Running Head: INDIVIDUAL DIFFERENCES IN ADULT SPELLERS Aging, Neuropsychology, and Cognition, 14, 529-544 Individual Differences in Young and Older Adults' Spelling: Do Good Spellers Age Better than Poor Spellers?

Young and older adults’ ability to retrieve the spellings of highand low-frequency words was assessed via tests of spelling recognition and production. One of the spelling production tests required participants to write down the correct spellings of auditorily presented words, and accuracy was used to categorize participants in both age groups as good or poor spellers. The results showed that individual spelling ability and word frequency contributed to age differences. Older adults who were poor spellers were less accurate in recognizing and producing correct spelling than young adults who were poor spellers. In contrast, no age differences occurred for good spellers. Furthermore, low-frequency words were especially difficult for young adults and poor spellers, relative to older adults and good spellers. These results indicate that aging alone is not detrimental to the processes underlying recognition or production of spelling but instead compounds existing problems caused by poor spelling. Individual Differences in Adult Spellers 3 Individual Differences in Young and Older Adults' Spelling: Do Good Spellers Age Better than Poor Spellers? To date, research on spelling in adults has focused almost exclusively on patients with brain damage (e.g., Badecker, Hillis, & Caramazza, 1990; Caramazza & Hillis, 1990; Katz, 1991; Levine, Mani, & Calvanio, 1988; Neils, Boller, Gerdeman, & Cole, 1989; Neils, Roeltgen, & Greer, 1995; Tainturier & Rapp, 2004). This literature documents decrements in spelling in people with neurological damage, such as Alzheimer's disease. Patients with Alzheimer's disease exhibit significant impairments in spelling ability, including increased production of spelling errors, especially for longer words, and a greater proportion of errors that are phonologically incompatible with the correct spelling (e.g., Neils et al., 1989; Neils et al., 1995). Although these declines in spelling ability are useful for understanding cognitive changes following brain damage, they do not further our knowledge about patients' age-matched counterparts, healthy older adults. Furthermore, these studies focus solely on written spelling production, with direct instruction to try and produce correct spelling. The purpose of the present experiment was to examine the effects of individual differences in spelling ability on young and older adults' recognition and production of spelling, using three spelling tasks different from the traditional single-word spelling production test. Very few studies concerning spelling in healthy older adults have been conducted (Abrams & Stanley, 2004; MacKay & Abrams, 1998; MacKay, Abrams, & Pedroza, 1999; Stuart-Hamilton & Rabbitt, 1997). These studies have shown that older adults' ability to recognize a word as correctly spelled is less susceptible to age declines than producing correct spelling (e.g., MacKay et al., 1999). Using spelling recognition tasks (Abrams & Stanley, 2004; MacKay et al., 1999), participants were asked to indicate whether briefly presented words Individual Differences in Adult Spellers 4 (ranging from 50-390 msec) were spelled correctly or incorrectly. In both studies, older adults were equivalent to young adults in recognizing words as correctly spelled, suggesting that spelling recognition processes are relatively intact in older adults. However, rapid presentation of targets that disappeared from the screen before a response was given is not the most naturalistic test of spelling recognition; people are not typically time-pressured to detect spelling during everyday tasks, such as reading. In contrast to recognition, MacKay and Abrams (1998) found that young-old (mean age 67.2) and old-old (mean age 77.0) adults were less able to write down the correct spellings of auditorily presented words than college students. Similarly, StuartHamilton and Rabbitt (1997) found that adults in their 70s produced written correct spellings less often than adults in their 60s, who were less accurate than adults in their 50s. Again, these tests of spelling differ from everyday tasks involving writing, such as taking notes in class or writing down phone messages, where whole sentences are produced, and attention is not directly focused on spelling. In conjunction with age, spelling retrieval is likely to be influenced by other factors, such as individual differences in spelling ability and word frequency. Individual differences in spelling have been studied in children and young adults by classifying individuals as good or poor spellers, usually by using scores on a written test of spelling production following auditorily presented words (e.g., Burden, 1989; Cobb, Kincaid, & Washburn, 1918; Dietrich & Brady, 2001; Holmes & Malone, 2004; Holmes & Ng, 1993; Kamhi & Hinton, 2000; Ormrod, 1990). The decrement in spelling production for poor spellers has been attributed to incomplete orthographic representations of words (e.g., Holmes & Malone, 2004), insufficient knowledge about the rules of spelling (e.g., Kamhi & Hinton, 2000), and inadequate phonological awareness (e.g., Allyn & Burt, 1998). One possibility is that these difficulties associated with poor spellers Individual Differences in Adult Spellers 5 are compounded by age, such that older adults who are poor spellers will have increased spelling difficulties relative to young adult poor spellers. In contrast, older adults who are good spellers may be able to offset age-related declines in spelling, analogous to Burt and Butterworth's (1996) suggestion that children who are good spellers may be able to maintain in adulthood their level of spelling knowledge and ability to use phonological information for spelling. This variable of spelling ability may explain the lack of age differences in spelling production observed in several unpublished studies (e.g., Abrams, White, McDermott, & Wolf, 2000; Kramer, Burke, & Taylor, 2000). Another factor shown to influence spelling retrieval is word frequency, or how often a word is encountered in a language (by use of reading, speech, etc.), but its interactions with aging on spelling have yielded mixed results. Stuart-Hamilton and Rabbitt (1997) compared adults in their 50s, 60s, and 70s (but no young adults), showing an age-linked decline in spelling production with respect to low-frequency words but not high-frequency words. However, MacKay and Abrams (1998) observed the opposite pattern, where old-old adults exhibited a spelling production deficit relative to college students for high-frequency words, but not lowfrequency words, although low-frequency words were spelled correctly less often than highfrequency words for all age groups. MacKay and Abrams (1998) suggested that the young adults were less familiar with the low-frequency words than older adults, masking the age-linked decline in spelling production that would have appeared if both age groups were equally familiar with the words. Nonetheless, the appearance of age differences in these studies as a function of frequency suggests the importance of manipulating word frequency in assessing older adults' spelling capabilities. Individual Differences in Adult Spellers 6 Two theories can be used to understand the processes underlying spelling production. One framework, dual route theories (e.g., Barry, 1994; Barry & Seymour, 1988; Coltheart, 1978; Coltheart, Curtis, Atkins, & Haller, 1993), proposes two routes to produce spelling: a lexical process and a nonlexical process. In the lexical process, spellings of words are retrieved directly from a storage base of words, i.e., a word's spelling is retrieved directly from memory. The nonlexical process functions by assembling the spellings of words from sound-to-spelling conversions, i.e., words are spelled by following rules that are stored in memory. The nonlexical route works well for words that are spelled the way they sound, but cannot be the sole route for retrieving spelling; using this route to spell irregularly-spelled words (words spelled differently from how they sound, e.g., yacht) would result in phonologically plausible misspellings (Barry, 1994). Therefore, the lexical process is used for words that are not spelled the way they sound, so that they can be spelled correctly by retrieving the whole word directly from memory. An alternate theory of spelling is derived from Node Structure Theory (NST; Burke, MacKay, & James, 2000; MacKay, 1987; MacKay & Abrams, 1998), a connectionist framework where words, their sounds (phonology), their spellings (orthography), and their meanings (semantics) are stored in nodes that are interconnected at multiple levels. The connections between a word and its orthography can become weakened, e.g., through low frequency of use or non-recent use, causing less accurate retrieval of a word’s spelling. The Transmission Deficit Hypothesis (TDH; MacKay & Burke, 1990) proposes that the normal aging process also weakens connections between nodes by decreasing the transmission of priming across connections, suggesting that older adults will be especially vulnerable to declines in spelling retrieval, consisting with previous research (e.g., MacKay & Abrams, 1998). Individual Differences in Adult Spellers 7 Whereas both theories predict greater spelling accuracy for (1) high-frequency words relative to low-frequency words, and (2) good spellers relative to poor spellers, dual route theories do not specifically address spelling in older adults. In contrast, NST and TDH propose interactions of these variables with age. NST and TDH predict that the spelling of low-frequency words will be more difficult to retrieve or recognize, especially for older adults, because these connections to orthography are weakened both by infrequent use and by aging. NST and TDH also predict that poor spelling ability will be especially detrimental to older adults. Poor spellers in general have weakened connections to orthography, whereas older adult poor spellers have exacerbated weakening of their connections due to age. In contrast, good spellers have strong orthographic connections, minimizing any weakening that may have been caused by age. In sum, any variable that adversely affects spelling retrieval will have a greater influence on older adults. Method