Abstract
Objective
The number of years of education an individual completes is related to their future morbidity and mortality. There are obvious drivers for educational attainment such as childhood intellect, parental intelligence and education attainment, as well as socioeconomic status; and associations may be age‐dependent. We investigated associations between intelligence across childhood (collected at two, four, seven and between eleven and thirteen years) and educational attainment (total years) by the late 20s in the Port Pirie Cohort Study, taking into account maternal intelligence, parental schooling and occupational prestige.
Method
There were 388 individuals from the population‐based longitudinal Port Pirie Cohort Study (South Australia) who provided educational attainment data in the 2008‐9 data collection wave. A Structural Equation Model was employed to test associations between educational attainment and childhood cognitive/IQ measures, taking into account parental factors of IQ, schooling and socioeconomic status.
Results
The vast majority of variables displayed significant simple correlations with each other in expected directions, e.g. child cognitive/IQ measures with maternal IQ. In the full structural equation model, paternal schooling and child intelligence at seven years were the only variables significantly related to educational attainment by the late 20s; maternal intelligence was strongly associated with early life but not adolescent intelligence.
Conclusions
These findings highlight the complex inter‐generational transmission of social advantages, and substantiate the independent effects of education and intelligence on later morbidly and mortality.
Background
Educational attainment, that is, how many years of schooling and any tertiary studies an individual completes, is a critical factor to understand within a public health context as it associated with morbidity and mortality. For example, those who complete more education in early life have better physical health (Wrulich et al., Citation2013) including a lower risk of cardiovascular disease and obesity in midlife (Chandola, Deary, Blane, & Batty, Citation2006; Lawlor, Clark, Davey Smith, & Leon, Citation2006; Lawlor, David, Clark, McIntyre, & Leon, Citation2008; Paile‐Hyvärinen et al., Citation2009; Richards et al., Citation2009; Yu, Han, Cao, & Guo, Citation2010), rate their midlife health as better (Hagger‐Johnson, Batty, Deary, & von Stumm, Citation2011; Wrulich et al., Citation2013), display improved health behaviours in midlife (Gale, Johnson, Deary, Schoon, & Batty, Citation2009), have higher cognitive performance (Clouston et al., Citation2012; Wilson et al., Citation2009) along with a lower risk of dementia in late life (EClipSE, Citation2010), and die later (Lager, Bremberg, & Vågerö, Citation2009), as compared with those with less education. These mortality and morbidity associations with education attainment are well established; however, the predictors of the accrual of educational attainment are not.
Childhood intelligence is an obvious driver for educational attainment in adulthood (Batty, Kivimaki, & Deary, Citation2010; Deary & Johnson, Citation2010; Lager, Modin, De Stavola, & Vågerö, Citation2012; McCall, Citation1977; Strenze, Citation2007). In a meta‐analysis, including nearly 85,000 individuals over 59 studies, intelligence correlated (sample size weighted average correlation) with educational attainment at r = .56 (95% confidence interval (CI) .53–.58), with intelligence measured between 3 and 23 years, and educational attainment between 20 and 78 years (Strenze, Citation2007). When age at intelligence testing was taken into account, it was found that the strength of the relationship between intelligence and educational attainment increased as childhood age increased—the sample size weighted correlation was r = .42 for intelligence testing taken 3–10 years, r = .57 for 11–15 years, r = .58 for 16–18 years, and r = .61 for 19–23 years (Strenze, Citation2007)—which may be due to increasing influence of genes and/or life experience. This analysis employed a cross‐sectional approach as data were collected from different individuals across childhood, and there appears a lack of longitudinal research where the contributions of multiple measures of childhood intelligence to later educational attainment are assessed within the same cohort. This is particularly important for teasing apart the contributions of childhood intelligence and educational attainment on midlife and late‐life factors such as health, as some variability in the intelligence scores obtained once the child is at school is likely due to the differential effects of education on cognitive development (Wrulich et al., Citation2013). By measuring cognitive ability or intelligence before entry into school, or early into schooling, the effects of education on scores obtained are minimised.
Parental factors such as intelligence, their own educational attainment, as well as socioeconomic status influence their child's intelligence (Der, Batty, & Deary, Citation2006; Johnson, Gow, Corley, Starr, & Deary, Citation2010; Lager et al., Citation2012; Sanson, Smart, & Misson, Citation2011). For example, maternal intelligence has been found to be strongly related to the cognitive development of the child (Der et al., Citation2006); and paternal socioeconomic status has been positively related, and paternal educational attainment negatively related, to childhood intelligence at 11 years of age (Johnson et al., Citation2010). Lager et al. (Citation2012), however, found a positive relationship between paternal schooling and intelligence of their children (measured at 10 and 20 years) and educational attainment at age 20. The differential importance of these various parental factors to the relationship between their child's intelligence and education is therefore unclear.
It was our aim to understand the predictors of education attainment by investigating relationships between childhood intellectual development measures (general cognitive ability at 2 and 4 years; intelligence at 7 and 11–13 years) and later adult educational attainment (collected in late 20s). Further, we aimed to assess the contributions of parental factors including years of schooling completed, socioeconomic status, and intelligence. In doing so, we sought to identify any age‐dependent associations between childhood intelligence and education attainment along with parental factors that may exist.
Methods
Participants
Individuals were from the population‐based Port Pirie Cohort Study (Baghurst et al., Citation1992; McFarlane et al., Citation2013; McMichael et al., Citation1988; Tong, Baghurst, McMichael, Sawyer, & Mudge, Citation1996). The cohort included 723 live births between September 1979 and October 1982 in the regional industrial town of Port Pirie (within 30-km) in South Australia (90% response rate).
All cohort participants were assessed periodically from birth to 7 years, and then a subsample of those who had completed the study at age 7 years was assessed between 11 and 13 years. All members of the cohort from birth were approached in their late 20s (2008–2009) and 56% agreed to participate (n = 402). Those who agreed to participate in 2008–2009 had a larger birth weight and gestational age, their mothers were older at their birth and had lived in Port Pirie for longer, and their parents were less likely to be smokers, as compared with those who did not agree to participate—all differences were of small effect size. For full details, please see McFarlane et al. (Citation2013).
This analysis includes the 388 individuals who provided educational attainment data in the 2008–2009 data collection wave. Fifty‐five per-cent were female (n = 213; n = 175 males), and individuals varied between 25 and 29 years at the adult data collection wave (M = 36.94, standard deviation (SD) = .84 years). Ninety per-cent of individuals were firstborns (n = 344), and the remainder were second‐borns (n = 39).
Measures
Research psychologists administered all of the testing sessions to the children and the mother, and were blinded to the results of previous cognitive assessments.
Measures collected from child
Bayley Mental Development Index (child at 2 years of age)
The mental scale of the Bayley Mental Development Index has 163 items and was used to assess the development of each child at 22–26 months of age (Bayley, Citation1969). It covers memory, learning, problem solving capacity, early language and speech development, and the understanding of object permanence. The raw Bayley Mental Scale is converted to the age‐normed Mental Development Index with a mean score of 100 and SD of 16.
McCarthy Scales of Children's Abilities (child at four years)
The McCarthy Scales of Children's Abilities were used to assess the developmental competence of each child at 4 years of age (MCarthy, Citation1972). The McCarthy Scales consists of five subscales including: verbal, perceptual performance, quantitative, memory, and motor skill tasks. The first three subscales are combined to produce a general cognitive index with a mean of 50 and a SD of 10, which provides an age‐normed index of cognitive functioning.
Wechsler Intelligence Scale for Children (child at 7 and 11–13 years)
The revised version of the Wechsler Intelligence Scale for Children (WISC‐R) was used to assess general intelligence at 7 years of age, and between 11 and 13 years (Wechsler, Citation1974). It was designed for children aged between 6 and 16 years of age and included 10 subtests. The age‐normed full‐scale intelligence quotient (IQ) had a mean score of 100, with an SD of 15.
Educational attainment (of child; 25–29 years)
The total number of years of schooling (primary and secondary school, as well as any tertiary studies) was calculated from questions asked at the 2008–2009 data collection wave.
Cumulative blood lead
The Port Pirie Cohort Study began primarily to investigate the effect of lead on physical and cognitive outcomes in children, as Port Pirie hosts a lead smelting plant (Baghurst et al., Citation1985; McMichael, Baghurst, Robertson, Vimpani, & Wigg, Citation1985; Vimpani, McMichael, Robertson, & Wigg, Citation1985). Methods for blood collection and lead calculation are detailed most recently in McFarlane et al. (Citation2013). Variables assessing cumulative blood lead concentration up to four assessment ages (e.g., up to 2 years, 4 years, 7, years and 11–13 years) were used within each subregression with child cognitive performance/intelligence as the outcome (2, 4, 7, and 11–13 years). Notably, within this sample blood lead concentration has been found to relate to small cognitive deficits (Baghurst et al., Citation1992; Tong et al., Citation1996), but not the trajectory of cognitive development (Tong, Baghurst, Sawyer, Burns, & McMichael, Citation1998) or adult psychiatric diagnosis in this cohort (McFarlane et al., Citation2013).
Measures collected from parents
Wechsler Adult Intelligence Scale‐Revised (maternal; when child was between 3 and 4 years)
Maternal intelligence was measured using the Wechsler Adult Intelligence Scale‐Revised (Wechsler, Citation1981) when the child was between 3 and 4 years. Notably, the cohort includes siblings, in which case maternal intelligence was only taken once (when the first child was 3–4 years).
Maternal and paternal schooling
The number of years of secondary school completed was reported for the mother and father of each child. Because of questionnaire design, any further study was not reordered however in this sample only 16% of mothers and 20% of fathers completed secondary school.
Maternal and paternal occupational prestige
Occupational prestige of both the mother and father was assessed via the Daniel Scale as a proxy of socioeconomic status (Daniel, Citation1984). The Daniel Scale was developed on an Australian sample and uses a 7‐point scale, from 1 to 7 to indicate occupational prestige, where a lower score equates to a more prestigious occupation, which is usually associated with high socioeconomic status. These data were collected at baseline, usually within the first trimester of pregnancy of the child later followed in the study.
Statistical analysis
There were missing data cross all variables except the outcome of education attainment for all n = 388. The percentage of missing data was (from n = 388 total) age 2 cognitive performance (3%), age 4 cognitive performance (9%), age 7 IQ (14%), age 11–13 IQ (38%), maternal IQ (25%), maternal occupational prestige (1%), paternal occupational prestige (1%), maternal schooling (2%), paternal schooling (9%), 2 years blood lead (19%), 4 years blood lead (26%), 7 years blood lead (29%), and 11–13 years blood lead (45%).
A structural equation model was considered to test associations between educational attainment and childhood cognitive/IQ measures (see Fig. 1 for a simplified version of the model fitted). The model was fitted in Mplus (Muthén & Muthén, Los Angeles, CA, USA) and estimated using full information maximum likelihood (FIML) estimation. Continuous variables with missing values were estimated within the model via the FIML method under the missing at random assumption. Model fit was assessed using standard fit indices such as the root mean square error of approximation (RMSEA), standardised root mean square residual (SRMR), and the comparative fit index (CFI).
Figure 1. Simplified version of the statistical model fitted. Notably, all parental factors and child sex were entered as separate variables in the full model.
Results
Descriptions of key variables included in the statistical model are given in Table , and correlations between these variables are presented in Table . It can be seen that the vast majority of variables were significantly correlated with each other in expected directions. For example, all child cognitive/IQ measures were significantly positively correlated and maternal IQ was positively correlated with all of these child assessments. Notably, the measure of maternal and paternal occupational prestige (Daniel Scale) uses a scale where lower numbers represents more prestige than higher numbers; therefore, maternal and paternal occupational prestige correlated negatively with cognitive and IQ assessments. The only exception from the broad pattern of significant correlations was maternal occupational prestige, which only significantly correlated with maternal IQ, maternal schooling, and paternal schooling.
Table 1. Participant characteristics of sample
Table 2. Correlations and their significance between key study variables
Results from the model are presented in Table , and significant effects are illustrated in Fig. 2. Fit indices indicated fair to good model fit (RMSEA index .06, 90% CI = .04, .075; SRMR = .052; CFI = .907).
Table 3. Full results from the SEM
Figure 2. Diagram showing significant (indicated by arrow) effects from the structural equation model (SEM), with standardised betas.
It was found that educational attainment (i.e., the number of years of education the child completed by their late 20s) was significantly associated with number of years of schooling completed by their father (standardised beta = .227, standard error = .065, p = .001) and IQ at 7 years (standardised beta = .218, standard error = .120, p = .032). That is, for each additional year of high school completed by the father, the child completed on average, nearly half a year more education (any schooling and tertiary studies). Further, for each additional IQ point at age 7, that an individual on average completed nearly half a month more education by their late 20s—translating to a 1‐year educational attainment difference between the bottom and top of the average IQ range (i.e., IQ score of 85–115). When taking into account all other model variables, educational attainment was not significantly associated with any other childhood cognitive ability or IQ measure (i.e., at 2 and 4 years, and in early adolescence; however, notably the effect was consistent but failed to reach conventional significance levels at early adolescence with p = .099), maternal IQ, or with maternal schooling, parental occupational prestige, and the sex of the child.
In terms of subregression results within the model, it was evident that childhood cognitive ability was a highly stable construct with each cognitive/IQ measurement related to the one taken prior (i.e., 4 years related to 2 years, standardised beta = .360, standard error = .052, p < .001; 7 years related to 4 years, standardised beta = .377, standard error = .050, p < .001; and 11–13 years related to 7 years, standardised beta = .742, standard error = .038, p < .001). Further, maternal IQ had a large effect on early childhood cognition and IQ, with positive relationships when the child was 2 years (standardised beta = .324, standard error = .070, p < .001), 4 years (standardised beta = .272, standard error = .073, p < .001), and 7 years (standardised Beta = .195, standard error = .069, p = .004); however, the association between maternal IQ and childhood IQ in early adolescence (11–13 years) was not significant (standardised beta = .038, standard error = .063, p = .541).
Although years of paternal schooling showed the strongest relationship with the child's educational attainment, paternal schooling was not significantly related to the child's IQ/cognitive ability in the model (at all measurement points; notably, paternal schooling was significantly correlated with all child cognitive/IQ measures). Maternal years of schooling and parental occupational prestige were not significantly related to any child measure (cognitive ability/IQ or educational attainment), and only related to maternal IQ, within the model.
Sex was related to two childhood cognitive ability/IQ scores—at 2 years (standardised beta = .394, standard error = .100, p < .001) and in early adolescence (standardised beta = −.221, standard error = .081, p = .006)—meaning that girls outperformed boys at 2 years and boys outperformed girls in early adolescence. Blood lead concentrations were not associated with cognitive performance at any age, when taking into account covariates, within the model.
Discussion
In this large Australian sample, the best predictors of education attainment by the late 20s were intelligence scores at 7 years and paternal schooling. Cognitive ability at 2 and 4 years, and intelligence in early adolescence, did not account for additional significant variance in education attainment in the full model. Despite the father's schooling having a large effect on educational attainment of the child, it was not significantly associated (taking into account other model variables) to the cognitive ability of the child across their development.
Intelligence appeared to be highly stable trait, which aligns with over a century of research into the construct (Deary, Citation2012) and recent published work in a longitudinal child sample (Bornstein et al., Citation2006). The predictive strength increased with age, that is, age 7 intelligence predicted early adolescent intelligence much better than age 4 to age 7, and age 2 to age 4. One reason for this is that age 7 and early adolescent function were both measured via the WISC‐R, whereas measures taken at ages 2 and 4 years related to cognitive function and not intelligence per se (there was no Wechsler scale suitable for these age groups at the time). Another reason for the increase in predictive power of intelligence/cognition with age is that is has been known for a long time that infant and preschool intelligence measurement is less stable than measures taken later in childhood (Anderson, Citation1939; Bayley, Citation1933). Similar to our results, intelligence has been reported to be relatively stable between 7 years of age and early adolescence (Moffitt, Caspi, Harkness, & Silva, Citation1993).
All childhood cognitive ability/intelligence measurements significantly correlated with educational attainment by the late 20s. Intelligence at 7 years was the only significant predictor within the model, and given cognitive and intelligence scores were highly correlated, this means that age seven intelligence explained most variance (note: children in this cohort would have started school at 5 years). The reason that age 7, over other ages, carried this weight may reflect slight variations in influences from the environment or age‐related genetic expression, as intelligence has an inherited component (Benyamin et al., Citation2013) that overlaps with educational achievement (Calvin et al., Citation2012). As a child ages, environmental influences on intelligence decrease and genetic influences increase (Bartels, Rietveld, Van Baal, & Boomsma, Citation2002), and the underlying stability of intelligence over childhood is driven by genetics (Bartels et al., Citation2002; Petrill et al., Citation2004).
It also may be that case that cognitive abilities measured at 2 and 4 years were not significant predictors of educational attainment because of the aforementioned measurement instability or that these measures taken prior to school entry do not relate to educational attainment as well as those taken after school entry (i.e., there is an interaction between intelligence and educational experience). The effect for early adolescent intelligence (i.e., in association with educational attainment) did approach conventional significance levels (standardised beta = .168, standard error = .102, p = .099). It is of note that social disadvantage leads to a reduction in intelligence between 6 and 11 years at a magnitude similar to this cohort (note average intelligence dropped from an average of 104 at 7 years to 100 in early adolescence in this cohort; see Table ; Breslau et al., Citation2001). There is a possibility that social disadvantages manifest in the regional industrial town cohort setting led to a drop in measured intelligence in adolescence, which meant that the earlier intelligence measurement at age 7 more closely reflected mental capacity and therefore later education attainment.
The influence of paternal factors on child intelligence are at odds with Johnson et al. (Citation2010) who reported that paternal socioeconomic status was positively related, and paternal educational attainment negatively related, to childhood intelligence at 11 years of age. In the subregression looking at predictors of adolescent intelligence in this sample, paternal (nor maternal) factors had an effect. Although generational, geographical, and cultural differences likely drove these differential findings in part (Johnson et al., Citation2010 employed the 1936 Scottish Lothian Birth Cohort), this difference is likely driven by our study including age 7 intelligence, which accounted for the majority of variance. Our results were somewhat in line with Lager et al. (Citation2012), who reported that paternal education was related to the educational attainment of the child at 20 years. However, Lager et al. (Citation2012) also reported paternal education was related to the intelligence of the child at 10 and 20 years. We found no relationships between paternal education and childhood cognitive ability/intelligence (in full model), rather only relationships between maternal intelligence—which Lager et al. (Citation2012) was unable to include in their model. Maternal schooling did not appear to affect educational attainment in the full model; however, maternal intelligence was strongly associated with child cognitive ability/intelligence (Bayley, Citation1955; Breslau et al., Citation2001; Der et al., Citation2006), with some attenuation in this relationship in adolescence.
The relationship between childhood cognitive function/intelligence and later educational attainment is obviously more complex than our model may suggest. It is likely that non‐cognitive factors play a large role in the intergenerational transmission of social advantages (Ackerman, Chamorro‐Premuzic, & Furnham, Citation2011; Furnham & Moutafi, Citation2012; Gale, Batty, & Deary, Citation2008; Pearson et al., Citation2011; Sanson et al., Citation2011; Slominski, Sameroff, Rosenblum, & Kasser, Citation2011). However, in a large New Zealand study, the positive relationship between intelligence at 7–9 years of age and later education achievement held when co‐varying for early conduct problems and family, social, and childhood circumstances (Fergusson, John Horwood, & Ridder, Citation2005).
This study is not without limitations. The two earliest measures of childhood cognitive development (i.e., at 2 years the Bayley Mental Development Index and at 4 years the McCarthy Scales of Children's Abilities) were not the same, nor the same as those employed at later assessments (WISC‐R). At the time of data collection (1980s), the Wechsler intelligence test for pre‐schoolers (the Wechsler Preschool and Primary Scale of Intelligence) was only valid from 4 and a half years; therefore, the Bayley and McCarthy scales were employed, which are measures of general cognitive ability and not intelligence per se. These scales however have shown good correlations with other standardised intelligence tests (Bayley, Citation1969; Hack et al., Citation2005), and the Bayley Scale has also shown good predictive validity in terms of later intelligence score, if taken after 18 months (Kopp & McCall, Citation1982), and our own results show that the McCarthy Scale score at 5 years was significantly related to the WISC‐R at 7 years.
Another limitation is that educational attainment was only ascertained in the late 20s (25–29 years of age), and at the time of data collection 14% of the sample were still in full‐time education and 5% in part‐time education. However, the timing of data collection is also a positive, as education attainment is usually measured in late adulthood in large cohort studies, which introduces sample biases (e.g., survivor effect) and issues with recall reliability. We also measured educational attainment as the number of years of school and tertiary studies completed, we did not measure academic performance/achievement (e.g., test scores), where relationships may be different (Deary, Strand, Smith, & Fernandes, Citation2007).
Our findings provide theoretical support for childhood intelligence and educational attainment having independent effects on later morbidity and mortality, in that we have shown that do not relate to one underlying mental‐ability construct. For example, it has been reported that educational attainment and not intelligence is associated with smoking while pregnant (Gale et al., Citation2009) and midlife obesity (Lawlor et al., Citation2006), and that intelligence is more strongly associated with dementia in late life than education (Schmand, Smit, Geerlings, & Lindeboom, Citation1997). Further, cognitive function in late life is independently related to both early adulthood intelligence and educational attainment (Plassman et al., Citation1995).
The number of years of education a child completed appeared most closely related to the number of years of high school that their father completed and their own intelligence at age 7. These findings demonstrate that when taking into account parental factors, the relationship between childhood intellect and education attainment by the mid‐20s is age dependent, with age 7 intelligence the best predictor. The relationships between parental factors and child intelligence/cognitive ability and educational attainment were complex. Maternal intelligence was strongly related to the child's early life cognitive ability and intelligence, but not to the child's educational attainment in adulthood—the reverse was true for paternal schooling. These findings highlight the complex intergenerational transmission of social advantages and mechanisms for the independent effects of education and intelligence on later morbidly and mortality.
References
- Ackerman, P. L., Chamorro‐premuzic, T., & Furnham, A. (2011). Trait complexes and academic achievement: Old and new ways of examining personality in educational contexts. British Journal of Educational Psychology, 81(1), 27–40.
- Anderson, J. E. (1939). The limitations of infant and preschool tests in the measurement of intelligence. Journal of Psychology, 8, 351–373.
- Baghurst, P. A., Oldfield, R., Wigg, N., Mcmichael, A., Robertson, E., & Vimpani, G. (1985). Some characteristics and correlates of blood lead in early childhood: Preliminary results from the Port Pirie Study. Environmental Research, 38(1), 24–30.
- Baghurst, P. A., Mcmichael, A. J., Wigg, N. R., Vimpani, G. V., Robertson, E. F., Roberts, R. J., & Tong, S.‐L. (1992). Environmental exposure to lead and children's intelligence at the age of seven years. New England Journal of Medicine, 327(18), 1279–1284.
- Bartels, M., Rietveld, M. J. H., Van baal, G. C. M., & Boomsma, D. I. (2002). Genetic and environmental influences on the development of intelligence. Behavioral Genetics, 32(4), 237–249.
- Batty, G. D., Kivimaki, M., & Deary, I. J. (2010). Intelligence, education, and mortality. British Medical Journal, 340, c563.
- Bayley, N. (1933). Mental growth during the first three years. Genetic Psychology Monographs, 14, 1–92.
- Bayley, N. (1955). On the growth of intelligence. American Psychologist, 10(12), 805–818.
- Bayley, N. (1969). Bayley scales of infant development. New York: The Psychological Corporation.
- Benyamin, B., Pourcain, B., Davis, O. S., Davies, G., Hansell, N. K., Brion, M. J., … Visscher, P. M. (2013). Childhood intelligence is heritable, highly polygenic and associated with FNBP1L. Molecular Psychiatry, 19, 253–258.
- Bornstein, M. H., Hahn, C.‐S., Bell, C., Haynes, O. M., Slater, A., Golding, J., … ALSPAC Study Team. (2006). Stability in cognition across early childhood: A developmental cascade. Psychological Science, 17(2), 151–158.
- Breslau, N., Chilcoat, H. D., Susser, E. S., Matte, T., Liang, K.‐Y., & Peterson, E. L. (2001). Stability and change in children's intelligence quotient scores: A comparison of two socioeconomically disparate communities. American Journal of Epidemiology, 154(8), 711–717.
- Calvin, C., Deary, I., Webbink, D., Smith, P., Fernandes, C., Lee, S., & Visscher, P. (2012). Multivariate genetic analyses of cognition and academic achievement from two population samples of 174,000 and 166,000 school children. Behavioral Genetics, 42(5), 699–710.
- Chandola, T., Deary, I. J., Blane, D., & Batty, G. D. (2006). Childhood IQ in relation to obesity and weight gain in adult life: The National Child Development (1958) Study. International Journla of Obesity, 30(9), 1422–1432.
- Clouston, S. A., Kuh, D., Herd, P., Elliott, J., Richards, M., & Hofer, S. M. (2012). Benefits of educational attainment on adult fluid cognition: International evidence from three birth cohorts. International Journal of Epidemiology, 41(6), 1729–1736.
- Daniel, A. (1984). The measurement of social class. Community Health Studies, 8(2), 218–222.
- Deary, I. J. (2012). Intelligence. Annual Review of Psychology, 63(1), 453–482.
- Deary, I. J., & Johnson, W. (2010). Intelligence and education: Causal perceptions drive analytic processes and therefore conclusions. International Journal of Epidemiology, 39(5), 1362–1369.
- Deary, I. J., Strand, S., Smith, P., & Fernandes, C. (2007). Intelligence and educational achievement. Intelligence, 35(1), 13–21.
- Der, G., Batty, G. D., & Deary, I. J. (2006). Effect of breast feeding on intelligence in children: Prospective study, sibling pairs analysis, and meta‐analysis. British Medical Journal, 333(7575), 945.
- EClipSE. (2010). Education, the brain and dementia: Neuroprotection or compensation? Brain: A Journal of Neurology, 133(8), 2210–2216.
- Fergusson, D. M., John horwood, L., & Ridder, E. M. (2005). Show me the child at seven II: Childhood intelligence and later outcomes in adolescence and young adulthood. Journal of Child Psychology and Psychiatry, 46(8), 850–858.
- Furnham, A., & Moutafi, J. (2012). Personality, age, and fluid intelligence. Australian Journal of Psychology, 64, 128–137.
- Gale, C., Batty, G. D., & Deary, I. J. (2008). Locus of control at age 10 years and health outcomes and behaviors at age 30 years: The 1970 British Cohort Study. Psychosomatic Medicine, 70(4), 397–403.
- Gale, C., Johnson, W., Deary, I., Schoon, I., & Batty, G. (2009). Intelligence in girls and their subsequent smoking behaviour as mothers: The 1958 National Child Development Study and the 1970 British Cohort Study. International Journal of Epidemiology, 38(1), 173–181.
- Hack, M., Taylor, H. G., Drotar, D., Schluchter, M., Cartar, L., Wilson‐costello, D., … Marrow, M. (2005). Poor predictive validity of the Bayley scales of infant development for cognitive function of extremely low birth weight children at school age. Pediatrics, 116(2), 333–341.
- Hagger‐johnson, G., Batty, G. D., Deary, I. J., & von Stumm, S. (2011). Childhood socioeconomic status and adult health: Comparing formative and reflective models in the Aberdeen Children of the 1950s Study (prospective cohort study). Journal of Epidemiology and Community Health, 65(11), 1024–1029.
- Johnson, W., Gow, A. J., Corley, J., Starr, J. M., & Deary, I. J. (2010). Location in cognitive and residential space at age 70 reflects a lifelong trait over parental and environmental circumstances: The Lothian Birth Cohort 1936. Intelligence, 38(4), 402–411.
- Kopp, C. B., & Mccall, R. B. (1982). Predicting later mental performance for normal, at‐risk, and handicapped infants. Life‐span Development and Behavior, 4, 33–61.
- Lager, A., Bremberg, S., & Vågerö, D. (2009). The association of early IQ and education with mortality: 65 year longitudinal study in Malmö (p. 339). Sweden: British Medical Journal.
- Lager, A., Modin, B. E., De stavola, B. L., & Vågerö, D. H. (2012). Social origin, schooling and individual change in intelligence during childhood influence long‐term mortality: A 68‐year follow‐up study. International Journal of Epidemiology, 41(2), 398–404.
- Lawlor, D., Clark, H., Davey smith, G., & Leon, D. A. (2006). Childhood intelligence, educational attainment and adult body mass index: Findings from a prospective cohort and within sibling‐pairs analysis. International Journal of Obesity, 30(12), 1758–1765.
- Lawlor, D., David, B., Clark, H., Mcintyre, S., & Leon, D. (2008). Association of childhood intelligence with risk of coronary heart disease and stroke: Findings from the Aberdeen Children of the 1950s cohort study. European Journal of Epidemiology, 23(10), 695–706.
- Mccall, R. B. (1977). Childhood IQ's as predictors of adult educational and occupational status. Science, 197(4302), 482–483.
- Mcfarlane, A. C., Searle, A. K., Van hooff, M., Baghurst, P. A., Sawyer, M. G., Galletly, C., … Clark, L. S. (2013). Prospective associations between childhood low‐level lead exposure and adult mental health problems: The Port Pirie cohort study. Neurotoxicology, 39(0), 11–17.
- Mcmichael, A. J., Baghurst, P. A., Robertson, E. F., Vimpani, G. V., & Wigg, N. R. (1985). The Port Pirie cohort study. Blood lead concentrations in early childhood. Medical Journal of Australia, 143(11), 499–503.
- Mcmichael, A. J., Baghurst, P. A., Wigg, N. R., Vimpani, G. V., Robertson, E. F., & Roberts, R. J. (1988). Port Pirie Cohort Study: Environmental exposure to lead and children's abilities at the age of four years. New England Journal of Medicine, 319(8), 468–475.
- Mcarthy, D. (1972). Manual for the McCarthy scales of children's abilities. New York: Psychological Corporation.
- Moffitt, T. E., Caspi, A., Harkness, A. R., & Silva, P. A. (1993). The natural history of change in intellectual performance: Who changes? How much? Is it meaningful? Child Psychology & Psychiatry & Allied Discipline, 34(4), 455–506.
- Paile‐hyvärinen, M., Kajantie, E., Räikkönen, K., Henriksson, M., Leskinen, J., Laaksonen, I., … Eriksson, J. G. (2009). Intellectual ability in early adulthood and type 2 diabetes in later life. Acta Diabetologica, 46(3), 249–252.
- Pearson, R. M., Heron, J., Melotti, R., Joinson, C., Stein, A., Ramchandani, P. G., & Evans, J. (2011). The association between observed non‐verbal maternal responses at 12 months and later infant development at 18 months and IQ at 4 years: A longitudinal study. Infant Behavior and Development, 34(4), 525–533.
- Petrill, S. A., Hewitt, J. K., Cherny, S. S., Lipton, P. A., Plomin, R., Corley, R., & Defries, J. C. (2004). Genetic and environmental contributions to general cognitive ability through the first 16 years of life. Developmental Psychology, 40(5), 805–812.
- Plassman, B. L., Welsh, K. A., Helms, M., Brandt, J., Page, W. F., & Breitner, J. C. S. (1995). Intelligence and education as predictors of cognitive state in late life: A 50‐year follow‐up. Neurology, 45(8), 8 1446–8 1450.
- Richards, M., Black, S., Mishra, G., Gale, C. R., Deary, I. J., & Batty, D. G. (2009). IQ in childhood and the metabolic syndrome in middle age: Extended follow‐up of the 1946 British Birth Cohort Study. Intelligence, 37(6), 567–572.
- Sanson, A., Smart, D., & Misson, S. (2011). Children's socio‐emotional, physical, and cognitive outcomes: Do they share the same drivers? Australian Journal of Psychology, 63, 56–74.
- Schmand, B., Smit, J. H., Geerlings, M. I., & Lindeboom, J. (1997). The effects of intelligence and education on the development of dementia. A test of the brain reserve hypothesis. Psychological Medicine, 27(06), 1337–1344.
- Slominski, L., Sameroff, A., Rosenblum, K., & Kasser, T. (2011). Longitudinal predictors of adult socioeconomic attainment: The roles of socioeconomic status, academic competence, and mental health. Development and Psychopathology, 23(01), 315–324.
- Strenze, T. (2007). Intelligence and socioeconomic success: A meta‐analytic review of longitudinal research. Intelligence, 35(5), 401–426.
- Tong, S., Baghurst, P., Mcmichael, A., Sawyer, M., & Mudge, J. (1996). Lifetime exposure to environmental lead and children's intelligence at 11–13 years: The Port Pirie cohort study. British Medical Journal, 312(7046), 1569–1575.
- Tong, S., Baghurst, P. A., Sawyer, M. G., Burns, J., & Mcmichael, A. J. (1998). Declining blood lead levels and changes in cognitive function during childhood: The Port Pirie Cohort Study. Journal of the American Medical Association, 280(22), 1915–1919.
- Vimpani, G., Mcmichael, A., Robertson, E., & Wigg, N. (1985). The Port Pirie Study: A prospective study of pregnancy outcome and early childhood growth and development in a lead‐exposed community—Protocol and status report. Environmental Research, 38(1), 19–23.
- Wechsler, D. (1974). Wechsler Intelligence Scale for Children‐Revised. New York: Psychological Corporation.
- Wechsler, D. (1981). Wechsler Adult Intelligence Scale‐Revised. San Antonio, TX: Psychological Corporation.
- Wilson, R. S., Hebert, L. E., Scherr, P. A., Barnes, L. L., Mendes de leon, C. F., & Evans, D. A. (2009). Educational attainment and cognitive decline in old age. Neurology, 72(5), 460–465.
- Wrulich, M., Brunner, M., Stadler, G., Schalke, D., Keller, U., Chmiel, M., & Martin, R. (2013). Childhood intelligence and adult health: The mediating roles of education and socioeconomic status. Intelligence, 41(5), 490–500.
- Yu, Z. B., Han, S. P., Cao, X. G., & Guo, X. R. (2010). Intelligence in relation to obesity: A systematic review and meta‐analysis. Obesity Reviews, 11(9), 656–670.