Developmental Science 16:1 (2013), pp 124–135
DOI: 10.1111/desc.12010
PAPER Environmental contributions to preschoolers’ semantic fluency Gitit Kav e,1 Moran Shalmon2 and Ariel Knafo2 1. Department of Education and Psychology, The Open University, Raanana, Israel 2. Department of Psychology, The Hebrew University of Jerusalem, Israel
Abstract Semantic fluency was examined in Hebrew-speaking 5-year-old monozygotic and dizygotic twins (N = 396, 198 pairs), 22% of them with mother-reported speech-related problems. There were positive correlations of similar magnitudes among monozygotic, same-sex dizygotic, and opposite-sex dizygotic twins. Analyses showed no genetic effects, alongside significant shared (39%) and non-shared environmental (61%) effects on fluency scores. The presence of speech-related problems in one twin affected the fluency score of the co-twin. A multivariate regression analysis revealed that parental education and length of stay at daycare significantly predicted fluency scores. We suggest that semantic fluency performance is highly affected by environmental factors at age 5 although genetic effects might emerge later on.
Research highlights
• • •
At age 5, performance on semantic fluency tasks is determined by environmental factors rather than by genetic factors. Parental education and length of stay at daycare predict semantic fluency scores. Environmental effects on fluency performance at age 5 might reflect reliance on executive abilities.
Introduction Word fluency tests are included in almost every neuropsychological evaluation (Ardila, Ostrosky-Solis & Bernal, 2006; Lezak, 1995), as well as in the evaluation of reading disorders (Chilosi et al., 2009; Menghini, Finzi, Benassi, Bolzani, Facoetti, Giovagnolic, Ruffino & Vicari, 2010). On these tasks individuals are asked to generate as many different words as possible in a given time according to a criterion set by the examiner. Verbal fluency performance has been assessed in diverse populations, including children born pre-term or of low birth weight (Aarnoudse-Moens, Weisglas-Kuperus, van Goudoever & Oosterlaan, 2009), children with Down syndrome (Nash & Snowling, 2008), children with specific language impairment (Henry, Messer & Nash, 2012), children who have sustained traumatic brain injury (Prigatano & Gray, 2008; Slomine, Gerring, Grados, Vasa, Brady, Christensen & Bridge-Denckla,
2002), children with ADHD (Pineda, Ardila, Rosselli, Cadavid, Mancheno & Mejia, 1998), and children with epilepsy (Gagliano, Ferlazzo, German, Calarese, Magaz, Sferro & Tortorella, 2007; Hernandez, Sauerwein, Jambaque, De Guise, Lussier, Lortie, Dulac & Lassonde, 2002). Tests of verbal fluency are widely used in clinical settings because they are highly sensitive to impairment in both language and executive functions. Furthermore, they do not require reading or meta-linguistic skills, and thus might be ideal for pre-surgical brain mapping in children (de Guibert, Maumet, Ferre, Jannin, Biraben, Allaire, Barillot & Le Rumeur, 2010). In addition to their clinical usefulness, verbal fluency tests have been used to document normal development in word retrieval across a variety of languages such as English (Delis, Kaplan & Kramer, 2001; Prigatano, Gray & Lomay, 2008), French (Sauzeon, Lestage, Raboutet, N’Kaoua & Claverie, 2004), Hebrew (Kave, 2006; Kave, Kukulansky-Segal, Avraham, Herzberg & Landa, 2010b), Italian (Riva, Nichelli & Devoti, 2000), and Swedish (Tallberg, Carlsson & Lieberman, 2011). Previous research suggests that successful word generation on fluency tests depends on the maturation of executive search strategies no less, if not more, than on the rise in word knowledge (Kave, Kigel & Kochva, 2008; Kave et al., 2010b; Koren, Kofman & Berger, 2005). Studying children aged 6–16, Hurks, Schrans, Meijs, Wassenberg, Feron and Jolles (2010) distinguished between automatic retrieval that takes place within the first 15 seconds and controlled processing that takes place within the remaining 45 seconds of the 1-minute interval. Hurks et al.
Address for correspondence: Ariel Knafo, Psychology Department, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israel; e-mail:
[email protected]
© 2012 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
Envionmental effects on semantic fluency
found that automatic processes reach mature levels earlier than do controlled processes. In normally developing children, executive skills continue to improve as the frontal lobes mature (Anderson, Anderson, Northam, Jacobs & Catroppa, 2001; Fuster, 2002), and it takes a relatively long time for performance on verbal fluency tests to reach adult level (Kave, 2006). In the current work, we set out to examine the genetic and environmental factors that affect individual differences in category fluency performance among preschoolers, using the twin design. Twin studies enable estimation of genetic and environmental contributions to individual differences by comparing monozygotic (MZ) twins, who share 100% of their genetic sequence, with dizygotic (DZ) twins, who share on average half of the variability in genetic sequence. Greater MZ than DZ covariance indicates genetic influence (heritability), covariance beyond this genetic effect indicates an effect of the environment shared by twins (shared environment), and further twin differences reflect the effect of the non-shared environment and measurement error (Plomin, DeFries, McClearn & McGuffin, 2008). Most studies report genetic effects on individual differences in the development of child language (Plomin & Kovas, 2005; Stromswold, 2001). Estimates of the heritability of language components vary widely. For example, Hayiou-Thomas, Kovas, Harlaar, Plomin, Bishop and Dale (2006) estimated the genetic effect on various language measures at .24–.41, whereas DeThorne, Petrill, Hart, Channell, Campbell, Deater-Deckard, Thompson and Vandenbergh (2008) estimated genetic effects at .37–67. This variability reflects differences in sample characteristics, such as child age or the inclusion of children with language disabilities (Kovas, Hayiou-Thomas, Oliver, Dale, Bishop & Plomin, 2005); differences in the linguistic tasks (conversational speech vs. formalized tests), and in the domains of language under study (syntax, phonology, vocabulary, see Stromswold, 2001); as well as differences in whether the assessment of linguistic skills is based on parental report or on direct testing. Hart, Petrill, DeThorne, Deater-Deckard, Thompson, Schatschneider and Cutting (2009) reported similar and relatively modest heritability (.29–.49) and shared environment (.27–.39) influence on a measure of vocabulary that involved word retrieval at age 6, with notable effects of the home literacy environment on the variance in performance. Kovas et al. (2005) as well as Hayiou-Thomas et al. (2006) examined semantic fluency in 4.5-year-old twin pairs along with eight other measures of articulation, phonology, grammar, vocabulary, and verbal memory. Kovas et al. (2005) reported a genetic contribution to semantic fluency of .40, with a shared environment contribution of .11 and a non-shared environment contribution of .49. HayiouThomas et al. (2006) reported that the estimate of genetic contribution to verbal fluency performance at
© 2012 Blackwell Publishing Ltd.
125
age 4.5 was .24, the estimate of shared environment was .21, and the non-shared environment estimate was .54. According to Hayiou-Thomas et al., the genetic contribution to performance on verbal fluency tests might be mediated by the same genetic factors that underlie the development of individual differences in a wide range of language skills. Thus, direct findings on the factors that contribute to individual differences in verbal fluency scores are rather scarce, they were derived from studies of English-speaking children alone, and each study resulted in slightly different estimates. Previous twin studies highlight several important issues that must be taken into account when investigating the genetic and environmental influence on verbal fluency performance at age 5. First, genetic effects are less prominent in early childhood than they are in adolescence and they continue to increase into young adulthood (Haworth, Wright, Luciano, Martin, de Geus, van Beijsterveldt, Bartels, Posthuma, Boomsma, Davis, Kovas, Corley, Defries, Hewitt, Olson, Rhea, Wadsworth, Iacono, McGue, Thompson, Hart, Petrill, Lubinski & Plomin, 2010), while effects of the shared environment decrease with age (Hoekstra, Bartels, van Leeuwen & Boomsma, 2009). Hence, at age 5, we should expect to see more modest heritability estimates and stronger environmental effects than in older children. Specifically, research that compared twins in the US, Australia, and Sweden showed weak (but significant) genetic influence on vocabulary at preschool, with strong shared-environment influence at this age (Byrne, Coventry, Olson, Samuelsson, Corley, Willcutt, Wadsworth & DeFries, 2009; Olson, Keenan, Byrne, Samuelsson, Coventry, Corley, Wadsworth, Willcutt, DeFries, Pennington & Hulslander, 2011; Samuelsson, Byrne, Olson, Hulslander, Wadsworth, Corley, Willcutt & DeFries, 2008). In addition, these studies report a substantial increase in genetic influence alongside a decrease in the influence of the shared environment from preschool to grade 2. Second, genetic estimates of general verbal ability are greater than equivalent estimates in more specific tests such as verbal fluency (Hoekstra et al., 2009), yet general and specific abilities show significant covariance, probably reflecting common genetic factors (Hayiou-Thomas et al., 2006). Third, estimates of the heritability of language differ when children with speech or language disabilities are included in the sample or excluded from it (Bishop & Hayiou-Thomas, 2008; Kovas et al., 2005), either because genetic effects on the disability itself are strong, overlapping the genetic effects on normal language development, or due to the co-occurrence of genetic and environmental effects. According to Stromswold (2001), parents of children with language disorders may use simpler language with their children, so that children with language disorders may receive an input that differs from the input that children with no disorders receive.
126 Gitit Kave et al.
Population-based studies emphasize that socioeconomic factors and parental characteristics have strong and significant associations with cognitive development (e.g. Bradley & Corwyn, 2002; Landry, Smith, MillerLoncar & Swank, 2002; Landry, Smith & Swank, 2006; Rodriguez, Tamis-LeMonda, Spellmann, Pan, Raikes, Lugo-Gil & Luze, 2009; Westerlund & Lagerberg, 2008, among many others). Children of low socioeconomic status (SES) perform below children of high SES on tests of intelligence, language proficiency, and academic achievement (Bradley & Corwyn, 2002). Furthermore, Noble, Norman and Farah (2005) show that SES is most related to performance in the domains of language and executive functions, probably because postnatal development in these domains is more prolonged than in other cognitive domains. SES effects on semantic fluency tasks have been found in normally developing children (Ardila, Rosselli, Matute & Guajardo, 2005; Hurks, Vles, Hendriksen, Kalff, Feron, Kroes, Van Zeben, Steyaert & Jolles, 2006, Hurks et al., 2010; Kishiyama, Boyce, Jimenez, Perry & Knight, 2009), and Gerrard-Morris, Taylor, Yeates, Walz, Stancin, Minich and Wade (2010) show that more optimal family environment (high SES as well as learning support and stimulation in the home) is associated with better fluency scores in children after traumatic brain injury. Thus, we expect that individual differences in performance on semantic fluency tasks will be at least partly genetically related, as is the case for other linguistic tasks. However, performance on this task relies on one’s vocabulary and executive functions and these domains might be determined in part by environmental influences, especially at age 5. The purpose of the present work is to examine the contributions of genetic and environmental effects on performance in a semantic fluency task at age 5, in both normally developing twins and in twins with speech-related problems.
Method Participants The sample included 396 5-year-old children (mean age in months, 61.79, SD = 2.18). Participants were members of the Longitudinal Israeli Study of Twins (LIST; Knafo, 2006; Knafo, Israel & Ebstein, 2011). All Hebrew-speaking Jewish families identified by the Israeli Ministry of the Interior as having twins born in Israel during 2004 and 2005 were contacted with mail surveys regarding children’s development close to the twins’ third birthday. The questionnaire included information on twins’ zygosity, birth and medical history, as well as additional information described by Knafo (2006). Responses from 953 pairs of twins were available following a mail survey conducted when the twins reached the age of 5. In addition, twins living in the
© 2012 Blackwell Publishing Ltd.
Greater Jerusalem area were invited to the lab to participate in an experimental session, in which children were observed performing a variety of tasks and mothers filled questionnaires regarding family and twin background. In all, 71% of those contacted agreed to participate in the lab visit. Family background At the time of the current study, mothers (mean age = 36.63 years, SD = 5.75) reported how many years of formal education they had received and, when applicable, the education level of the other parent. A mean parental education score was computed by averaging the number of years of education of both parents (M = 15.64, SD = 2.72). In addition, mothers reported the family’s relative income by referring to the official average income in Israel, rating whether their family income was (1) much below to (5) highly above the national average (M = 3.13, SD = 1.36). Mothers also reported whether they had other children and the ages of those children. On average, mothers had 1.91 (SD = 1.90) additional children, and twins had 1.42 (SD = 1.88) older siblings and .50 (SD = .69) younger siblings. The families tested in the lab were not significantly different from the larger national sample in terms of income and number of additional children, but parental education was slightly higher in this group relative to the entire sample, t (df = 952) = 2.79, D = 0.18. Twin zygosity Twin zygosity was assessed through DNA tests. When DNA tests were not available, zygosity was assessed through a parent questionnaire of physical similarity, shown to be 95% accurate when compared to DNA testing (Price, Freeman, Craig, Petrill, Ebersole & Plomin, 2000). If DNA information was not available and zygosity was uncertain from the questionnaires, zygosity was estimated from videos of the twins (the video assessment was in 94% agreement with DNA results). The final sample included 51 MZ, 83 DZ same-sex, 60 opposite-sex twin pairs, and four pairs whose zygosity was not established. Table 1 presents further demographic information about the sample. Developmental problems For each twin, mothers were asked ‘would you say that your child has any developmental problems?’ Mothers who responded positively were asked to describe the nature of the problem. In addition, mothers were asked if they have consulted a physician, a physical therapist, or a speech therapist, either in the present or in the past, concerning this problem. While most twins (71%) had no known developmental problems, a substantial minority (29%) did present with such problems. Of those, 7% had
Envionmental effects on semantic fluency
Table 1
Sample characteristics by twin type MZ (N = 101)
Age in months 61.61 (1.99) Parental education (years) 15.18 (2.79) Family income (1–5 scale) 2.85 (1.35) Mean number of older siblings* 1.84 (2.22) Mean number of younger siblings .49 (.61) Caretaker–child ratio at daycare 12.77 (3.85) Length at daycare (% full day) 52% Distribution of speech-related problems* No speech-related problem in either 63% twin Speech-related problem in both twins 22% Speech-related problem in one twin 14% but not the other
Table 2 Number (and %) of twin pairs with developmental and speech-related problems DZ (N = 285) 61.82 15.8 3.23 1.23 .51 12.61 68%
(2.23) (2.71) (1.37) (1.71) (.72) (4.01)
71% 9% 20%
*p < .05, for a difference between MZ and DZ twins. There were no differences between same-sex and opposite-sex DZ twins on any of the variables.
difficulties such as (but not limited to) physical disability or hypotonia, but no speech-related problem. The developmental problems of the remaining 22% included speech-related problems. Speech-related problems If the mother indicated that a child had been treated by a speech therapist, or had hearing or ear problems, the child was identified as having a speech-related problem. Children with a mild cognitive delay were also classified as having speech-related problems (importantly, none of the twins were reported to have cognitive disability such as mental retardation). Table 2 presents the frequency of developmental and speech-related problems in all twin pairs. Daycare information Mothers reported whether their children attended daycare and if so whether they stayed at daycare for a half day (approximately 30 hours a week) or a full day (approximately 45 hours a week). Reports on this question were available for 340 (86%) of the twins. One hundred and twenty-four children (31% of all children, 36% of the children for whom information was available) attended daycare for half a day, and 216 children (55% of all children, 64% of children for whom information was available) attended daycare for a full day. In addition, mothers reported the number of children in the group with each child, as well as the number of daycare providers, and the ratio of caretaker to child was computed (see Table 1 for further details). Procedure Families (usually mother and twins, sometimes accompanied by additional family members) came to the lab where they met two female experimenters (trained to
© 2012 Blackwell Publishing Ltd.
127
Description of problem
Prevalence
No specified problems for either twin Both twins present speech-related problems Both twins present other (e.g. motor), no speech-related problems One twin presents speech-related problems One twin presents other problem, no speech-related problem One twin presents speech-related problems, co-twin presents other problem
119 (62.6%) 24 (12.6%) 9 (4.7%) 29 (15.2%) 3 (1.6%) 6 (3.2%)
Note: Mothers of eight pairs of twins (4%) did not provide data on twins’ developmental problems. Twins were considered as presenting speech-related problems regardless of whether they had additional developmental problems. The group with other (e.g. motor) problems includes only children who do not have mother-reported speech-related problems.
administer the test in a standardized way, to avoid experimenter bias). Visits were scheduled at a time when mothers estimated that children were likely to be at their best. Each twin was concurrently tested in a separate room by one of the two experimenters. Most visits were completed in less than two hours. Semantic fluency was assessed by asking children to generate as many different words as possible within 1 minute, for each of the following three semantic categories: animals, fruit and vegetables, and vehicles (as in Kave, 2006). Fruit and vegetables were treated as one category in order to avoid the ambiguity between botanical definitions and common usage (as in ‘avocado’). It was specified that for the category of vehicles only types of transportation should be provided, whereas brand names were unacceptable. Because our focus was on individual differences, we held the order of semantic categories (animals, fruit and vegetables, and vehicles) constant across participants. Responses were recorded verbatim, with repetitions, the same word with a different ending, or perseverations subsequently excluded from the total score. Words provided in both masculine and feminine forms (e.g. par-para for ‘bull-cow’) were counted as one, whereas an animal and its offspring were counted as separate words (e.g. para-egel for ‘cow-calf’). Names of subcategories (e.g. bird) were not given credit if specific items within that subcategory (e.g. dove, eagle) were also provided. When a questionable response was provided, clarifications were invited at the end of the 1-minute interval. One MZ twin and one DZ twin did not complete the semantic fluency test and were dropped from the analyses. Mean number of correct items generated by the entire sample was 9.00 (SD = 3.99) on the animal category, 5.80 (SD = 2.30) on the fruit and vegetable category, 3.93 (SD = 2.22) on the vehicle category, and 18.75 (SD = 6.77) on all three categories together, in line with data from Hebrew-speaking singletons (Kave, unpublished). Figure 1 presents the distribution of total
128 Gitit Kave et al.
raw scores (sum of three categories) for MZ twins, samesex DZ twins, and opposite-sex DZ twins. Children’s scores on the three categories inter-correlated positively, rs = .39 to .48, p < .01. A factor analysis yielded a single factor on which all categories loaded .79 or higher. This factor accounted for 64% of the variance in test scores. To allow for equal category weights, we first standardized the score of each participant on each category on the basis of the mean and variance of scores on that category across all participants. We then averaged the three standardized category scores to a single semantic fluency score (a = .72). All further analyses were conducted on mean standardized scores. To further examine the reliability and validity of the scale, we analyzed data from a sub-sample of children (N = 52 pairs) who were tested more than a year later, at age 6.5 (mean age 78.36 months, SD = 2.98). Again, children’s scores on the three categories inter-correlated positively, rs = .50 to .63, p < .01, and a single-factor solution emerged, with loadings of .74 or higher, accounting for 65% of the variance in test scores. A longitudinal stability of .62 (p < .001) in test scores pointed to substantial test–retest reliability. Finally, in this sub-sample the total number of words increased significantly from age 5 (M = 20.38, SD = 6.97) to age 6.5 (M = 26.62, SD = 8.08), t(df = 51) = 6.73, p < .001. Such an increase (D = 0.81) is in line with previous reports of developmental trends, adding further validity to this measure.
Results As data from the two twins in each pair are correlated, in the non-genetic analyses the data were treated as
MonozygoƟc twins
DizygoƟc, same-sex twins
non-independent and twins were considered as nested within twin pairs in Mplus (Muthen & Muthen, 1998– 2010). Preliminary analyses showed no differences in fluency levels between girls and boys. MZ twins’ fluency scores were slightly lower (D = .31) than those of DZ twins, b = .23, p < .01. In addition, children’s fluency scores were strongly predicted by their age in months, b = .35, p < .001, an effect accounting for 12% of the variance. Because twins share their age, age effects can inflate the observed twin covariance. Thus, we conducted a regression analysis in which twins’ age and zygosity were used as predictors of fluency scores, and all the following analyses were performed on standardized residual scores based on this analysis. We report three sets of analyses. First, we compare similarity in fluency scores in all MZ and DZ twins. Next, we examine the relation between speech-related problems and fluency scores. Finally, we look at environmental correlates of fluency scores in the entire sample. Genetic and environmental influences on semantic fluency We first compared twin correlations obtained within MZ and DZ pairs (see Table 3). There were positive correlations of similar magnitudes among MZ twins, r = .35, p < .05, same-sex DZ twins, r = .42, p < .001, and opposite-sex DZ twins, r = .37, p < .01, indicating no genetic effect on semantic fluency. The positive twin correlations obtained in all zygosity groups suggest that twin covariance results from the effects of the environment that they share. The moderate size of the correlation across all zygosity groups points to non-shared environmental contributions, in addition to measurement error. To examine genetic and environmental effects more directly, we used the model-fitting Mx structural equation modeling software (Neale, Boker, Xie & Maes, 1999). Including opposite-sex DZ dyads necessitates testing sex-limitation models to see that the model applies similarly to girls and boys and that the genetic and/or environmental factors overlap similarly for samesex and opposite-sex DZ twins (e.g. Galsworthy, Dionne, Dale & Plomin, 2000). Preliminary sex-limitation Table 3
Twin correlations in fluency scores MZ twins
DZ twins
.35** (50) .33* (31)
.40** (142) .39** (98)
DizygoƟc, opposite-sex twins
Total number of words generated
Figure 1 Distribution of raw fluency scores (number of words) summed across three semantic categories.
© 2012 Blackwell Publishing Ltd.
Full sample No speech-related problem in either twin Speech-related problem in both twins Speech-related problem in one twin but not the other
.42 (10)
.36 (12)
.37 (7)
.36* (28)
Note: *p < .05, one-tailed; **p < .01. The number of twin pairs in each analysis appears in parentheses.
Envionmental effects on semantic fluency
analyses showed that setting the genetic correlation between opposite-sex twin pairs at .50 as for same-sex DZ twins, and equating genetic and environmental estimates for boys and girls did not decrease model fit. We therefore estimated a single set of genetic and environmental effects with all combinations of male and female MZ and DZ twins. The initial model estimated genetic effects as accounting for 0% of the variance. Dropping the genetic effect from the model did not affect model fit, Dv2 (df = 1) = .00, ns (Confidence Interval (CI): .00–.58). A model without the genetic component had a good fit to the data, as indicated by several fit indices: v2 (df = 12) = 10.69, ns; AIC (Akaike’s information criterion) = 13.31; RMSEA (root mean square error of approximation) = .05. Shared environment accounted for a substantial proportion of the variance (.39, CI: .26–.51), indicating that the environment shared by twins was associated with individual differences in semantic fluency. The remaining variance (.61, CI: .49–.74) was accounted for by non-shared environment and error. Speech-related problems and semantic fluency No significant differences in fluency scores were found between children with developmental problems other than speech (M = .18, SD = .84) and children without motherreported developmental problems (M = .07, SD = 1.02). We therefore combined all children who had no speechrelated problems and compared them to children whose mother reported that they had speech-related problems. As would be expected, children with mother-reported speech-related problems scored lower (M = .29, SD = .90) than did children without such problems (M = .09, SD = 1.01). Thus, children’s fluency scores were strongly predicted by the presence of a speech-related problem, b = .34, p < .001, an effect accounting for 12% of the variance within the entire sample. The presence versus absence of speech-related problems was more similar for MZ twins than for DZ twins. The probandwise concordance rate (number of probands in concordant pairs as a ratio of the total number of probands, indicating the probability that a co-twin of a twin with a speech-related problem also has a problem) was .76 for MZ twins (N = 18 pairs with at least one proband), .48 for same-sex DZ twins (N = 22 pairs), and .43 for opposite-sex DZ twins (N = 18 pairs). We next used Mx to run a liability threshold model (an analysis in which data on presence or absence of an ordinal variable such as a disorder are assumed to reflect an underlying, normally distributed continuous liability that gives rise to the disorder once a certain threshold is passed; e.g. Kovas et al., 2005). This analysis enables estimation of environmental and genetic contributions to variation in liability in the population. Based on the contingency tables of MZ and DZ twins’ having or not having a speech-related problem, Mx estimated that 57% of the variation in liability to speech-related problems was
© 2012 Blackwell Publishing Ltd.
129
attributed to genetic factors, 29% of the variation was due to shared environment effects, with the remaining 14% attributable to non-shared environment and error. As speech-related problems were found to be heritable in our own data, and because fluency performance was not, it was important to investigate the genetic and environmental contributions to semantic fluency separately for twins with and without speech-related problems. In pairs in which both twins had no reported speech-related problems (N = 130, same-sex and opposite-sex DZ twins were combined as no sex differences were found in earlier sex limitation analyses), semantic fluency scores correlated similarly (see Table 3). In twin pairs in which both twins had a speech-related problem (N = 23), the correlation between semantic fluency scores within the MZ sample was slightly higher than the correlation between semantic fluency scores within the DZ sample. Mx estimated the genetic effects at zero for no-speech-related-problem pairs, and at a low .13 (CI: 0–.76) for twins sharing speech-related problems; these genetic effects could be dropped without affecting model fit, Dv2 (df = 2) = .04, ns. The shared environment effect was estimated at .37 for twins with no known speechrelated problems, and at .29 for twins with such problems. These effects did not differ significantly, as equating the parameter estimates for the two groups did not affect model fit, Dv2 (df = 4) = 0.95, ns. In the final model that included all concordant pairs (with or without speech-related problems), 38% of the variance (CI: .23–.51) was accounted for by the shared environment, with the remaining variance (.62, CI: .49–.77) accounted for by non-shared environment and error. Thus, when speech-related problems are taken into account, twin covariance still appears to be the result of shared environment rather than genetics. Next, we looked at twin pairs discordant for speechrelated problems (N = 35). There were only seven discordant MZ pairs, but the twin correlation in semantic fluency in these pairs was identical to the correlation found in the 28 discordant DZ pairs. These findings further highlight the importance of the shared environment. Figure 2 presents mean fluency scores based on presence or absence of speech-related problems in twins and their co-twins. As can be seen in the first and second columns, twins with speech-related problems scored low, regardless of the problem status of their co-twin, and substantially lower than twins in pairs in which both twins had no speech-related problems (third column). Importantly, in twin pairs discordant for speech-related problems (fourth column), the twin without problems also had a low fluency score, M = .21, SD = 1.13, not significantly better than his or her co-twin with the speech-related problem, M = .30, SD = 0.99, t(34) = 0.48, ns. This result suggests that the presence of speech problems in one twin may affect the fluency level of the co-twin, even when the co-twin does not have the same speech-related problems.
130 Gitit Kave et al.
problems or in children without such problems. The length of stay in daycare related marginally to fluency score, r = .15, p = .099, within the entire sample (Table 4). No effect was found for twins with speechrelated problems, r = .03, ns, while for twins with no speech-related problems length of stay in daycare positively related to fluency scores, r = .25, p < .05. Joint contribution of environmental correlates
Figure 2 Mean standardized fluency scores (and 95% confidence intervals) based on presence or absence of speechrelated problems in twins and their co-twins.
Environmental correlates of semantic fluency In order to better understand the effect of shared environment on fluency scores that was documented in the previous analyses, we first analyzed the correlation between each demographic variable and fluency scores separately (see Table 4), and then analyzed their combined effect through a multivariate regression model. Family background As can be seen in Table 4, a significant positive correlation was found between parental education level and fluency scores when all twins were examined together. When these correlations were examined in twins with and without speech-related problems, it appeared that parental education was not related to fluency scores among twins with speech-related problems, r = .09, ns, but the relation between parental education and fluency scores was quite high for twins with no known speech-related problems, r = .46, p < .001. Fluency scores did not correlate with motherreported family income, r = .07, ns. Siblings While the number of younger siblings was not significantly related to fluency scores in children either with or without speech-related problems, the number of older siblings was associated with lower fluency scores within the entire sample (Table 4). In twins with speech-related problems, the number of older siblings was not related to fluency scores, r = .12, ns, while for twins with no speech-related problems the number of older siblings was significantly associated with lower fluency scores, r = .23, p < .05. Daycare Caretaker–child ratio at daycare did not correlate with fluency scores either in children with speech-related
© 2012 Blackwell Publishing Ltd.
Because the number of older siblings was negatively associated with length of stay at daycare, r = .20, p < .05, it was important to estimate the joint contribution of the three environmental correlates (parental education, number of older siblings, length at daycare) to semantic fluency. An Mplus multivariate regression model showed that the number of older siblings had no significant contribution to fluency scores in children without speech-related problems over and above the other two predictors, b = .19, ns, indicating that the significant correlation with fluency scores might have reflected the variance that this variable shares with length of stay in daycare. We next ran the regression with the two remaining environmental correlates. Parental education, b = .46, p < .001, and length of stay in daycare, b = .25, p < .05, both significantly related to fluency scores, were found to account together for 30% of the variance in fluency scores in children with no known speech-related problems.
Discussion The present study suggests that at age 5, the scores on semantic verbal fluency tests are affected by a child’s environment rather than by genetic factors. Support for this conclusion comes from several different analyses. First, we compared the strength of the correlation between the fluency scores of the two twins and found that scores of both twins were similarly related to one another, regardless of zygosity. That is, correlations of similar magnitude were found between twins’ scores whether they were MZ twins, same-sex DZ twins, or opposite-sex DZ twins. Second, when fluency scores were analyzed for all children through an Mx structural equation modeling (Neale et al., 1999), no genetic effects were revealed, whereas the effect of the shared environment, as well as the effect of the non-shared environment, was highly significant. Third, environmental effects were seen both in twins who had no motherreported speech-related problems and in twins with speech-related problems. Finally, a multivariate regression model found that parental education and length of stay at daycare accounted for 30% of the variance in fluency scores in children with no known speech-related problems. The fact that there were few genetic effects on verbal fluency scores is somewhat surprising given previous
Envionmental effects on semantic fluency
Table 4
131
Correlations between environmental factors and fluency scores Fluency score
Parental education Family income Older siblings Younger siblings Length at daycare Caretaker–child ratio
.35** .07 .19* .04 .15 .12
Parental education
Family income
.06 .04 .09 .07 .09
.07 .03 .01 .03
Older siblings
.15* .20* .02
Younger siblings
.11 .10
Length at daycare
.16*
Note: *p < .05; **p < .01. Results are based on 297 to 390 individual children in 198 twin pairs.
findings of a genetic effect on various aspects of child language around age 5 (Hayiou-Thomas et al., 2006; Kovas et al., 2005; Plomin & Kovas, 2005; Stromswold, 2001). According to Hayiou-Thomas et al. (2006), the same genetic and environmental factors underlie the development of individual differences in a wide range of linguistic skills. We thus expected to find at least some genetic effects in our own data as well. Although the inclusion or exclusion of children with speech problems could in principle affect heritability conclusions, Kovas et al. (2005) showed that the same pattern of results was seen regardless of whether the entire sample or a subsample of children with speech-related problems was analyzed. Our results suggest that even in children with speech-related problems, verbal fluency performance is much more strongly determined by environmental factors than by genetic determinants, mirroring our findings in problem-free children. We also note that although it is quite reasonable to assume that the source of genetic effects on individual differences in conversational language might overlap with the source of individual differences in more formalized language tasks (DeThorne et al., 2008), such as verbal fluency, our findings suggest that this is not necessarily the case. Importantly, we note that previous studies reported that the environment had a strong effect on language skills that complemented genetic effects (e.g. Hayiou-Thomas et al., 2006). The discrepancy between our results of no genetic effects in the normal range of the distribution and Kovas et al.’s (2005) results of significant genetic effects on verbal fluency might stem from the different procedures used in the two studies. Specifically, Kovas et al. (2005) had children generate words in each category for 20 seconds whereas in our study children were given 1 minute per category. Hurks et al. (2006, 2010) argue that children engage in automatic activation of the semantic system within the first 15 seconds, and executive functions that drive controlled search begin to affect performance only later on within the 1-minute interval. It could thus be the case that we failed to find genetic effects on fluency performance because our procedure encouraged children to rely more heavily on executive functions than did the procedure used by Kovas et al. (2005). In fact, Hayiou-Thomas et al. (2006) found genetic effects on verbal fluency but argued that environmental factors were no less important than were
© 2012 Blackwell Publishing Ltd.
genetic factors. This conclusion is supported by research that compared early literacy skills in twins living in the US, Australia, and Sweden (Byrne et al., 2009; Olson et al., 2011; Samuelsson et al., 2008). These crosscultural comparisons indicate that vocabulary at preschool is highly depended on shared environment and more weakly affected by genetic factors, and that the pattern of results is further affected by preschool literacy policies. Previous research has shown that performance on cognitive tasks that rely on both language and executive functioning might be more strongly associated with environmental characteristics, such as SES, than is performance on tasks that involve other cognitive domains, such as visual processing (Kishiyama et al., 2009; Noble et al., 2005). According to Noble et al. (2005), language abilities and executive functions are developed over a prolonged period of time and are thus more susceptible to environmental effects. Specifically, phonological, morphological, or grammatical aspects of conversational speech may develop prior to the completion of the development of executive abilities, which develop much later (Anderson et al., 2001). While the verbal fluency task relies on linguistic abilities, primarily on the availability of vocabulary, it also requires executive control and cognitive flexibility. In fact, Kave et al. (2008) found that the fluency component that improves most through childhood is switching from one subcategory to another. This component reflects strategic search, response initiation, monitoring, shifting, and flexibility (Troyer, Moscovitch & Winocur, 1997; Troyer, Moscovitch, Winocur, Alexander & Stuss, 1998; Troyer, 2000), and is often compromised in individuals with frontal brain damage (Kave, Heled, Vakil & Agranov, 2011). Hurks et al. (2006) found that the level of parental occupational achievement was related more strongly to the controlled processing component of the fluency task than to the automatic component (but see different conclusions in Hurks et al., 2010). Our findings that verbal fluency performance is largely environmentally determined might thus be explained by its reliance on executive functions. The fact that individual differences in executive functions are heavily influenced by environmental factors does not mean that they have no genetic origin, but genetic effects tend to emerge later for these skills. Indeed, Schachar, Forget-Dubois, Dionne, Boivin and
132 Gitit Kave et al.
Robaey (2011) as well as Friedman, Miyake, Young, DeFries, Corley and Hewitt (2008) have documented strong genetic influences on executive functions. These studies measured response inhibition, reaction time, updating working memory representations, and shifting between task sets, but not word retrieval. As executive functions are multifaceted, genetic effects might not be apparent in all of them to the same extent, so that effects that are found for inhibition might not apply to controlled word search. More importantly, while the children who participated in the current research were 5 years old, the participants studied by Schachar et al. (2011) were 8 years old and those tested by Friedman et al. (2008) were 17 years old. Indeed, Schachar et al. (2011) found that genetic effects accounted for 27–50% of the variance in performance on the tasks that they administered, whereas Friedman et al. (2008) found almost 100% heritability. This pattern of results fits well with Haworth et al.’s (2010) findings that genetic effects on general cognitive abilities increase from early childhood to adulthood. With regard to fluency performance, extensive verbal knowledge and lengthy practice in searching one’s mental lexicon are necessary for reaching adult level on word retrieval tasks (Kave, 2006; Kave et al., 2008; Kave, Knafo & Gilboa, 2010a). Thus, genetic effects on verbal fluency performance might be more prominent as executive functions mature, once the relevant skills are already fully developed. Having established that environmental factors account for a significant share of the variance in fluency performance, it remains to define which factors are most influential. Earlier studies implicated parental characteristics such as education and occupation level, as well as parental approach to children, including responsiveness or verbal support that enhance cognitive abilities in general (Bradley & Corwyn, 2002; Landry et al., 2002, 2006; Westerlund & Lagerberg, 2008). Parental education was the strongest predictor of verbal fluency scores in the current study. Nevertheless, parental education is not only an environmental factor but it is also closely related to heritable aspects of cognitive abilities. Our data suggest that children’s verbal fluency performance is more affected by those aspects of parental education that are not directly heritable in the sense of within-family differences (although genetic differences between DZ twins may affect the way they react to influences of parental education). Whether parental education reflects parental genetics, environmental factors, or both, it may affect children as an environmental factor. Put differently, the effect of education might be related to parental genetics but not to genetic differences between siblings, and thus the child experiences this effect as an environmental factor. For instance, the frequency of children’s participation in literacy activities, the quality of parental engagement with their children, and the provision of ageappropriate learning materials have been found to affect children’s language and cognitive skills during the first three years of life (Rodriguez et al., 2009). According to
© 2012 Blackwell Publishing Ltd.
Rodriguez et al., more highly educated mothers are more likely to provide an advantageous home environment, but each aspect of the literacy environment uniquely contributes to the prediction of children’s language and cognitive skills beyond other demographic characteristic such as maternal education. Furthermore, Hammond, Muller, Carpendale, Bibok and Liebermann-Finestone (2012) have recently shown that parental scaffolding (e.g. assistance in problem solving) significantly affects the development of executive functions in preschoolers. It is plausible, then, that these parental behaviors also determine performance on the verbal fluency task studied here, especially in preschoolers, a possibility that should be further explored in future research. In addition to finding environmental effects on fluency performance within a group of twins who were free of speech-related problems, we found that similar environmental effects were also documented in children with speech-related problems. Moreover, twins whose co-twin had a speech-related problem performed more poorly on fluency tasks than did twins whose co-twin had no such problems. As speech impairment itself is heritable, the co-twin and even the parent might have had mild, unrecognized speech-related problems, resulting in lowered fluency scores in the problem-free twin. This finding might also stem from environmental effects on verbal fluency performance. As Stromswold (2001) points out, shared environmental factors include the linguistic input that children receive (assuming parents of twins speak the same way to both). It is thus possible that parents of twins who have speech-related problems use simpler language while speaking with both twins, play less demanding games with their children, or interact with them in a way that leads to a more limited development of the skills necessary for successful performance on verbal fluency tasks. Note that one limitation of our study is the fact that speech-related problems were mother-reported rather than objectively diagnosed, as well as the fact that speech-related problems could have resulted from diverse etiologies, including both speech disorders and language impairments. Although children with reported speech-related problems performed more poorly than did children without such problems, suggesting that subjective reports were not without an objective basis, Bishop and Hayiou-Thomas (2008) point out that self-reports might refer to speech rather than to language problems. The use of twin data may pose another limitation when language tasks are concerned, because even when impaired twins are excluded, twins’ language development appears to lag behind that of singletons (Stromswold, 2001). However, with regard to semantic fluency scores, we found no difference between twins and singletons (Kave, unpublished). Our analyses demonstrated a relatively large effect of non-shared environment on fluency performance. This effect probably reflects measurement error as well as experiences that are unique to each twin. We cannot distinguish between these possibilities but we note that
Envionmental effects on semantic fluency
our measure is both reliable and valid. Psychometric features of test–retest and inter-test (three semantic categories) correlations demonstrated reliability over time as well as within each administration. Moreover, the validity of our measure is seen in the fact that it shows the expected developmental increase in performance over time, as well as in the fact that it was sensitive to speech-related problems. Identifying twinspecific experiences that contribute to differences between twins included in the non-shared environment estimates is an important goal for future research. In conclusion, our analyses show that at age 5 the performance on semantic fluency tasks is determined by environmental factors rather than by genetic factors, unlike the case for other, previously studied verbal abilities. We suggest that our findings reflect the fact that the semantic fluency task relies on executive functions whose development is more prolonged than the development of other aspects of language. In addition, we find that although children with unspecified speech-related problems perform more poorly than do problem-free children, the same pattern of environmental but not genetic effects on performance is seen in this population as well. It remains to be seen whether the inclusion of children with more specific language impairments will lead to different conclusions, and what pattern of findings will be documented for older children or even adults, whose executive functions have already developed.
Acknowledgements The authors are indebted to the parents and twins in the Longitudinal Israeli Study of Twins (LIST) for making the study possible. We also thank Naama Gilat, Noa Gordon Assayag, Adi Oz, Dana Zahar, and the research assistants who collected and coded the data. We thank Frank Spinath and Heike Maas for their advice. LIST was funded by grant no. 31/06 from the Israel Science Foundation to the last author. Preparation of this manuscript was supported by Starting Grant no. 240994 from the European Research Council to the last author.
References Aarnoudse-Moens, C.S., Weisglas-Kuperus, N., van Goudoever, J.B., & Oosterlaan, J. (2009). Meta-analysis of neurobehavioral outcomes in very preterm and/or very low birth weight children. Pediatrics, 124, 717–728. Anderson, V.A., Anderson, P., Northam, E., Jacobs, R., & Catroppa, C. (2001). Development of executive functions through late childhood and adolescence in an Australian sample. Developmental Neuropsychology, 20, 385–406. Ardila, A., Ostrosky-Solis, F., & Bernal, B. (2006). Cognitive testing toward the future: the example of semantic verbal fluency (animals). International Journal of Psychology, 41, 324–332.
© 2012 Blackwell Publishing Ltd.
133
Ardila, A., Rosselli, M., Matute, E., & Guajardo, S. (2005). The influence of the parents’ educational level on the development of executive functions. Developmental Neuropsychology, 28, 539–560. Bishop, D.V.M., & Hayiou-Thomas, M.E. (2008). Heritability of specific language impairment depends on diagnostic criteria. Genes, Brain and Behavior, 7, 365–372. Bradley, R.H., & Corwyn, R.F. (2002). Socioeconomic status and child development. Annual Review of Psychology, 53, 371–399. Byrne, B., Coventry, W.L., Olson, R.K., Samuelsson, S., Corley, R., Willcutt, E.G., Wadsworth, S., & DeFries, J.C. (2009). Genetic and environmental influences on aspects of literacy and language in early childhood: continuity and change from preschool to grade 2. Journal of Neurolinguistics, 22, 219– 236. Chilosi, A.M., Brizzolara, D., Lami, L., Pizzoli, C., Gasperini, F., Pecini, C., Cipriani, P., & Zoccolotti, P. (2009). Reading and spelling disabilities in children with and without a history of early language delay: a neuropsychological and linguistic study. Child Neuropsychology, 15, 582–604. de Guibert, C., Maumet, C., Ferre, J.C., Jannin, P., Biraben, A., Allaire, C., Barillot, C., & Le Rumeur, E. (2010). FMRI language mapping in children: a panel of language tasks using visual and auditory stimulation without reading or metalinguistic requirements. NeuroImage, 51, 897–909. Delis, D.C., Kaplan, E., & Kramer, J.H. (2001). The DelisKaplan executive function system. San Antonio, TX: The Psychological Corporation. DeThorne, L.S., Petrill, S.A., Hart, S.A., Channell, R.W., Campbell, R.J., Deater-Deckard, K., Thompson, L.A., & Vandenbergh, D.J. (2008). Genetic effects on children’s conversational language use. Journal of Speech, Language, and Hearing Research, 51, 423–435. Friedman, N.P., Miyake, A., Young, S.E., DeFries, J.C., Corley, R.P., & Hewitt, J.C. (2008). Individual differences in executive functions are almost entirely genetic in origin. Journal of Experimental Psychology: General, 137, 201–225. Fuster, J.M. (2002). Frontal lobe and cognitive development. Journal of Neurocytology, 31, 373–385. Gagliano, A., Ferlazzo, E., German, E., Calarese, T., Magaz, A., Sferro, C., & Tortorella, G. (2007). Neuropsychological deficits in monozygotic twins with childhood epilepsy with occipital paroxysms. Journal of Clinical and Experimental Neuropsychology, 29, 488–495. Galsworthy, M.J., Dionne, G., Dale, P.S., & Plomin, P. (2000). Sex differences in early verbal and non-verbal cognitive development. Developmental Science, 3, 206–215. Gerrard-Morris, A., Taylor, H.G., Yeates, K.O., Walz, N.C., Stancin, T., Minich, N., & Wade, S. (2010). Cognitive development after traumatic brain injury in young children. Journal of the International Neuropsychological Society, 16, 157–168. Hammond, S.I., Muller, U., Carpendale, J.I.M., Bibok, M.B., & Liebermann-Finestone, D.P. (2012). The effects of parental scaffolding on preschoolers’ executive function. Developmental Psychology, 48, 271–281. Hart, S.A., Petrill, S.A., DeThorne, L.S., Deater-Deckard, K., Thompson, L.A., Schatschneider, C., & Cutting, L.E. (2009). Environmental influences on the longitudinal covariance of expressive vocabulary: measuring the home literacy environment in a genetically sensitive design. Journal of Child Psychology and Psychiatry, 50, 911–919.
134 Gitit Kave et al.
Haworth, C.M.A., Wright, M.G., Luciano, M., Martin, N.G., de Geus, E.J.C., van Beijsterveldt, C.E.M., Bartels, M., Posthuma, D., Boomsma, D.I., Davis, O.S., Kovas, Y., Corley, R.P., Defries, J.C., Hewitt, J.K., Olson, R.K., Rhea, S.A., Wadsworth, S.J., Iacono, W.G., McGue, M., Thompson, L. A., Hart, S.A., Petrill, S.A., Lubinski, D., & Plomin, R. (2010). The heritability of general cognitive ability increases linearly from childhood to young adulthood. Molecular Psychiatry, 15, 1112–1120. Hayiou-Thomas, M.E., Kovas, Y., Harlaar, N., Plomin, R., Bishop, D.V.M., & Dale, P.S. (2006). Common aetiology for diverse language skills in 4½-year-old twins. Journal of Child Language, 33, 339–368. Henry, L.A., Messer, D.J., & Nash, G. (2012). Executive functioning in children with specific language impairment. Journal of Child Psychology and Psychiatry, 53, 37–45. Hernandez, M.T., Sauerwein, H.C., Jambaque, I., De Guise, E., Lussier, F., Lortie, A., Dulac, O., & Lassonde, M. (2002). Deficits in executive functions and motor coordination in children with frontal lobe epilepsy. Neuropsychologia, 40, 384–400. Hoekstra, R.A., Bartels, M., van Leeuwen, M., & Boomsma, D.I. (2009). Genetic architecture of verbal abilities in children and adolescents. Developmental Science, 12, 1041– 1053. Hurks, P.P.M., Schrans, D., Meijs, C., Wassenberg, R., Feron, F.J.M., & Jolles, J. (2010). Developmental changes in semantic verbal fluency: analyses of word productivity as a function of time, clustering, and switching. Child Neuropsychology, 16, 366–387. Hurks, P.P.M., Vles, J.S.H., Hendriksen, J.G.M., Kalff, A.C., Feron, F.J.M., Kroes, M., Van Zeben, T.M.C.B., Steyaert, J., & Jolles, J. (2006). Semantic category fluency versus initial letter fluency over 60 seconds as a measure of automatic and controlled processing in healthy school-aged children. Journal of Clinical and Experimental Neuropsychology, 28, 684– 695. Kave, G. (2006). The development of naming and word fluency: evidence from Hebrew-speaking children between ages 8 and 17. Developmental Neuropsychology, 29, 493–508. Kave, G., Heled, E., Vakil, E., & Agranov, E. (2011). Which verbal fluency measure is most useful in demonstrating executive deficits after traumatic brain injury? Journal of Clinical and Experimental Neuropsychology, 33, 358–365. Kave, G., Kigel, S., & Kochva, R. (2008). Switching and clustering in verbal fluency tasks throughout childhood. Journal of Clinical and Experimental Neuropsychology, 30, 349–359. Kave, G., Knafo, A., & Gilboa, A. (2010a). The rise and fall of word retrieval across the lifespan. Psychology and Aging, 25, 719–724. Kave, G., Kukulansky-Segal, D., Avraham, A., Herzberg, O., & Landa, J. (2010b). Searching for the right word: performance on four word-retrieval tasks across childhood. Child Neuropsychology, 16, 549–563. Kishiyama, M.M., Boyce, W.T., Jimenez, A.M., Perry, L.M., & Knight, R.T. (2009). Socioeconomic disparities affect prefrontal function in children. Journal of Cognitive Neuroscience, 21, 1106–1115. Knafo, A. (2006). The longitudinal Israeli study of twins (LIST): children’s social development as influenced by genetics, abilities, and socialization. Twin Research and Human Genetics, 9, 791–798.
© 2012 Blackwell Publishing Ltd.
Knafo, A., Israel, S., & Ebstein, R.P. (2011). Heritability of children’s prosocial behavior and differential susceptibility to parenting by variation in the Dopamine D4 Receptor (DRD4) gene. Development and Psychopathology, 23, 53–67. Koren, R., Kofman, O., & Berger, A. (2005). Analysis of word clustering in verbal fluency of school-aged children. Archives of Clinical Neuropsychology, 20, 1087–1104. Kovas, Y., Hayiou-Thomas, M.E., Oliver, B., Dale, P.S., Bishop, D.V.M., & Plomin, R. (2005). Genetic influences in different aspects of language development: the etiology of language skills in 4.5-year-old twins. Child Development, 76, 632–651. Landry, S.H., Smith, K.E., Miller-Loncar, C.L., & Swank, P.R. (2002). The role of early parenting in children’s development of executive processes. Developmental Neuropsychology, 21, 15–41. Landry, S.H., Smith, K.E., & Swank, P.R. (2006). Responsive parenting: establishing early foundations for social, communication, and independent problem-solving skills. Developmental Psychology, 42, 627–642. Lezak, M.D. (1995). Neuropsychological assessment (2nd edn.). New York: Oxford University Press. Menghini, D., Finzi, A., Benassi, M., Bolzani, R., Facoetti, A., Giovagnolic, S., Ruffino, M., & Vicari, S. (2010). Different underlying neurocognitive deficits in developmental dyslexia: a comparative study. Neuropsychologia, 48, 863–872. Muthen, L.K., & Muthen, B.O. (1998–2010). Mplus user’s guide (6th edn.). Los Angeles, CA: Muthen & Muthen. Nash, H.M., & Snowling, M.J. (2008). Semantic and phonological fluency in children with Down syndrome: atypical organization of language or less efficient retrieval strategies? Cognitive Neuropsychology, 25, 690–703. Neale, M., Boker, S., Xie, G., & Maes, H. (1999). Mx: Statistical modeling. VCU Box, 900126. Noble, K.G., Norman, M.F., & Farah, M.J. (2005). Neurocognitive correlates of socioeconomic status in kindergarten children. Developmental Science, 8, 74–87. Olson, R.K., Keenan, J.M., Byrne, B., Samuelsson, S., Coventry, W.L., Corley, R., Wadsworth, S.J., Willcutt, E.G., DeFries, J.C., Pennington, B.F., & Hulslander, J. (2011). Genetic and environmental influences on vocabulary and reading development. Scientific Studies of Reading, 15, 26–46. Pineda, D., Ardila, A., Rosselli, M., Cadavid, C., Mancheno, S., & Mejia, S. (1998). Executive dysfunctions in children with attention deficit hyperactivity disorder. International Journal of Neuroscience, 96, 177–196. Plomin, R., DeFries, J., McClearn, G., & McGuffin, P. (2008). Behavioral genetics (5th edn.). New York: Worth Publishers. Plomin, R., & Kovas, Y. (2005). Generalist genes and learning disabilities. Psychological Bulletin, 131, 592–617. Price, T.S., Freeman, B., Craig, I.W., Petrill, S.A., Ebersole, L., & Plomin, R. (2000). Infant zygosity can be assigned by parental report questionnaire data. Twin Research: The Official Journal of the International Society for Twin Studies, 3, 129–133. Prigatano, G.P., & Gray, J.A. (2008). Predictors of performance on three developmentally sensitive neuropsychological tests in children with and without traumatic brain injury. Brain Injury, 22, 491–500. Prigatano, G.P., Gray, J.A., & Lomay, V.T. (2008). Verbal (animal) fluency scores in age/grade appropriate minority children from low socioeconomic backgrounds. Journal of the International Neuropsychological Society, 14, 143–147.
Envionmental effects on semantic fluency
Riva, D., Nichelli, F., & Devoti, M. (2000). Developmental aspects of verbal fluency and confrontation naming in children. Brain and Language, 71, 267–284. Rodriguez, E.T., Tamis-LeMonda, C.S., Spellmann, M.E., Pan, B.A., Raikes, H., Lugo-Gil, J., & Luze, G. (2009). The formative role of home literacy experiences across the first three years of life in children from low-income families. Journal of Applied Developmental Psychology, 30, 677–694. Samuelsson, S., Byrne, B., Olson, R.K., Hulslander, J., Wadsworth, S., Corley, R., Willcutt, E.G., & DeFries, J.C. (2008). Response to early literacy instruction in the United States, Australia, and Scandinavia: a behavioral-genetic analysis. Learning and Individual Differences, 18, 289–295. Sauzeon, H., Lestage, P., Raboutet, C., N’Kaoua, B., & Claverie, B. (2004). Verbal fluency output in children aged 7–16 as a function of the production criterion: qualitative analysis of clustering, switching processes, and semantic network exploitation. Brain and Language, 89, 192–202. Schachar, R.J., Forget-Dubois, N., Dionne, G., Boivin, M., & Robaey, P. (2011). Heritability of response inhibition in children. Journal of the International Neuropsychological Society, 17, 238–247. Slomine, B., Gerring, J.P., Grados, M.A., Vasa, R., Brady, K.D., Christensen, J.R., & Bridge-Denckla, M. (2002). Performance on measures of ‘executive function’ following pediatric traumatic brain injury. Brain Injury, 16, 759–772.
© 2012 Blackwell Publishing Ltd.
135
Stromswold, K. (2001). The heritability of language: a review and metaanalysis of twin, adoption, and linkage studies. Language, 77, 647–723. Tallberg, I.M., Carlsson, S., & Lieberman, M. (2011). Children’s word fluency strategies. Scandinavian Journal of Psychology, 52, 35–42. Troyer, A.K. (2000). Normative data for clustering and switching on verbal fluency tasks. Journal of Clinical and Experimental Neuropsychology, 22, 370–378. Troyer, A.K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two components of verbal fluency: evidence from younger and older healthy adults. Neuropsychology, 11, 138–146. Troyer, A.K., Moscovitch, M., Winocur, G., Alexander, M.P., & Stuss, D.T. (1998). Clustering and switching on verbal fluency: the effects of focal-frontal and temporal-lobe lesions. Neuropsychologia, 36, 499–504. Westerlund, M., & Lagerberg, D. (2008). Expressive vocabulary in 18-month-old children in relation to demographic factors, mother and child characteristics, communication style and shared reading. Child: Care, Health and Development, 34, 257–266. Received: 21 May 2012 Accepted: 19 August 2012
