adults to complete word fragments with words studied previously, but were less likely to produce category examples that were on the studied list than young adults. LaVoie and Light (1994) conducted a meta-a nalysis of 39 implicit memory studies in aging and replicated Jelicic et al. ( 1996) with a dissociation in older adultsÂ’ performance on tests assessing implicit retrie val. Some studies found no age differences on indirect memory tests, such as wo rd fragment completion (e.g., Light, Singh, & Capps, 1986), word-stem completion (e.g., Li ght & Singh, 1987; Park & Shaw, 1992), lexical decision (e.g., Moscovitch, 1982), and word naming (e.g., Light & LaVoie, 1992), whereas other studies showed significant age differences on these same tests (e.g., Chiarello & Hoyer, 1988; Davis et al., 1990; Howard, 1988; Hu ltsch et al., 1991). Fleischman and Gabrieli (1998) proposed that these inconsistent findings of age differences among implicit retrieval studies ma y be explained by whet her or not the test requires recognition or production. Recognition involves the process of identification (Fleischman & Gabrieli, 1998) and requires fa miliarity with an item (Light & LaVoie,
3 1993), whereas production is the process of outpu t following retrieval of words and their phonology or orthography (Burke, MacKay, & James, 2000). Fleischman and Gabrieli (1998) suggested that older adul ts decline in their ability to produce items from memory, so when an indirect test relies on production, age differences will arise, contrary to other indirect memory tests that s how age equivalence. Neverthele ss, whereas older adults are sometimes able to implicitly retrieve info rmation as well as young adults, there is consistently a decline in older adultsÂ’ abil ity to explicitly retrieve information from memory (e.g., Fleischman & Gabrieli, 1998; Je licic et al., 1996; La Voie & Light, 1994). Given that explicit retrieval is generally higher than implicit retrieval (e.g., Chiarello & Hoyer, 1988; Roediger, 1990), it is important to test older adults under conditions that maximize their implicit retrieval from memory. Independent of implicit retrieval, MacK ay and Abrams (1996) and Burke et al. (2000) proposed more generally that there is an age-li nked asymmetry between older adultsÂ’ recognition and production from memo ry, where production is impaired more than recognition. For example, older adults were less accurate at recalling words and phrases learned during study than young adu lts, but were similar in accuracy in recognizing those items (Craik & McDowd, 1987). In addition, older adults were slower and less accurate then young adults in nami ng pictures, indicating more difficulty in retrieving and producing the phonology of words that were familiar to them (e.g., Bowles, Obler, & Poon, 1989; Mitchell, 1989). Studies of spontaneous speech (e.g., Heller, Dobbs, & Rule, 1992) and recollection of personal experiences (e.g., Ulatowska, Hayashi, Cannito, & Flemming, 1986) revealed that older adults used more pronouns and ambiguous speech than young adults. In contra st to these language production studies,
4 studies of language comprehension and recogni tion showed little or no decline with age, using tasks involving semantic priming a nd lexical decision (e.g., Burke & Harrold, 1988; Howard & Wiggs, 1993; Tainturier, Tremblay, & Lecours, 1989). For example, Tainturier et al. (1989) examined older adul tsÂ’ recognition using a lexical decision task and showed that young and older adults were equivalent in accurat ely deciding whether a string of letters was a real word. In the memory literature, another variab le that interacts with implicit/explicit retrieval and recognition/production is word frequency, which can be described as how often a word is encountered in a language ( by use of reading, speech, etc.). A survey of the literature concerning word frequenc y and its impact on various tasks involving language demonstrates that low-frequency words are more difficult to recognize and retrieve than high-frequency words on indir ect memory tests, including those using recognition, such as lexical decision (e.g., Balota & Ferrar o, 1996; Hino & Lupker, 2000; Tainturier et al., 1989), and those using pr oduction, such as word naming (e.g., Hino & Lupker, 2000; Spieler & Balota, 2000), hom ophone spelling (e.g., Davis et al., 1990; Gomez, 2002; Rose, Yesavage , Hill, & Bower, 1986), and word-stem completion (e.g., Gomez, 2002). With respect to direct memo ry tests, high-freque ncy word are better recalled than low-frequency words (e.g., Deese, 1960; Delosh & McDaniel, 1996; Krishnan & Shapiro, 1996; Tan & Ward, 2000; Ward, Woodward, Stevens, & Stinson, 2003), but low-frequency words are recognized more accurately than high-frequency words (e.g., Chalmers, Humphreys, & Denni s, 1997; Earhard, 1982; Kinsbourne, 1974). In terms of aging, older adultsÂ’ performa nce is more negatively affected than young adultsÂ’ by low-frequency words, but only wh en the indirect memory test involves a
5 production component. Spieler and Balota ( 2000) found that in a naming task, word frequency accounted for a signi ficant amount of variance in naming latencies, the time between showing the word and naming the word, such that low-frequency words were named more slowly. However, frequency accoun ted for more variance in older adultsÂ’ naming times. However, in lexical decision, a recognition task, Tainturier et al. (1989) found that both young and older adults were sl ower and less accurate in deciding if a letter string was a real word, when the word was low in frequency, and the size of the word frequency effect was equivalent for both age groups. No studies of interactions between aging and word frequency in explicit memory have been done to date. Spelling and Aging The implicit/explicit and recognition/produc tion distinctions described above can be applied to the examination of spelling in aging. To date, studies of spelling in older adults have been conducted only with direct methods, i.e., recognizing or producing the correct spelling of words. With respect to written production of spelling, the few studies that have been conducted suggest an age declin e in the ability to e xplicitly retrieve a wordÂ’s correct spelling from memory. Using auditorily presented words, MacKay and Abrams (1998) found that older (mean age 67.2 years) and oldest (mean age 77.04) adults retrieved and wrote down the correct spelli ng of these words less often than college students, and Stuart-Hamilton and Rabbitt (1997) found that adults in their 70s retrieved the correct spelling less often than adults in their 60s, who were less accurate than adults in their 50s. These findings suggest an overall decline in explicit spelling production with age. In contrast, previous resear ch has shown intact recogni tion of words as correctly spelled in older adults (Abrams & Stanley, 2004; MacKay et al., 1999). In their spelling
6 recognition tasks, participants were asked to indicate whether words briefly presented on a computer screen were spelled correctly or in correctly. In both studi es, older adults were equivalent to young adults in recognizing words as correctly spelled. Older adults' recognition of words as misspelled has shown less consensus. Whereas MacKay et al. (1999) found no age differences in detecting words as misspelled, Abrams and Stanley (2004) found that older adults, especially ol d-old adults, were less accurate than young adults in recognizing words as misspelled. Howe ver, Abrams and Stanley used very rapid presentation rates (e.g., 50 msec for young adul ts and 70 msec for older adults), which may have explained the age declines. Across al l of these studies, the results suggest that spelling recognition processes are relative ly intact in older adults under most circumstances. Furthermore, in both of the above studie s (Abrams & Stanley, 2004; MacKay et al., 1999), spelling recognition and production were examined within the same experiment. Thus, older adults wrote down the correct sp elling of words less of ten than young adults, even when they were as accurate as young adul ts in recognizing those words as correctly spelled. Together, these results on explicit spelling and aging suggest an asymmetry similar to memory: spelling recognition re mains intact, whereas production of those spellings declines with age. Burke et al. ( 2000) explained this asym metry in spelling by proposing that as a word is accessed, it become s activated. However, the longer it takes to produce the spelling of that word, the more time passes, and the more the activation of that word decays. By the time the spelling is completed, the activation may have completely decayed, resulting in inaccurate production of that spelling. This decay in activation is especially likely to occur fo r older adults, who are generally slower to
7 respond than young adults (Salthouse, 1996). Ther efore, older adults are more likely to have errors producing a word's correct spe lling because their slow er speed of production will enable the word's activation to decay before the entire word's spelling was produced. In contrast, spelling recognition does not have this problem of decay because recognition of a word as correctly or in correctly spelled happens immedi ately. Since there is no time for decay of activation, older adultsÂ’ slowi ng does not negatively affect their recognition of spelling. In all of the tests of spelling describe d above, spelling was only measured using explicit retrieval, i.e., spel ling was recognized and/or retr ieved through deliberate effort to access spelling. For example, participants were either specifically asked to spell a word or decide if a word was spelled correctly. In contrast, implicit retrieval of spelling in older adults has never been examined. If indi rect memory tests are representative of all implicit retrieval tests, then older adults ma y exhibit little or no decline in implicit retrieval of spelling, relative to explicit retrieval. However, older adultsÂ’ implicit retrieval of spelling may decline when producti on is used, relativ e to recognition. Similar to indirect memory tests, indirect tests of spelling do not ask participants to consciously recall a wordÂ’s spelling, but knowledge of spelling can be assessed indirectly. For example, Burt and Tate (2002) used a lexical decision task as an indirect test of spelling recognition, showing that people took longer to respond and were less accurate in classifying words that they had misspelled in a prior written spelling test. Their explanation was that participants we re implicitly checking for spelling when making lexical decisions and therefore respond ed more slowly and less accurately for words they could not spell prev iously. If lexical decision tasks are using an implicit spell-
8 checking process, then the results of Taintu rier et al. (1989), w ho showed that older adultsÂ’ accuracy was similar to young adults on a lexical decision task, suggests that implicit retrieval of spelli ng via recognition may remain intact in old age. In contrast, implicit production of spelling has not been st udied, one of the goals of the present research. The methodology for implicit spelling production in the present experiment presented sentences auditorily and asked participants to write down those sentences as quickly as possible, which did not allow participants to have time to focus on their spelling. Even though spelling retrieval is indirect, it is possibl e that older adults will show deficits in spel ling accuracy compared to young adults because of the production component. However, if age differe nces occur in implic it spelling production, they are expected to be smaller than age di fferences in explicit sp elling production tasks. Similar to memory, it seems likely that word frequency will impact older adultsÂ’ spelling, and this hypothesis has been supported in the literature, but only using direct tests of spelling production. Stuart-Hamilton a nd Rabbitt (1997) examined older adultsÂ’ spelling production of participan ts in three age groups (people in their 50s, people in their 60s, and people in their 70s). Participants co mpleted a spelling test, in which they wrote down the spelling of auditorily presented word s, and were given 15 seconds to complete each spelling. Results showed a decline in spelling production as age increased, and the decline was specific to low-frequency words; the spelling of high-frequency words did not decline with age. MacKay and Abrams (1998) also examin ed spelling production in older adults. Young (ages 17-23), young-old (ages 60-71), and old-old (73-88) listened to words read aloud and wrote down the spelling of each wor d. The old-old adults exhibited a spelling
9 production deficit for high-frequency words re lative to young adults, but there were no age differences in spelling low-frequency wo rds. The conflicting results for spelling of low-frequency words presented by Stuart-H amilton and Rabbitt (1997) and MacKay and Abrams (1998) can be explained with what MacKay and Abrams (1998) call the Â“familiarity hypothesis.Â” Since older adults had larger vocabularies than young adults, they were more familiar with the low-freque ncy words than were the young adults tested in MacKay and Abrams (1998). Therefore, th e young adults were less able to spell lowfrequency words because of their lesser fam iliarity with these words in general. In contrast, older adultsÂ’ reduced accuracy in spelling low-frequency words relative to highfrequency words was caused by a decline in spelling production ability. Thus, the lack of age differences found in MacKay and Abrams (1998) may be an artifact of the wordsÂ’ familiarity, not intact spelling abilities in olde r adults. In any case, the appearance of age differences in direct spelling production for both highand low-frequency words under some circumstances suggests the importance of word frequency in spelling retrieval for older adults. In contrast to these studies of explicit production of spelling in aging, no studies designed to assess the role of word fr equency in older adultsÂ’ implicit spelling recognition (although see Tainturier et al., 1989 ), implicit spelling production, or explicit spelling recognition have been done to date. Spelling Theories Before we attempt to make predictions a bout how older adults would perform using implicit and explicit retrieval for spelling in recognition and production tasks, we must first examine theories that attempt to explain how we produce and recognize spelling. There are extensive theories concerning spelling and language research in general, but
10 only the most applicable with respect to ag ingÂ’s relationship to spelling and memory are described here. The first of these theories, called Node Structure Theory (NST) was developed by MacKay (1987), with the representation of spelling incorporated into the theory in MacKay and Abrams (1998) and Burke et al. (2000). NST is a connectionist model where words, their sounds, their spellings, and thei r meanings are stored in nodes that are interconnected at multiple levels (see Figure 1-1). With respect to spelling, a wordÂ’s spelling is stored in orth ographic nodes, which are connected to a wordÂ’s sounds (phonological nodes) as well as the entire wo rd itself (lexical node), under some circumstances. For example, the word Â“fluor ideÂ” is connected to e ach of the sounds and letters contained in that word (i .e., Â“f,Â” Â“l,Â” Â“u,Â” etc.). When a word is presented in some way (e.g., visually or auditor ily), its correspondi ng lexical node becomes activated, which then spreads node priming across its connections to other nodes. NST also explains that the connections between a word and its ort hography can become weak ened, especially if a word is low in frequency or not recently used. These weakened connections cause less accurate retrieval and slower recogn ition of the spelling of a word. In an extension of NST, MacKay and Bu rke (1990) explained the changes that occur with age via the Transmission Deficit Hypothesis (TDH). As people get older, the connections between all nodes w eaken in general. TDH predic ts that the spelling of lowfrequency words will be more difficult to retrieve or recognize, especially for older adults, because these connections are weaken ed both by low frequency and by aging. For example, a low-frequency word like Â“fluorideÂ” is difficult to spell because the word is not accessed very often, either through writing, reading, or speech, which allows the
11 connections to its orthographic nodes (spelling) to weaken. In contrast, a high-frequency word like Â“hospitalÂ” is used on a regular basis, which strengthens the connections to its spelling. In addition, TDH predicts greater age di fferences for spelling production tasks, relative to recognition tasks. If the connections between a word and its orthography are weakened, difficulty in producing a spelling w ould result because a wordÂ’s spelling is produced sequentially via top-down priming, and only a single connection can be used to retrieve each letter. Thus, if any of t hose single connections are weakened, the orthographic node(s) will not receive enough prim ing to enable activation and retrieval of the letter. In contrast, if a word is pr esented for recognition of its spelling, the orthographic nodes are activated by the presen tation of the written word, all of which send bottom-up priming to the connected lexi cal node. There would be fewer age-related declines in spelling recogniti on because any one weakened connection can be offset by other stronger connections. In terms of implicit and e xplicit retrieval from memory, MacKay and Burke (1990) explain how NST and TDH would account for la rge age differences in explicit retrieval but small age differences in implicit retrieva l. In NST and TDH, these different patterns arise because direct memory tests generally require new learning, i.e., learning a list of words, where new connections need to be formed between the words studied and the context in which they were studied. New l earning results in newly formed and weak connections between the nodes, pa rticularly for older adults. In contrast, indirect memory tests require less, if any, ne w learning, which results in smaller or no age differences. However, an implicit retrieval task that requires retrieval of many new connections, e.g.,
12 between unrelated words, will result in age differences. In terms of spelling, the same argument would not apply, as spelling retrie val always relies on preexisting connections, regardless of whether the retrieval is implic it or explicit. No new connections are made when we are asked to spell a particular word. Because wordsÂ’ spelling are retrieved through previously formed connections, NST and TDH predict similar findings for older adults on both direct and indirect tests of spelling retrieval. Another theory concerning spelling produc tion is the Dual Route Model (Barry & Seymour, 1988). There are two processes that wo rk to produce spelling: a lexical process and a nonlexical process. In the lexical process, spelli ngs of words are retrieved holistically from a storage base of words, i.e ., spelling is directly retrieved from memory. The nonlexical process functions by assembli ng the spellings of words from sound-tospelling conversions. In other words, the wo rds are spelled by following rules that are stored in memory. It is important to note that these processes are independent of each other, one of the key characteristics in di stinguishing the Dual R oute Model from NST, where words and rules to spell are in terconnected through the same nodes. This nonlexical route works well for Â“regul arlyÂ” spelled words (i.e., words that are spelled the way they sound), but would cause phonologically plausible misspellings for Â“irregularlyÂ” spelled words (i.e., words that ar e spelled different from how they sound) if the nonlexical process was used exclusively (e.g., Barry, 1994). In addition to the idea of regularly and irregularly spelled words, Ba rry and Seymour (1988) also postulate that words have a contingency, which is the fre quency within the Eng lish language that a wordÂ’s spelling patterns represent partic ular vowel sounds. A high-contingency word would have a spelling pattern that repres ents vowel sounds in many words, e.g., the
13 spelling Â“eaÂ” for the sound / /; a low-contingency word woul d have a spelling pattern that represents vowel sounds in a few words, e.g., the spelling Â“oaÂ” for the sound / / (Barry, 1994). A high-contingency word could be regu lar (e.g., leaf) or irregular (e.g., wool); a low-contingency word could al so be regular (e.g., soap) or irregular (e.g., brooch). Regularly-spelled words can be easi ly spelled through spelling-to-sound conversions, whereas irregularly-spelled word s can be correctly spelled only through retrieval of that specific word from memory. By this notion, any word, regular or irregular, could be spelled with ease as long as there is memory for those spelling rules/specific spellings. In terms of fre quency, high-frequency words would be more easily spelled than low-frequency words because they are encountered more often and would be more easily retrieved from memory, whether they are highor low-contingency. In addition, a Dual Route theory for r eading and word recognition exists, which could be used to understand spelling rec ognition. Shankweiler (1999) proposes two routes by which a person would see and recogn ize or read a word. The first route is a visual to semantic route, where a person s ees a word and immediately recognizes it as well as its meaning. The second route, the prim ary route for children learning to read, is the phonologically analytic route, where a person sees a word, transcri bes it to its sounds, and then accesses its meaning, thus identifyi ng the word. For an experienced reader, the visual to semantic route would be used to recognize very fam iliar (high-frequency) words, but the phonologically analytic route w ould be used to recognize newer and less familiar (low-frequency) words. By this theory, low-frequency words are recognized slowly because they are tran slated to their phonology before they are recognized. Highfrequency words are recognized quickly becau se they are immediately accessed via their
14 visual presentation. The more quickly a word is recognized, the soone r its spelling can be assessed to determine whether or not it is spelled correctly. Hypotheses These theories as well as pr evious research on spelling, memory, and aging lead to hypotheses about older and young adultsÂ’ ability to spell. The variables used in this experiment, i.e., implicit versus explic it retrieval of spelling, recognition versus production of spelling, and highversus low-fr equency, have never been studied together in relation to spelling in e ither young or older adults. NST and TDH predict that older adults wi ll have poorer accuracy than young adults on tasks that are dependent on weakened connec tions, especially thos e that are vulnerable to weakening from multiple sources or that cannot be offset by other, stronger connections. In sum, older adults are expect ed to show poorer accu racy than young adults on tests of spelling production, with this dec line exacerbated for low-frequency words. In contrast, spelling performance is not expected to decline in recognition tasks for either highor low-frequency words. As stated earlier, NST a nd TDH predict similar findings for older adults on both direct and in direct tests of spelling retrieval. The Dual Route Model can make predictions about spelling in general but does not specifically address spelling in older adults, i.e., when older adultsÂ’ spelling would or would not exhibit declines. Th e Dual Route Model predicts that the spelling of highfrequency words will be more easily recogni zed and produced than the spelling of lowfrequency words. Furthermore, since implicit retrieval does not necessarily decline with age according to the memory literature, implicit retrieval of spelling or spelling rules that are stored in memory would remain intact, whereas explicit retrieval of spelling or spelling rules would show age declines. In terms of age differen ces in recognition of
15 spelling, the Dual Route Model of reading w ould predict that sin ce both young and older adults are experienced readers, simila r accuracy should be observed in spelling recognition for the age groups. In terms of sp elling production, the Du al Route Model of spelling would predict age differences only if word regularity and contingency, the two variables postulated to influence spelli ng production, affect young and older adults differently. A review of the literature concerning spelling has al ready demonstrated that explicit recognition remains intact, whereas e xplicit production of spelling declines with age (Abrams & Stanley, 2004; MacKay & Ab rams, 1998; MacKay et al., 1999; StuartHamilton & Rabbitt, 1997), although the impact of word frequency on production is less clear. Whether this pattern holds for implicit sp elling retrieval, or wh ether it changes for recognition versus production tasks, will allo w us to assess whether explicit methods of retrieval, rather than produc tion per se, are a greater determ inant of age-related declines in spelling.
16 Figure 1-1. A representation of Node Struct ure Theory. The dashed lines represent weakened connections. This occurs for lowfrequency words such as Â“fluorideÂ” (s hown), but does not occur for high-freque ncy words such as Â“hospitalÂ” (not shown). The arrows show that recognition spreads node priming in a bottom-up manner, and production spreads in a top-down direction. Semantic System Phonological System Orthographic System Propositional Nodes Lexical Nodes Syllable Nodes Letter Nodes /Â’flor/ /Id/ /fl/ /o/ /I/ /d/ f l u o r i d e fluoride chemical in toothpaste /r/ prevents cavities Quasi-irregular Phonological Nodes
17 CHAPTER 2 METHOD Participants Sixty-four younger and 64 older adults pa rticipated in this experiment. The young adults ranged from 17 to 24 years ( M = 19.58, SD = 1.59, 21 males and 43 females), were recruited from general psychology and sensory processes courses, a nd received partial course credit for their participation. The older adults were selected from the University of Florida Cognition and Aging Laboratory pa rticipant pool, a group of 600 communitydwelling older adults who were recruited from churches, clubs, libraries, the University of Florida Alumni Association, as well as other laboratorie s in the Psychology Department. Their ages ranged from 61 to 91 years ( M = 75.32, SD = 6.49, 30 males and 34 females), and these participants received $8 per hour as compensation for their participation. All participants were fluent speakers of American English and reported normal or corrected-to-normal vision and hearing. Young and older participants were tested on several measures of cognitive ability, including a 25-item multiple-choice vocabular y test, and forward and backward digit span tests to measure working memory capacit y, where the span record ed is the greatest number of digits the participant was able to repeat back (either forward or backward). In addition, older participants completed a test of mental status known as the Mini Mental State Examination (Folstein, Folstein, & McHugh, 1972), a 30-point questionnaire surveying older adultÂ’s orientation, atte ntion and calculation, recall, and language processing abilities. Participants were also aske d to give ratings of their health, compared
18 to other people their age, as well as ratings of their spelling traini ng and spelling abilities. Other background demographics, such as age and years of education, hours each day spent reading, writing, watching television, and doing crossword puzzles, were also recorded. T-tests revealed that the older adults had greater vocabulary scores than young adults, t (128) = 10.15, p < .001, more years of education, t (128) = 6.18, p < .001, higher ratings of spelling training, t (128) = 3.02, p < .003, more time watching television, t (128) = 3.99, p < .001, and more time doing crossword puzzles, t (128) = 3.29, p < .001. Young adults reported spending more tim e writing than older adults, t (128) = -4.61, p < .001. However, young and older adults did not differ on raw forward digit span, t (128) = -.94, p > .35, raw backward digit span, t (128) = 1.09, p > .28, or health rating t (128) = .14, p > .89. Table 2-1 presents the means and standard deviations for the above measures. Materials One hundred twenty-eight target words were used in this experiment (32 in each of four tasks, discussed below). Half of them were categorized as low frequency using Kucera and FrancisÂ’s (1967) word frequency norms ( M = 3.36, SD = 2.75, Range = 0-9), whereas the other half were designated as high frequency ( M = 76.28, SD = 37.40, Range = 40-191). A misspelling was created for each targ et word using 41misspellings that were tested/produced in previous research st udies (MacKay & Abrams, 1998, Katz & Frost, 2001). Seventy-nine misspelled words were al so taken from WebsterÂ’s Bad SpellerÂ’s Dictionary (1992). The remaining 8 misspellings were generated by the experimenter. All the misspellings were phonologically plausible and did not change more than two letters of the original spelling of the word (see Appendix A). In addition, 20 correctly-spelled filler words (see Appendix B) and 20 pseudowords (taken from Seidenberg, Plaut, Petersen, McClelland, and McRae, 1994; see Appendix C), which were nonwords
19 containing plausible spellings of words in or thographically legal combinations, were used in tasks where misspellings were presented. Participants engaged in four separate spelling-related tasks: lexical decision, spelling recognition, spelling pr oduction, and sentence genera tion. The lexical decision and spelling recognition tasks tested spelli ng via recognition, where participants made judgements about words and pseudowords based on their visual representations. Each of these tasks presented 52 words, including 16 correctly-spelled targets, 16 misspelled targets, 10 pseudowords, and 10 fillers. Sp elling production and se ntence generation assessed spelling via production, which requi red participants to produce spelling by writing down auditorily presented words. Each of these tasks presented 32 target words, either embedded in a sentence (sentence ge neration) or alone (spelling production). Each of these four tests ca n also be classified into one of two types of spelling retrieval, implicit or explicit. The lexical decision and sentence generation tasks were indirect tests of spelling because they did not require particip ants to consciously retrieve spelling; instead, the focus was on other aspect s of the presented words, such as lexicon membership and placement in a sentence. In contrast, the spelling recognition and spelling production test were direct tests of spelling b ecause participants had to intentionally attend to spelling to perform the tasks correctly. Design A 2x2x2x2 factorial design was used, with age group (young and older), word frequency (high and low), method of respons e (recognition and producti on), and type of test (direct and indirect) as factors. All factors, except age group, were within subjects. The dependent variables were mean decisi on times and percent correct accuracy to
20 correctly classify the word s in the recognition tasks. In the production tasks, the percentage of correctly spelled words was the dependent measure. Procedure Upon beginning the experiment, participan ts were asked to read and sign an informed consent form. Then, participants were given instructions for the lexical decision task and the spelling generation task, with or der of these tasks c ounterbalanced across participants. These indirect tests were always completed first so that the experimentÂ’s focus on spelling would remain hidden to partic ipants. In the lexical decision task, words, misspelled words, pseudowords, or fillers we re presented one at a time on the computer screen for a maximum of 4000 msec, or until the participant responded, whichever came first. Participants were asked to decide as quickly as possible whethe r the string of letters was a real word. The participants res ponded by pressing a key marked Â“yÂ” on the keyboard if they thought the word was real and a key marked Â“nÂ” if they thought the word was not real. After giving their response, the next string of letters immediately came onto the screen. If they gave no response af ter 4000 msec, the word disappeared, and they received a message indicating th at they lost their opportunity to respond to that stimulus and had to press the space bar to continue. If they pressed an incorrect key, the computer produced a beep, and participants were reminde d again of the correct keys to make their responses. After four practice trials consisting of one of each type of stimulus (correctly spelled word, misspelled word, filler word, and pseudoword), participants began the experimental trials. In the sentence generation task, sentences were presented audito rily via computer one at a time, and participants were asked to write the sentence as quickly and neatly as possible. They were told that printing wa s preferred for legibili ty; however, many older
21 adults were uncomfortable printing and th erefore used cursive. The experimenter observed the participantsÂ’ writi ng to make sure it was legible. If it was not legible, the experimenter asked the participant to clarify what was written. Participants had the opportunity to hear each sentence as many ti mes as they wanted before they began writing by clicking a button on th e computer screen, but on ce they began writing, they had only 20 seconds to complete each sentence . After 20 seconds, the computer produced a beep, and the participants were requ ired to stop writing. The time limit was implemented to restrict participantsÂ’ abil ity to attend to spelling while writing the sentences. Participants were given three pract ice trials before data collection began to ensure that they understood the instructi ons. After completing both implicit tests, a questionnaire was administered that assessed spelling strategies used during the indirect tests (see Appendix D). Following the indirect tests a nd the questionnaire, particip ants completed the direct tests, spelling recognition a nd spelling production, with or der counterbalanced across participants. Similar to lexical decision, the spelling recognition task presented words one at a time for a maximum of 4000 msec, or until the participant ma de their response, whichever came first. Participants instead determined whether or not the word was spelled correctly, indicated by pr essing keys marked either Â“y Â” or Â“nÂ” on the keyboard as quickly as possible. After pressing either key, the next string of letters immediately came onto the screen. If no response was given afte r 4000 msec, the stimulus disappeared, and the participants received a message indicating that they lost their opportunity to respond to that stimulus and would have to press th e space bar to continue . If they pressed an incorrect key, the computer produced a b eep, and the experimenter reminded the
22 participant of the correct keys to make their responses. Participants completed four practice trials with one of each type of s timulus to ensure their understanding of the instructions before data collection began. The spelling production task required partic ipants to write as neatly as possible words that were presented aud itorily one at a time. Particip ants were able to listen to these words as many times as they wanted by clicking on a button on the computer screen both before and while they were writing th e spelling of each word. No time limit was instituted, although most participants took le ss than 30 seconds on average to write each word. If participants wanted to change the sp elling of a word they had written down, they were asked not to cross a nything out and to circle their final spelling. When the participants were satisfied w ith their spelling, they were in structed to press the Â“enterÂ” key on the keyboard to hear the next word. Th ree practice trials were given before data collection began. After the participants completed all four ta sks, they were presented with a written debriefing of the experiment. This debriefing was discussed with each participant, and the experimenter answered any questi ons that the part icipants had.
23 Table 2-1. Means and Standard Deviations of ParticipantsÂ’ Bac kground Characteristics Young Adults Older Adults Mean Standard Mean Standard Deviation Deviation Vocabulary (out of 25)* 15.26 3.52 20.82 2.66 Years of Education 13.87 1.39 16.69 3.41 Forward Digit Span 7.32 1.09 7.14 1.14 Backward Digit Span 4.98 1.19 5.22 1.23 Health Rating (out of 10) 7.72 1.52 7.76 1.69 MMSE (out of 30) n/a n/a 28.45 1.75 Hours Spent Writing* 2.46 1.83 1.28 0.96 Hours Spent Reading 2.88 1.76 2.79 1.48 Hours Watching T.V.* 1.92 1.37 3.05 1.81 Hours Doing Crosswords* 0.17 0.45 0.62 1.00 Spelling Training* (out of 10) 5.58 2.11 6.83 2.58 Spelling Ability Now (out of 10) 5.81 1.76 5.56 2.26 Spelling Ability at Age 20 (out of 10) n/a n/a 6.65 2.25 Note . * p < .05
24 CHAPTER 3 RESULTS Spelling Accuracy A 2 (Age Group) x 2 (Word Frequency) x 2 (Method of Response) x 2 (Type of Test) repeated-measures ANOVA was conduc ted on percent accuracy of correctly judging (in recognition) or retrieving (in production) a word Â’s spelling (see Table 3-1 for means and standard deviations). There we re main effects of method of response, F (1, 122) = 39.02, MSE = 2.90, p < .001, type of test, F (1, 122) = 31.24, MSE = 1.17, p < .001, and word frequency, F (1, 122) = 465.36, MSE = 1.39, p < .001, where these effects respectively showed higher accuracy for recogni tion tasks relative to production tasks, for explicit retrieval relative to implicit retrieva l, and for high-frequency words relative to low-frequency words. The main effect of age group was not significant, p > .689. These main effects, however, were mode rated by several two-way interactions. There was a significant Word Frequency x Age Group interaction (see Figure 3-1), F (1, 122) = 6.41, MSE = 1.39, p < .013, with young adults spelling high-frequency words marginally more accurately than older adults, p < .093, unlike low-frequency words, which were spelled equiva lently by both age groups, p > .608. For both age groups, there was a significant effect of word frequency su ch that high-frequency words were spelled more accurately than low-frequency words, with a larger effect for young adults, p < .001, than for older adults, p < .001.
25 Table 3-1. Means and Standard Devi ations for Spelling Accuracy (%). Age Type of Type of Word Mean Standard Group Response Test Frequency Deviation Older Recognition Indirect Low 76.4 12.9 High 83.7 13.7 Direct Low 80.3 14.6 High 85.8 13.9 Production Indirect Low 59.6 22.0 High 83.6 14.2 Direct Low 66.0 22.1 High 86.4 14.3 Young Recognition Indirect Low 73.4 10.0 High 84.1 9.7 Direct Low 78.7 9.8 High 89.0 8.6 Production Indirect Low 61.2 20.8 High 87.3 10.5 Direct Low 64.3 18.7 High 89.3 10.7 There was also a significant Method of Response x Word Frequency interaction (see Figure 3-2), F (1, 123) = 109.07, MSE = 1.35, p < .001, such that there was greater accuracy on recognition tasks re lative to production tasks for low-frequency words, p < .001, but there was no difference in method of response for high-frequency words, p > .384. The effect of word frequency, where accuracy was higher for high-frequency words, was significant for both types of tasks, p s < .001, but was larger for production tasks than for recognition tasks. The highest interaction revealed by th is ANOVA was a marginally significant Method of Response x Type of Test x Age Gr oup interaction (see Figu re 3-3 and Figure 3-4), F (1, 122) = 3.23, MSE = 9.07, p < .075. Follow-up tests on this interaction showed that while the Method of Response x Type of Test interaction was si gnificant for both age
26 groups, p s < .001, the nature of this interaction differed for young and older adults. For young adults, accuracy was greater for recogniti on tasks than production tasks for both types of test, p s < .022, with a larger method of re sponse effect for direct tests. Frequency x Age Group Interaction for Accuracy0 20 40 60 80 100 Low FrequencyHigh Frequency Age GroupAccuracy (%) Old Young Figure 3-1. Spelling accuracy as a func tion of age group and word frequency. Method of Response x Frequency Interaction for Accuracy0 20 40 60 80 100 Low FrequencyHigh Frequency FrequencyAccuracy (%) Recognition Production Figure 3-2. Spelling accuracy as a functi on of method of response (recognition or production) and word frequency. Alternatively, young adultsÂ’ accuracy was greater for the direct te sts than indirect tests on the recognition tasks, p < .001, but there was no differe nce in type of test on the production tasks, p > .108. Older adults showed a di fferent interaction: accuracy was higher for the direct tests relative to i ndirect tests on both th e recognition tasks, p < .009, and the production tasks, p < .002, although this effect was greater for the production tasks. Similar to young adults, older adults exhibited greater accuracy for recognition
27 tasks than for production tasks wh en either implicit retrieval, p < .001, or explicit retrieval, p < .001, was used, but with a larger eff ect when using implicit retrieval. No other interactions were significant, p s > .15. Figure 3-3. Young adultsÂ’ spelling accuracy as a function of method of response (recognition or production) and type of test (direct or indirect). Figure 3-4. Older adultsÂ’ spelling accuracy as a function of method of response (recognition or production) and type of test (direct or indirect). Accuracy for Recognition Tasks Because the recognition tasks involved multip le types of stimuli (i.e., correctly spelled words, misspelled words, and pseudow ords), a 2 (Age Group) x 2 (Type of Test) x 3 (Type of Stimulus) ANOVA was conducted to examine differences in recognition Method of Response x Type of Test Interaction for Accuracy for Young Adults 45 55 65 75 85 95 RecognitionProduction Method of ResponseAccuracy (%) Indirect Direct Method of Response x Type of Testl Interaction for Accuracy for Older Adults 45 55 65 75 85 95 RecognitionProduction Method of ResponseAccuracy (%) Indirect Direct
28 accuracy1 between correctly spelled words, mi sspelled words, and pseudowords (see Table 3-2 for means and standard deviations ). An accurate response in lexical decision and spelling recognition meant that the participants said Â“yesÂ” to a correctly spelled word and Â“noÂ” to a misspelled word or a pseudoword. Because type of stimulus is the only new variable in this ANOVA, only effects relevant to this variable were tested and reported here. This ANOVA revealed a T ype of Stimulus main effect, F (2, 244) = 208.71, MSE = 3.10, p < .001, where judgments about correctly sp elled words were more accurate than judgments about pseudowords, which were more accurate than judgments about misspelled words, p s < .002. A Type of Stimulus x Age Group interaction (see Figure 3-5), F (2, 244) = 5.21, MSE = 3.10, p < .006, was also found. Follow-up tests showed no significant age differences for correctly spelled or misspelle d words, but a significant age difference was found for pseudowords, p < .001, where older adults were more accurate in recognizing pseudowords as nonwords or as being misspelled, relativ e to young adults. Significant effects of type of stimulus were found for both young and older adults, p s < .001, where correctly spelled words had the highest accuracy followe d by pseudowords and then by misspelled words, all of which were si gnificantly different from one another, p s < .002. Type of stimulus did not signifi cantly interact with type of te st, either with or without age group, p > .122, demonstrating that the effects fo r type of stimulus were similar whether the task was lexical decision or spelling recognition. 1 Recognition via lexical decision refers to identifying a letter string as a real word, whereas recognition via spelling recognition mean s to identify whether or not a word was correctly spelled.
29 Figure 3-5. Recognition accuracy as a function of type of stimulus (correct, misspelled, and pseudoword) and age. The next analyses were de signed to examine whether wo rd frequency differentially affected recognition of correctly spelled and misspelled words. Frequency was not able to be included in the previous analyses because pseudowords did not have a word frequency and therefore could not be categorized into groups by frequency. A 2 (Age Group) x 2 (Word Frequency) x 2 (Type of Test) x 2 (Type of Stimulus ) ANOVA was conducted on percentage of recognition (see Table 3-2 for means and standard deviations). There was a significant Word Frequency x T ype of Stimulus interaction (see Figure 3-6), F (1, 122) = 9.78, MSE = 1.59, p < .002, where correctly spelled words were recognized more accurately than misspelled words for both highand lowfrequency words, p s < .001, but this difference was more pronounced for low-frequency words. There was greater accuracy for high-frequenc y words relative to low-frequency words for both correctly spelled and misspelled words, p s < .001. The other relevant interaction, Age Group x Word Frequency x Type of Stimulus, was not significant, p > .265. Type of Stimulus x Age Group Interaction for Recognition Accuracy0 20 40 60 80 100 Correctly Spelled MisspelledPseudowords Type of StimulusAccuracy (%) Old Young
30 Table 3-2. Means and Standard Deviations for Recognition Task Accuracy (%). Age Word Type of Type of Mean Standard Group Frequency Test Spelling/ Deviation Stimulus Older Low Indirect Correct 86.9 13.9 Misspelled 56.9 26.7 Direct Correct 92.2 13.9 Misspelled 59.1 24 High Indirect Correct 95.3 12.3 Misspelled 64.1 28.8 Direct Correct 95 13.4 Misspelled 69.6 24.4 N/A Indirect Pseudoword 91.1 15.6 Direct Pseudoword 95.2 12.1 Young Low Indirect Correct 87 8.3 Misspelled 51 25.6 Direct Correct 89.7 10 Misspelled 60.7 20.8 High Indirect Correct 94.9 7.4 Misspelled 65.5 24.6 Direct Correct 97.3 5 Misspelled 75.2 19.9 N/A Indirect Pseudoword 81.3 18.9 Direct Pseudoword 87.9 14.3
31 Figure 3-6. Recognition accuracy as a function of word frequency and type of stimulus (correct or misspelled). Response Times for Recognition Tasks Response time data were selected to in clude only those responses that were accurate, i.e., where participants correctly recognized words as correctly spelled or misspelled (spelling recognition) or as real or not real (lexical decision), which excluded 15.4% of all young adultsÂ’ responses and 17.3% of older adultsÂ’ responses. In addition, response times lower than one standard de viation and greater than two standard deviations were also excluded for each age group (12.5% of young adultsÂ’ data and 13.2% of older adultsÂ’ data) to rule out a ny responses which were not produced as quickly as possible or where a key was pressed inadvertently. Similar analyses to those conducted on percent accuracy were also conducted on response times. A 2 (Age Group) x 2 (Type of Test) x 3 (Type of Stimulus) ANOVA was conducted to examine differences in recogniti on tasksÂ’ response times between correctly spelled words, misspelled words, and pseudowor ds (see Table 3-3 for means and standard deviations). This analysis showed a main effect of type of stimulus, F (2, 240) = 65.73, MSE = 30964.81, p < .001, where correctly spelled words had faster responses than pseudowords, which were faster than the responses to misspelled words, p s < .001. A Frequency x Type of Stimulus Interaction for Recognition Accuracy40 50 60 70 80 90 100Low FrequencyHigh FrequencyWord FrequencyAccuracy (%) Correctly Spelled Misspelled
32 main effect of age was also revealed, F (1, 120) = 30.61, MSE = 295037.08, p < .002, where older adultsÂ’ response times were slower than young adultsÂ’ response times. A main effect of type of test was not significant, p > .530. Table 3-3. Means and Standard Deviations for Recognition Task Response Times. Age Word Type of Type of Mean Standard Group Frequency Test Spelling/ (msec) Deviation Stimulus (msec) Older Low Indirect Correct 1319.8 280.5 Misspelled 1592.2 361.4 Direct Correct 1324.9 263.2 Misspelled 1518.6 385.9 High Indirect Correct 1218.8 283.7 Misspelled 1544.1 335.7 Direct Correct 1269.6 240 Misspelled 1438.2 320.8 N/A Indirect Pseudoword 1384.3 325.8 Direct Pseudoword 1278.4 282.2 Young Low Indirect Correct 1101.2 205.1 Misspelled 1226.8 249.9 Direct Correct 1139.1 237 Misspelled 1268.3 254.7 High Indirect Correct 1000.5 166.2 Misspelled 1165.1 232.4 Direct Correct 1056.1 194.9 Misspelled 1173.7 204.1 N/A Indirect Pseudoword 1154.1 251 Direct Pseudoword 1185.3 307.9 A significant Type of Stimulus x Ag e Group interaction was revealed, F (2, 240) = 10.44, MSE = 30964.81, p < .001, as well as a significant Type of Test x Type of
33 Stimulus interaction, F (2, 240) = 10.75, MSE = 12364.44, p < .001, but these were qualified by a Type of Test x Type of Stim ulus x Age Group interaction (see Figure 3-7 and Figure 3-8), F (2, 240) = 7.131, MSE = 12364.44, p < .001. Follow-up tests revealed a significant Type of Test x Type of Stimulus interaction for older adults, p < .001, where responses on indirect tests we re significantly slower than direct tests for both misspelled and pseudowords, p s < .001, but responses on direct test s were marginally slower than indirect tests for correctly spelled words, p < .084. Young adults had no significant Type of Test x Type of Stimulus interaction, p > .184, suggesting equivalent response times for direct and indirect tests of each type of stim ulus. In addition, a signifi cant Type of Test x Age Group interaction was revealed for misspelled words, p < .001, and pseudowords, p < .002, but not for correctly spelled words, p > .974. The nature of these interactions was that older adults' responses to misspelled wo rds and pseudowords were especially slowed relative to young adults fo r indirect tests compared with direct tests. Figure 3-7. Recognition task response times as a function of type of test (indirect or direct) and type of stimul us (correctly spelled, missp elled, or pseudoword) for young adults.
34 Figure 3-8. Recognition task response times as a function of type of test (indirect or direct) and type of stimul us (correctly spelled, missp elled, or pseudoword) for older adults. To examine word frequencyÂ’s effect on response times, a 2 (Age Group) x 2 (Word Frequency) x 2 (Type of Test) x 2 (Type of Stimulus ) ANOVA was conducted on response times for recognizing correctly a nd misspelled words (see Table 3-3 for means and standard deviations). A main effect of word frequency was found, F (1, 116) = 52.44, MSE = 27389.07, p < .001, such that responses to highfrequency words were faster than responses to low-frequency words. No signifi cant Word Frequency x Type of Stimulus interaction was revealed, either with, p > .996, or without age group, p > .455.
35 CHAPTER 4 DISCUSSION The purpose of the present experiment was to assess whether the age-related dissociation often found for imp licit and explicit memory retr ieval would also apply to the area of spelling retrieval. Previous research (e.g., Jelici c et al., 1996; LaVoie & Light, 1994; Light & Singh, 1987; Light, Singh, & Capps, 1986; Moscovitch, 1982; Park & Shaw, 1992) has demonstrated that explicit re trieval from memory declines with age, whereas implicit retrieval remains relatively in tact. In contrast, the present research found no age declines for either implicit or explic it retrieval when retrieving correct spelling. These results suggest that age-related declin es in spelling are le ss affected by explicit retrieval than are age-related declines in memory. One explanation, NST and TDH (Burke et al., 2000; MacKay & Burke, 1990), suggests that spelling represents preexisting knowledge that has been stored in memory for many years, unlike memory for recently presented word lists that are of ten used in studies (e.g., Chiarello & Hoyer, 1988; Jelicic et al., 1996; LaVoie & Light, 1994). Within this framework, age declines in explicit and sometimes implicit memory retrie val depend on the extent to which the task requires new connections to be formed. Sin ce spelling does not require new connections to be formed, older adults were able to exp licitly retrieve spelling to the same degree as young adults in the present experiment. Whereas the lack of age differences in implicit spelling retrieval is consistent with many studies of implicit memory (e.g., Light, Singh, & Capps, 1986; Light & Singh, 1987; Moscovitch, 1982; Park & Shaw, 1992), the lack of age differences in explicit
36 retrieval of spelling contradi cts previous research, whic h showed age declines (e.g., MacKay & Abrams, 1998; Stuart-Hamilton & Ra bbitt, 1997). The failure to replicate an age decline in explicit retrieval of spelling c ould be due to several factors. One possible factor is differences in the kind of words used in the different expe riments. Both MacKay and Abrams (1998) and Stuart-Hamilton and Rabbitt (1997) used many irregularlyspelled words, such as Â“rhythm,Â” Â“pneumonia,Â” Â“queue,Â” and Â“epitome,Â” whereas the present experiment used many regularly-spe lled words as well as some irregularlyspelled words (less than 10% were irregularly spelled). Wo rds that are more irregular may be the ones that are more susceptible to age declines, reducing the probability of seeing age declines in the present experiment. Another possibility is that the process of explicit retrieval is not necessa rily what defines age declines in previous studies. The present experiment suggests that older adultsÂ’ ability to spell may be influenced more by other factors, such as word frequency a nd method of response (recognition/production). The factor that had the greatest impact on older adultsÂ’ spelling retrieval in the present research was word frequency. Inde pendent of other factors, older and younger adults had similar accuracy in recognizi ng and producing low-frequency words, but young adults were marginally more accurate than older adults for high-frequency words. This result replicates the findings of MacK ay and Abrams (1998), who used only explicit production of spelling, and extends their result to indirect tests of spelling as well as recognition tasks. The frequency with which people encounter words affects their ability to recognize and produce the correct spelli ng of those words. As proposed by MacKay and Abrams (1998), older adultsÂ’ increased vocabularies enabled th em to recognize and produce the spellings of rare words that young adults were unfamiliar with or may have
37 never seen. Young adults therefore did not know the correct spellings of those lowfrequency words and were less able to judge whether they were real words or were correctly spelled, resulting in no age differences for low-frequency words. Frequency also interacted with the method of response, such that for low-frequency words, there was evidence that recognition of spelling was more accurate than production of spelling, for both age groups, as seen in past research on spelling and aging (e.g., Abrams & Stanley, 2004; MacKay & Abrams , 1996; MacKay et al., 1999). This asymmetry can be explained by NST (e.g., Bu rke et al., 2000; MacKay & Abrams, 1998), where bottom-up priming spreads from the orthographic nodes to the lexical node, allowing spelling recogn ition to occur. In contrast, top-down priming spreads from the lexical node to the orthographic nodes, whic h facilitates produc tion if all of the orthographic nodes are activated correctly and in the prope r sequence. However, highfrequency words were equivalently recogni zed and produced. In NST, high-frequency words do not have weakened connections because of their frequency of use, allowing topdown node priming to activate all of the ort hographic nodes effectively. The advantage in retrieving high-frequency words over low-freque ncy words is consistent with previous research in both the spelling literature (e.g., MacKay & Abrams, 1998; Stuart-Hamilton & Rabbitt, 1997) and the implicit and explicit memory literatures (e.g., Balota & Ferraro, 1996; Davis et al., 1990; Deese, 1960; Delosh & McDaniel, 1996; Gomez, 2002; Hino & Lupker, 2000; Krishnan & Shapiro, 1996; Rose et al., 1986; Spie ler & Balota, 2000; Tainturier et al., 1989; Tan & Ward, 2000; Wa rd et al., 2003; but see Chalmers et al., 1982; Earhard, 1982; Kinsbourne, 1974).
38 Despite differences between recognition a nd production of spelling, age differences did not emerge differentially for recognition a nd production in indirect tests of spelling, contrary to the age dissociation seen in implicit retrieval of words from memory. Fleischman and Gabrieli (1998) proposed that si nce older adults declin e in their ability to produce items from memory, age differences w ill arise when an indirect test relies on production. The present experiment did not support this claim for retrieval of spelling. There were no age differences in spelling ac curacy for either indirect recognition or indirect production, although the difference be tween indirect recognition and production was larger for older adults than young adults, due to a decline in ol der adultsÂ’ production. This provides support for a decline in production with age but does not support Fleischman and GabrieliÂ’s claim that the recognition/production di stinction causes age differences in implicit retrieval. These resu lts are yet another indicator that memory retrieval differs from spelling retrieval, where aging more profoundly affects memory than spelling. The analyses that focused solely on rec ognition tasks revealed that older and young adults were equivalently able to recogni ze spelling for both correctly spelled and misspelled words, independent of whether re cognition was tested indirectly (lexical decision) or directly (spelling recognition). These results ar e consistent with previous research on explicit recognition of spelling in young and older adults (e.g., MacKay et al., 1999; correct spellings in Abrams & Stanle y, 2004). This age equivalence in accuracy came at a cost of slower response times for ol der adults, which potentially could indicate a speed-accuracy tradeoff, wher e older adults took longer to respond to obtain the same accuracy as young adults. Interestingly, despite slower response times, older adults had
39 higher accuracy than young adults when rec ognizing pseudowords, ei ther as nonwords or as being misspelled. Consistent with the ear lier explanation about word frequency and aging, older adultsÂ’ increased vocabulary allows them to discriminate pseudowords from real words more than young adults, who possess less knowledge of words. Older adultsÂ’ response times to the different types of stimuli depended on the type of test, unlike young adults. Olde r adultsÂ’ lexical decisions were faster than spelling recognition for correctly spelled words but were slower than spelling recognition for misspelled and pseudowords. A possible explan ation is that indirect recognition of spelling via lexical decision may require a two-stage process for recognition, where two questions must be answered before recognition oc curs: (1) Is this a word?, and (2) Is this word spelled correctly? This idea is consistent with Burt and Tate (2002)'s proposal that an implicit spelling-checking process occurs in lexical decision. However, the present experiment suggests that this two-stage pr ocess would be differentially slower as a function of type of stimulus. For correctly sp elled words, the answers to question 1 and to question 2 would be Â“yes,Â” resulting in a very quick recognition. For pseudowords, the answers to question 1 and to que stion 2 would be Â“no,Â” resulti ng in a slower response, as negative responses take longer to pr oduce (e.g., MacDonald & Just, 1989). For misspelled words, the answers would be Â“y esÂ” for question 1 and Â“noÂ” for question 2, resulting in conflicting responses which would then slow response times to the greatest degree. In contrast, explicit recognition of spelling does not use this two-stage process because only spelling must be checked to answ er the question, "is this spelled correctly?Â” Within this framework, older adults may have been slower to recognize misspellings, i.e., pseudowords and misspelled words, implicitly than explicitly because
40 of an increased difficulty in processing mo re complex responses, such as those with negation or response conflicts. This age-relate d slowing in indirect recognition of spelling does not occur for correctly spelled words b ecause of the greater frequency with which words are seen as correctly spelled relative to misspelled. Since words are seen most often in their correctly spelled form, answer ing the second question occurs automatically and is facilitated by the response to question 1 (i.e., since the word is real, it should be spelled correctly). In terms of the two theories used to gene rate hypotheses for the present experiment, the results demonstrated some support as well as counterevidence for both theories. With respect to the lack of age differences in expl icit versus implicit retr ieval of spelling, this result was consistent with NST and TDH (Burke et al., 2000; MacKay & Burke, 1990), but not the Dual Route Mode l (e.g., Barry, 1994; Shankwei ler, 1999) as applied in conjunction with prior research on implicit memory (e.g., Jeli cic et al., 1996; LaVoie & Light, 1994; Light & Singh, 1987; Light, Si ngh & Capps, 1986; Moscovitch, 1982; Park & Shaw, 1992), which predicted increased age de clines for explicit retrieval of spelling rules. In terms of recognition versus producti on, both theories predic ted greater accuracy in recognizing spelling than producing spelling as well as no age declines in recognition, both of which were supported. However, the finding of no age decl ines in production of spelling was unexpected and was not consistent with the theory that specifically predicted age declines in spelling production, NST a nd TDH. Lastly, the results regarding frequency supported both theories' claims th at the spelling of high-frequency words are more easily recognized and produced than lo w-frequency spellings, but showed partial support for NST and TDH's predictions of age declines, as age decl ines occurred for
41 high-frequency words but not low-frequency word s. However, this result replicated the Â“familiarityÂ” effect (MacKay & Abrams, 1998), where young adults were unable to retrieve the spelling of words they did not kno w, resulting in age equivalence in spelling low-frequency words. One major conclusion that can be drawn fr om the present research is that although spelling does share some characteristics with memory, such as a decline in production relative to recognition, an advantage in retrieving high-frequency words over lowfrequency words, and no age decline in implic it retrieval, the retrieval of spelling differs in fundamental ways from retrieval from me mory. Specifically, the retrieval of spelling does not decline with age in the same way that memory retrieval does , as age declines in spelling are smaller in magnitude and are most dependent on the variable of word frequency. Spelling may be viewed as a specializ ed form of memory, one that is practiced over many years and is so engrained that distinctions drawn in traditional memory research, e.g., implicit vs. explicit, do not n ecessarily apply to spelling. However, it is important to note that these re sults may not necessarily genera lize to all older adults, but rather represent a best case s cenario of spelling in old age. In practical use, the pres ent experiment suggests that since spelling acts like memory on some levels, the strategies for keeping memory strong in old age can be applied to preserving spelling retrieval long-term. Spelling retrieval must be practiced, both directly and indirectl y, through language tasks such as reading, doing crossword puzzles, and writing letters, in or der to strengthen the connections that weaken in old age. In a sense, we can control a wordÂ’s fre quency, the primary fact or affecting spelling retrieval in old age. The more often we use a word, the higher its frequency and the
42 stronger its connections to its orthography. With stronger connections, we will be able to preserve spelling retrieval for words of any fr equency as well as offset memory declines that accompany the normal aging process.
43 APPENDIX A TARGET WORDS, FREQUENCIES, MISSPELLINGS AND SENTENCES Table A-1. Target Words, Frequenc ies, Misspellings, and Sentences Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production abolish 8 abollish The king wi ll abolish the outdated and inappropriate laws. abstain 1 abstane I will abstain from eating dessert while I am on my diet. accomplice 2 accompliss The accomplice in the crime did not receive a punishment. acquitted 2 aquitted The court acquitted the prisoner of the alleged crime. adamant 5 adament He is adamant about leaving the party at 10 oÂ’clock. adjourn 4 adjurn The judge will adjourn the court session until tomorrow. aggravate 1 aggrevate You will aggravate your siblings if you forget their birthdays. amplify 1 amplefy Amplify your voice when you are speaking in an auditorium. anatomy 9 anatamy Anatomy is not a difficult class if you are good at science. apricot 1 appricot Apricot jam is enjoyable if it is served with toast. aspirin 3 asprin I took aspirin to ease the pain of my broken arm. astronaut 2 astronat The astronaut had to learn to walk on the moon. avocado 9 avacado I bought an avocado to make some guacamole. banana 4 bannana I ate a banana and yogurt with my cereal for breakfast. bankruptcy 8 bankrupcy He was threat ened with bankruptcy when the deal fell through. bedspread 2 bedspred The bedspread was too large to fit in the washing machine. bourbon 4 burbon She had a glass of bourbon at the cocktail party.
44 Table A-1. Continued Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production bungalow 1 bungaloe We rented a bungalow on the beach during our vacation. calligraphy 1 caligraphy The calligraphy used for the wedding invitations received many compliments calorie 1 calarie Counting your calorie intake and exercising are good weight-loss strategies. camouflage 3 camaflage The sold ier wore camouflage during his mission in the woods. canoe 7 canue The canoe sank after the rainstorm filled it with water. cardboard 5 cardbord Over one hundred cardboard boxes were packed before moving day. cauliflower 1 cauliflour Cauliflower is on sale at the grocery store this week. chronological 7 chronalogicalWhen so rted in chronological order, the records are easier to access. counterfeit 1 counterfit Using count erfeit money is a serious crime, punishable by time in jail. cupboard 2 cuboard In the cupboard, there are all of the ingredients for tonightÂ’s dinner. decipher 0 decifer I will decipher the poet's true meaning within a few days. defiance 7 defience The childÂ’s defiance about practicing the piano was no surprise. deodorant 2 deoderant Her new deodorant is scented like baby powder. descendant 2 descendent He is a de scendant of many of our countryÂ’s leaders. dichotomy 1 dicotomy The feud caused a dichotomy between the families. dinosaur 1 dinosoar The only dino saur I remember learning about is the T-Rex. disappoint 0 dissapoint They will disappoint me if they do not come to the party. dissatisfied 6 disatisfied She was di ssatisfied with yesterdayÂ’s opera performance. ecstasy 6 ecstacy A state of ecstasy overcame her when she received an A on her exam.
45 Table A-1. Continued Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production embarrassed 7 embarassed He was embarrassed when he fell in front of the audience. emerald 3 emrald The emerald matched the color of the ocean water in the Bahamas. endeavor 5 endever Her new endeavor is to change the cityÂ’s policy on recycling. erratic 3 irratic The rabbit had an erratic path as it escaped the fox. evaporate 1 evaperate The water will evaporate if it is left in the sun for too long. flourish 5 flurish I will flourish in my new job as soon as I learn the basics. fluoride 2 flouride Toothpaste with fluoride is good for rebuilding enamel on teeth. fulfill 9 fullfill The catering company will fulfill the order for the luncheon. fuschia 0 fusha Everyone noticed her fuschia dress and asked where she bought it. ghoul 1 ghool There was a ghoul in the haunted house we visited on Halloween. girdle 2 gurdle A girdle must be worn under my new dress. jeopardy 4 jepardy He is in jeop ardy of losing all of his money to his ex-wife. labyrinth 1 labrynth This buildin g is a labyrinth, which causes many people to get lost. lacrosse 0 lacross Serious lacro sse players often practice for a minimum of 5 hours daily. lonesome 2 lonesum If you are lonesome, invite some people to come and visit you. mackerel 2 mackeral I will serve mackerel for dinner when my parents come to visit. maneuver 3 manuver The Army maneuver was hard to complete within the time limit. nickel 7 nickle I found a nickel under the couch cushions. niece 8 neice My niece loves to read books about kings and queens. ninetieth 1 nintieth On her nineti eth birthday, she wants to have a celebration. nuisance 5 nuicance There is a childhood nuisance that stands out in everyoneÂ’s memory.
46 Table A-1. Continued Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production pageant 2 pagent The Miss Florida pageant will be taking place in my hometown. plumber 4 plummer I called a plumber when the pipe broke, and the basement flooded. porcelain 2 porcelin She has been collecting porcelain dolls since she was five years old. preference 9 preferance His preference in cars is very different from mine. prestigious 0 prestigous A prestigio us new job was offered to him yesterday. sapphire 0 saphire She wore the sapphire ring that her mother gave to her. satellite 7 satelite This satellite is used to transmit information from one location to another. absence 53 absense Her weeklo ng, unexpected absence worried me. academic 56 accademic Her academic career began after she earned her doctoral degree. achieve 51 acheive Everyone should strive to achieve his or her goals. agency 56 agensy Her insurance agency paid all of her medical bills after the accident. amount 172 amownt The amount of the grocery bill was sixty dollars. ancient 69 anceint Ancient Greek authors wrote many stories with interesting plots. apartment 79 appartment My new apartment is much larger than my old one. apparent 57 aparent It was appare nt that the home team was not going to win. approval 51 approvel Your bossÂ’s approval is necessary before you can take a vacation. argument 63 arguement I avoided the argument that my brothers were having. artery 46 artory A clogged artery can cause a heart attack or stroke. assistance 83 assistence He needed assistance in filing his income taxes. background 67 backround The judgeÂ’s political background is strong and extensive. billion 62 bilion Bill Gates is worth several billion dollars.
47 Table A-1. Continued Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production campaign 75 campain The advertising campaign was the most original he had ever seen. candidate 73 canidate The candidate I voted for was the best for the position. capacity 83 capasity Exceeding the buildingÂ’s capacity is a fire hazard. captain 85 captin The captain of the Titanic went down with the ship. chairman 67 chareman The chairman of the board made the final decision. character 118 caracter The main char acter in the play was a blind woman named Helen. civilization 42 civilizasion The Roma n civilization existed a very long time ago. comfort 43 comfert Comfort is a top priority when buying new furniture. commerce 58 commerse Commerce is important for our countryÂ’s well being. committee 188 commitee The service committee has a meeting at six oÂ’clock tonight. congregation 47 congragation The congr egation was very excited for the festivities to begin. continuous 44 continous An exampl e of a continuous line is the horizon. dictionary 55 dictionery The dictiona ry is a useful tool when writing a paper. director 101 directer The dire ctor critiqued the actorÂ’s performance. doctrine 46 doctrin The legal doctrine called for punishment for all crimes. domestic 63 damestic Dogs and cats are domestic animals, while sharks are not. economy 79 econamy The United StatesÂ’ economy has a potentially positive future. elsewhere 45 elswhere He move d his books elsewhere to study without interruption. environment 42 envirament Caring for the environment is important for all of us to do. existence 106 existance The existe nce of life on other planets is under investigation. familiar 72 familier This building looks familiar, even though I have never seen it before.
48 Table A-1. Continued Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production fashion 69 fashon The latest fashion is frilly skirts with sandals. fought 46 faught They fought until the police arrived and stopped them. friend 133 freind Her friend went to the doctor with her for support. governor 41 governer The current governor must resign, and a new one will be elected. heavily 60 hevily He was heavily saddened by the death of his father. hospital 110 hospetal She went to th e hospital as soon as she went in to labor. immediately 116 imediately They called me immediately when they heard the news. impossible 84 imposible It is almost impossible to solve this puzzle. industry 171 industrey The textile industry employs many people each year. instrument 47 instrament Playing an instrument is a good use of spare time. laboratory 40 labratory The laborator y is where the chemists spend all of their time. language 149 langauge Learning a foreign language like German can be very difficult. marriage 92 marraige The average age of marriage has been increasing steadily. neither 141 niether Neither of th e chairs in the living room matches the new curtains. numerous 47 numerus She had num erous bad dreams last night, resulting in a poor nightÂ’s sleep. occasion 80 occassion To celebrate the occasion, we went out to dinner. occurrence 40 occurance The occurren ce of side effects with the new medicine is low. permitted 43 permited The only people permitted to use the gym are members and guests. personnel 70 personell The office personnel were helpful on my first day of work. private 191 privite The private pr operty had a sign that said Â“no trespassing.Â” received 96 recieved She receiv ed the package that she was expecting in the mail.
49 Table A-1. Continued Target Word Francis and Kucera (1967) Frequency Misspelling for Recognition Tasks Sentence for Spelling Production regular 83 reguler The regular bus schedule is not in effect today. salary 43 salery His salary increased when he was promoted to manager. scientific 86 sientific The scientific experiment revealed the pattern of DNA. secretary 41 secretery The secretary will deliver the message as soon as she finishes lunch. separate 64 seperate I will separate the meat from the vegetables in the refrigerator. shoulder 61 sholder The quarterback injured his shoulder in last weekÂ’s game. successful 94 sucessful She hopes to be a successful doctor by the time she is 30. testimony 47 testamony Her testimony helped him win the case against the thief.
50 APPENDIX B FILLER WORDS AND FREQUENCIES Table B-1. Filler Words and Frequencies Filler Words Francis and Kucera (1967) Frequency add 88 cyclone 0 display 41 emperor 0 gasp 2 greeting 4 hear 153 ideal 45 infinite 0 island 167 main 119 newspaper 64 oatmeal 1 phobia 0 population 136 queen 41 sailor 4 spinach 2 twilight 3 writing 77
51 APPENDIX C PSEUDOWORDS Table C-1. Pseudowords beese chig choad creighth faunch glealth herge jauce konze malve modge nooth plewd relte skose snurr tymph vyme weg zuss
52 APPENDIX D POST EXPERIMENT QUESTIONNAIRE Instructions: Please administer these questi ons verbally, and write down their responses in the space provided. 1. In the task where you were asked to deci de if the words that were shown to you would appear in the dictionary, did you us e any strategies to help you make this decision? If your answer is yes, please explain what. Did you use spelling to help you make this judgment? 2. In the task in which you were asked to wr ite the sentences that you heard, were you concentrating on anything specific? Did you pay attention to spelling? Did you pay attention to grammar? Or were you just trying to get the sentence down as fast as possible
53 LIST OF REFERENCES Abrams, L., & Stanley, J.H. (2004). The detec tion and retrieval of spelling in older adults. In Shohov, S.P. (Ed), Advances in psychology research (Vol. 33, pp. 87109). Hauppauge, NY: Nova Science Publishers, Inc. Balota, D., & Ferraro, F.R. ( 1996). Lexical, sublexical, and im plicit memory processes in healthy young and healthy older adults and in individuals with dementia of the AlzheimerÂ’s type. Neuropsychology , 10 , 82-95. Barry, C. (1994). Spelling Routes (or roots or rutes). In G.D.A. Brown & N.C. Ellis (Eds), Handbook of spelling: Theory, process and intervention . New York: John Wiley & Sons Ltd. Barry, C., & Seymour, P.H.K (1988). Lexical priming and sound-to-s pelling contingency effects in nonword spelling. The Quarterly Journal of Experimental Psychology . 40A , 5-40. Bowles, N.L., Obler, L.K., & Poon, L.W. (1989) . Aging and word retrieval: Naturalistic, clinical and laboratory data. In L.W. Poon, D.C. Rubin, & B.A. Wilson (Eds.), Everyday cognition in adulthood and late life . Cambridge: Cambridge University Press. Burke, D.M., & Harrold, R.M. (1988). Automatic and effortful semantic processes in old age: Experimental and naturalistic appr oaches. In L.L. Light & D.M. Burke (Eds.), Language, memory, and aging (pp 100-116). New York: Cambridge University Press. Burke, D.M., MacKay, D.G., & James, L.E. (2000). In T.J. Perfect & E.A. Maylor (Eds.), Models of cognitive aging (pp 204-237). New York: Oxford University Press. Burt, J.S., & Tate, H. (2002). Does a reading lexicon pr ovide orthographic representations for spelling? Journal of Memory and Language , 46 , 518-543. Chalmers, K.A., Humphreys, M.S., & Dennis, S. (1997). A naturalistic study of the word frequency effect in episodic recognition. Memory & Cognition , 25 , 780-784. Chiarello, C. & Hoyer, W.J. (1988). Adult age differences in implicit and explicit memory: Time course and encoding effects. Psychology and Aging , 3 , 358-366.
54 Craik, F.I.M. (1983). On the transfer of in formation from temporary to permanent memory. Philosophical Transactions of the Royal Society of London , B302 , 341359. Craik, F.I.M , & McDowd, J.M. (1987). Ag e differences in recall and recognition. Journal of Experimental Psychol ogy: Learning, Memory, & Cognition , 13 , 474479. Davis, H.P., Cohen, A., Gandy, M., Colombo, P., VanDusseldorp, G., Simolke, N., & Ramano, J. (1990). Lexical priming de ficits as a function of age, Behavioral Neuroscience , 104 , 288-297. Deese, J. (1960). Frequency of usage and numbe r of words in free r ecall: The role of assocation. Psychological Reports , 7 , 337-344. Delosh, E.L., & McDaniel, M.A. (1996). The ro le of order informati on in free recall: Application to the word frequency effect. Journal of Experimental Psychology: Learning Memory, and Cognition , 22 , 1136-1146. Earhard, B. (1982). Determinants of the word -frequency effect in recognition memory. Memory & Cognition , 10 , 115-124. Fleischman, D.A. & Gabrieli, J.D.E. (1998) . Repetition priming in normal aging and AlzheimerÂ’s disease: A review of findings and theories. Psychology and Aging , 13 , 88-119. Folstein, M.F., Folstein, S.E., & McHugh, P. R. (1975). Mini-mental state: A practical method for grading the state of patients for the clinician. Journal of Psychiatric Research , 12 , 189-198. Gomez, R. (2002). Word frequency effects in priming performance in young and older adults. Journal of Gerontology : Psychological Sciences , 57B , P233-P240. Hino, Y., & Lupker, S.J. (2000). Effects of word frequency and spelling-to-sound regularity in naming with and without preceding lexical decision. Journal of Experimental Psychology: Hum an Perception and Performance , 26 , 166-183. Heller, R.B., Dobbs, A.R., & Rule, B.G. ( 1992). Age differences in communication: Evidence from an on-line video descript ion task. Paper presented to the 4th biennial Cognitive Aging Conference, Atlanta, GA. Howard, D.V. (1988). Implicit and explicit a ssessment of cognitive aging. In M.L. Howe & C.J. Brainerd (Eds.), Cognitive development in childhood: Progress in cognitive development research (pp 3-37). New York: Springer-Verlag.
55 Howard, D.V., & Wiggs, C.L. (1993). Aging and learning: Insight s from implicit and explicit tests. In J. Cerella, W.J. Ho yer, J. Rybash, & M. Commons (Eds.), Adult information processi ng: Limits on loss . New York: Academic Press. Hultsch, D.F., Masson, M.E., & Small, B.J. (19 91). Adult age differences in direct and indirect tests of memory. Journal of Gerontology: Psychological Sciences , 46 , P22-P30. Jelicic, M., Craik, F.I.M., & Moscovitch, M. (1996). Effects of ageing on different explicit and implicit memory tasks. European Journal of Cognitive Psychology , 8 , 225-234. Katz, L., & Frost, S.J. (2001). Phonology constrains the internal orthographic representation. Reading and Writing: An In terdisciplinary Journal , 14 , 297-332. Kinsbourne, M., & George, J. (1974). The mehanism of the word-frequency effect on recognition memory. Journal of Verbal Learning and Verbal Behavior , 13 , 63-69. Krishnan, H.S., & Shapiro, S. (1996). Comparing implicit and explicit memory for brand names from advertisements. Journal of Experiment al Psychology: Applied , 2 , 147-163. Kucera, H., & Francis, W. (1967). Computational analysis of present day American English . Providence, RI: Brown University Press. LaVoie, D., & Light, L.L. (1994). Adult age di fferences in repetition priming: A metaanalysis. Psychology and Aging , 9 , 539-553. Light, L.L., & LaVoie, D. (1992, April). Repetition priming of nonwords in young and older adults: Evidence for formation of new associations? Paper presented at the Cognitive Aging Conference, Atlanta, GA. Light, L.L., & LaVoie, D. (1993). Direct and indirect measures of memory in old age. In P. Graf & M.E.J. Masson (Eds.), Implicit memory: New directions in cognition, development, and neuropsychology (pp. 207-230). Hillsdale, NJ: Lawrence Erlbaum. Light, L.L., & Singh, A. (1987). Implicit and explicit memory in young and older adults. Journal of Experimental Psychol ogy: Learning, Memory, and Cognition , 13 , 531541. Light, L.L., Singh, A., & Capps, J.L. (1986). Di ssociation of memory and awareness in young and older adults. Journal of Clinical and Ex perimental Neuropsychology , 8 , 62-74.
56 MacDonald, M.C., & Just, M.A. (1989). Cha nges in activation levels with negation. Journal of Experimental Psychol ogy: Learning, Memory, and Cognition , 15 , 633-642. MacKay, D.G. (1987). The organization of perception and action: A theory for language and other cognitive skills . New York: Springer-Verlag. MacKay, D.G., & Abrams, L. (1996). Language, memory, and aging: Distributed deficits and the structure of new-versus-old conn ections. In J.E. Birren & K.W. Schaie (Eds.), Handbook of the psychology of aging (4th ed., pp. 251-265). San Diego: Academic Press. MacKay, D.G., & Abrams, L. (1998). Age-lin ked Declines in re trieving orthographic knowledge: Empirical, practical, and theoretical implications. Psychology and Aging , 13 , 647-662. MacKay, D.G., Abrams, L., & Pedroza, M.J. (1999). Aging on the input versus output side: Theoretical implications of ag e-linked asymmetries between detecting versus retrieving orthographic information. Psychology and Aging , 14 , 3-17. MacKay, D.G., & Burke, D.M. (1990). Cognition and aging: A theory of new learning and the use of old connections. In T.M. Hess (Ed.), Aging and cognition: Knowledge organization and utilization (pp. 213-263). Amsterdam: NorthHolland. Mitchell, D.B. (1989). How many memo ry systems? Evidence from aging. Journal of Experimental Psychology: Learning, Memory, and Cognition , 15 , 31-49. Moscovitch, M. (1982). A neuropsychological ap proach to perception and memory in normal and pathological aging. In F. I.M. Craik & S. Trehub (Eds.), Aging and cognitive processes (pp 55-78). New Yorke: Plenum Press. Park, D.C., & Shaw, R.J. (1992). Effect of envi ronmental support on implicit and explicit memory in younger and older adults. Psychology and Aging , 7 , 632-642. Roediger, H.L. (1990). Implicit memory: Retention without remembering. American Psychologist , 45 , 1043-1056. Rose, T.L., Yesavage, J.A., Hill, R.D., & Bower, G.H. (1986). Priming effects and recognition memory I young and elderly adults. Experimental Aging Research , 12 , 31-37. Salthouse, T.A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review , 103 , 403-428.
57 Shankweiler, D. (1999). Words to meanings. Scientific Studies of Reading , 3 , 113-127. Small, B.J., Hultsch, D.F., & Masson, M.E. (1995) . Adult age differences in perceptually based, but not conceptually ba sed implicit tests of memory. Journals of Gerontology: Series B: Psycholog ical Sciences & Social Sciences , 50B , P162P170. Spieler, D.H., & Balota, D.A. (2000). Fact ors influencing word naming in younger and older adults. Psychology and Aging , 15 , 225-231. Stuart-Hamilton, I., & Rabbitt, P. (1997). Age -related decline in sp elling ability: A link with fluid intelligence? Educational Gerontology , 23 , 437-442. Tainturier, M., Tremblay, M., & Lecours, A.R. (1989). Aging and the word frequency effect: A lexical decision investigation. Neuropsychology , 27 , 1197-1203. Tan, L., & Ward, G. (2000). A recency-based acc ount of the primacy effect in free recall. Journal of Experimental Psychol ogy: Learning, Memory, and Cognition , 26 , 1589-1625. Ulatowska, H.K., Hayashi, M.M, Cannito, M. P., & Flemming, S.G. (1986). Disruption of reference in aging. Brain and Language , 28 , 24-41. Ward, G., Woodward, G., Stevens, A., & Stinson, C. (2003). Using overt rehearsals to explain word frequency e ffects in free recall. Journal of Experimental Psychology:Learning, Memory, and Cognition , 29 , 186-210.
58 BIOGRAPHICAL SKETCH I was born in Trumbull, Connecticut, and lived there until I wa s ten years old. I moved to Boynton Beach, Florida, and gra duated from Santaluces Community High School in 1999. In the fall of 1999, I began the pursuit of my psychology degree at the University of Florida. I graduated in 2002 with honors with my Bachelor of Science degree in psychology and a minor in mathematic s. In the fall of 2002, I began my work in the cognitive psycho logy program at the Univ ersity of Florida.