Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jul 1.
Published in final edited form as: Soc Sci Med. 2021 May 18;280:114044. doi: 10.1016/j.socscimed.2021.114044

The Effects of Education on Cognition in Older Age: Evidence from Genotyped Siblings

Jason Fletcher 1, Michael Topping 2, Fengyi Zheng 3, Qiongshi Lu 4
PMCID: PMC8205990  NIHMSID: NIHMS1705857  PMID: 34029863

Abstract

A growing literature has sought to tie educational attainment with later-life cognition and Alzheimer’s disease outcomes. This paper leverages sibling comparisons in educational attainment as well as genetic predictors (polygenic scores) for cognition, educational attainment, and Alzheimer’s disease to estimate effects of educational attainment on cognition in older age in the United Kingdom. We find that the effects of education on cognition are confounded by family background factors (~40%) and by genetics (<10%). After adjustments, we continue to find large effects of education. College graduates have cognition scores that are approximately 0.75 SD higher than those who report no credentials. We also find evidence that educational effects on cognition are smaller for those with high polygenic scores for Alzheimer’s disease.

Keywords: Education, Old Age Cognition, Gene-Environment Interactions: Sibling Fixed Effects, UK Biobank

Introduction

Cognitive health is one of the most critical aspects of health as people begin to enter the later stages of life (Evans et al., 2018). Individuals face subtle changes to how they function as they age, along with alterations in their overall cognitive ability. There is a great deal of variation in how cognition levels and trajectories vary in old age. Some people experience and retain a high level of cognition in older age, whereas others in late life experience a decline (Gow et al., 2007). This has many implications because the cognitive ability that a person has in their older age has the potential to influence their autonomy, well-being, and quality of life (Díaz-Venegas et al., 2019). As populations continue to get older, it is imperative to look at specific indicators of cognition in later life, and consider how early life factors, such as educational attainment, continue to shape population-level cognition and health in old age.

There is a myriad of social, economic, and health-related consequences for older adults tied to their cognition. Failing cognition could lead to increased dependence on other people to complete tasks that were once simple (Gill et al., 2010), increased costs of potential treatments or adjustments to lifestyle practices (Vossius et al., 2011), and lead to disability, increased hospital visits, other morbidities and even death (Robertson, Savva, & Kenny, 2013). Research has shown that cognitive decline may be indicative for early phases of dementia or Alzheimer’s disease in later life, thus leading to greater consequences (Ritchie et al., 2016).

Cognition in later life is influenced by a multitude of predictors throughout many stages of the life course. Level of cognition has also been positively influenced by lack of illness and mental health problems (Ritchie et al., 2016), greater amounts of exercise and physical activity throughout life (Sofi et al., 2011), being in a relationship with a significant other (Boss, Kang, & Branson, 2015), specific personality traits (Luchetti et al., 2016), more favorable neighborhood contexts (Ailshire, Karraker, & Clarke, 2017), and number of social connections (Evans et al., 2018; Kuiper et al., 2016), among many other factors. Conversely, cognition has been negatively associated with greater prevalence of illness (Ritchie et al., 2016), depression (van den Kommer et al., 2013; Zaninotto et al., 2018), lack of exercise or relationships (Sofi et al., 2011; Boss et al., 2015), disadvantaged neighborhood contexts (Ailshire et al., 2017), and age (Plassman et al. 2010). All these factors have implications not just for cognition, but other outcomes that involve health as well. However, despite all these factors that have the potential to influence cognition in later stages of the life course, one that gets the most attention is the role of education.

The influence of education in later life outcomes is well established in many disciplines and is the focus of many health outcomes (Galama, Lleras-Muney, & Kippersluis, 2018). People with higher levels of education tend to be far healthier on average, have greater longevity, and have fewer instances of morbidity throughout the life course (Davies et al., 2018; Lager & Torssander, 2012). Moreover, education is a key factor in the social mobility of an individual, which directly has the potential to impact health across the life span. Simandan (2018) specifically highlights the role of education in ones future identity, and its impact on social mobility, which has implications for one’s mental health in later life. This extends to the association that education has with cognition. Despite this, there is a debate about the relationship of education with many outcomes in later life, particularly if education influences these outcomes, or if the outcomes are a result of behavioral or genetic differences. While education has been shown to have positive effects on economic outcomes later in the life course, chiefly income (Clouston et al., 2012), there is a question of how strong the role is regarding cognition.

Research on educational attainment in early life and later cognition is vast (Chen et al. 2019; Clouston et al., 2012; Clouston et al., 2020; Lövdén et al., 2020; Schneewei, Skirbekk, & Winter-Ebmer, 2014; Wilson et al., 2019). An individual’s cognitive ability plays a vital role in the quality of life towards the end of the life course, along with other outcomes such as lifestyle choices and social interactions, which makes understanding this relationship critical (Langa et al., 2009; Ritchie et al., 2016). However, there have been some disputes regarding the determinants of cognition among adults in later life, particularly when it comes to the role of education. Some studies have shown that educational attainment is positively associated with cognition in later life (Richards & Hatch, 2011; Schneewei, et al., 2014; Yount, 2008). Chen and colleagues (2019) found that higher education among individuals in early life served as a defensive factor in aging, which in turn helped delay potential cognitive decline later on (Stern, 2009; Stern 2012). Conversely, other research has shown that education is associated with cognitive level in later life, but not consistently associated with change in cognition (Glymour, Tzourio, & Dufouil, 2012; Zahodne, et al., 2011).

A major component of this debate in educational attainment’s role in cognition is due to whether education is a correlational or causal factor (Schneewei et al., 2014). For example, the association between these two factors could be a result of reverse causation, where high levels of cognitive ability in early life may influence educational attainment. However, education is an outcome that could be influenced by a myriad of unobserved characteristics, which in turn may potentially impact cognition in later life. A large literature has focused on the causal role education plays in childhood cognition, which ultimately could influence these same characteristics later in life (Clouston et al., 2020; Davies et al., 2018; Deary & Johnson, 2010; Richards & Sacker, 2011). Furthermore, there are several suggestions that higher levels of education in life not only affect cognitive ability but also may attenuate aging-associated declines in cognition, although there is doubt regarding the latter (Lövdén et al., 2020).

Research on policies and laws that are implemented at key points in the life course has been shown to have causal effects on later-life cognition. Banks and Mazzonna (2012) studied when compulsory school laws were implemented, specifically the age at which students could leave school. They found that educational attainment increases memory test scores in old age for students and, additionally, executive functioning in male students. Similarly, Glymour and colleagues (2008) discovered that increases in mandatory schooling of students led to improvements in performance on cognition tests, particularly memory scores, decades after they completed their schooling. However, they do point out that unobserved genetic variation may be unlikely to account for these results. Contrarily, previous work has outlined educational attainment as a symptom of pre-existing intelligence in life, and that genetic differences are dominant in these outcomes, opposed to them purely being consequences of social determinants (Gottfredson, 2005; Gottfredson, 2011). Other studies have shown that higher educational attainment can aid in coping with age-related brain deterioration, which allows people to better handle cognitive tasks in later life (Lenehan et al., 2015).

In addition to issues of reverse causality, potential confounding between cognition and educational attainment has also been raised as an important empirical issue. One typical exploration of the role of confounding due to family background and genetics is to compare siblings. Data limitations have not allowed this technique in general, although some studies have examined the question in early life. These studies have shown that genetics appears to explain a large share of the similarity in cognition between siblings (Moorman, Carr, & Greenfield, 2018; Reynolds et al., 2014). Other research has studied sibling differences with regard to educational attainment and cognition but focused on how education is influenced by cognition, not the inverse (Polderman et al., 2015). Differential investments by parents may also lower levels of sibling similarity (Baier, 2019).1

This paper seeks to examine how sibling differences in educational attainment are related to differences in cognition in older age. This design, which is novel in the study of old age cognition, allows family background and shared genetics to be controlled. We add measured genetics that is unshared between siblings to the sibling difference design to further probe the possibility of genetic confounding between educational attainment and later-life cognition.

Data

We used data from the UK Biobank (UKB) project. The participants, aged between 37 and 74 years, were originally recruited between 2006 and 2010.2 Cognition (fluid intelligence3) is measured in the UKB by summing the number of 13 logic puzzles that the participants could answer correctly in two minutes (Davies et al. 2018). This measure is only available for ~200,000 of the 500,000 UKB participants, as it was added as a module toward the end of the recruitment window.

Educational levels of the UK Biobank participants were measured by mapping each major educational qualification that can be identified from the survey measures to an International Standard Classification of Education (ISCED) category and imputing a years-of-education equivalent for each ISCED category (Lee et al., 2018)4.

Although siblings are not identified in the survey, respondents’ genetics can be used to measure genetic relatedness among all pairs of respondents. We first use the UKB provided kinship file, listing all pairwise kinships among 100,000 pairs in the sample of nearly 500,000 individuals. We first choose pairs with kinship >0.2, which reflects first-degree biological relatives (parents/siblings). We then choose the remaining pairs who are <13 years apart in age, leaving ~22,000 sibling dyads. We then chose to keep only one dyad from any family with more than one dyad, leaving ~17,600 dyads. The number of dyads who also have non-missing cognition scores (only available for ~1/3 of the full UKB sample) and educational attainment information is 4,138 (8,276 respondents). We include only respondents of European ancestry in our analysis.

We constructed polygenic scores (PGS) for three traits for which large genome-wide association studies (GWAS) are publicly available and do not contain UK Biobank samples: Alzheimer’s disease (Kunkle et al. 2019), cognition (Rietveld et al. 2014), and educational attainment (Lee et al., 2018). These three studies had sample sizes of 63,926, 106,736, and 324,162, respectively. Summary statistics for AD and cognition are publicly available while the summary statistics for EA with UKB and 23&me samples excluded was obtained from authors of (Lee et al., 2018). We used consistent pipelines to process these summary data. We used 1000 Genomes Project Phase III European samples as a reference for linkage disequilibrium (LD) and LD-clumped all GWAS summary statistics using PLINK (Purcell et al 2007). We used an LD window size of 1Mb and a pairwise r2 threshold of 0.1. In addition, strand-ambiguous SNPs and SNPs that do not exist in the UKB imputed genotype data were removed. After processing, 110,639 SNPs were included in the AD score, 73,167 SNPs remained in the cognition score, and 83,516 SNPs were in the EA score. No additional filtering based on GWAS p-values was applied (Choi & O’Reilly, 2019)5. All three scores produced by PRSice-2 are normally distributed. PGS of each trait was standardized to have a mean of zero and a standard deviation of one. We also control for 20 genetic principal components in our analysis provided by the UK Biobank (Fletcher et al. 2020).

Table 1 presents descriptive statistics for our sample. Panel 1 shows the sample in the UK Biobank who have European ancestry, showing over 400,000 observations and 34,000 observations of siblings. Panel 2 limits this sample to those with a cognitive test available, where we have over 160,000 observations including 13,000 observations of individuals who have a sibling in the sample. Panel 3 restricts the sample to those sibling pairs where each member also has a cognitive test as well as educational attainment information available, providing nearly 8,300 observations. Comparing our analysis sample to the full sample, we see a high level of similarity among the variables in our analysis.

Table 1.

Summary Statistics

Variable Obs Mean Std Dev Min Max
Sample European Ancestry
Cognition 164,150 6.3 2.1 0.0 13.0
Education 404,436 13.8 5.1 7.0 20.0
EA-PGS (std) 408,248 0.0 1.0 −4.7 5.0
AD-PGS (std) 408,248 0.0 1.0 −5.0 4.6
Cog-PGS (std) 408,248 0.0 1.0 −4.9 4.8
Age 408,248 56.9 8.0 39.0 73.0
Female 408,248 0.5 0.5 0.0 1.0
Family ID 34,516
Sample: Available Cognitive Test
Cognition 162,812 6.3 2.1 0.0 13.0
Education 162,812 14.3 5.0 7.0 20.0
EA-PGS (std) 162,812 0.0 1.0 −4.1 4.6
AD-PGS (std) 162,812 0.0 1.0 −5.0 4.5
Cog-PGS (std) 162,812 0.0 1.0 −4.9 4.3
Age 162,812 56.9 7.9 40.0 73.0
Female 162,812 0.5 0.5 0.0 1.0
Family ID 13,397
Sample: Sibling Pairs
Cognition 8,276 6.2 2.1 0.0 13.0
Education 8,276 13.8 5.0 7.0 20.0
No Qualifications / Years=7 8,276 0.2 0.4 0.0 1.0
CSE/O-Levels / Years =10 8,276 0.3 0.5 0.0 1.0
A/AS Levels / Years = 13 8,276 0.1 0.3 0.0 1.0
Other Professional / Years =15 8,276 0.1 0.2 0.0 1.0
NVQ/HNC / Years =19 8,276 0.1 0.3 0.0 1.0
College Degree / Years =20 8,276 0.3 0.5 0.0 1.0
EA-PGS (std) 8,276 0.0 1.0 −3.7 3.9
AD-PGS (std) 8,276 0.0 1.0 −3.7 4.0
Cog-PGS (std) 8,276 0.0 1.0 −4.7 3.8
Age 8,276 57.5 7.2 40.0 70.0
Female 8,276 0.6 0.5 0.0 1.0
Family ID 8,276
Open in a new tab

Notes: Family ID refers to the family identification number of each respondent and is missing for those without biological siblings in the UKB.

Education is coded as follows: no qualifications= 7 years of education; CSEs or equivalent = 10 years; O levels/GCSEs or equivalent = 10 years; A levels/AS levels or equivalent = 13 years; other professional qualification = 15 years; NVQ or HNC or equivalent = 19 years; college or university degree = 20 years of education

Results

Table 2 presents our main results. We include separate indicator variables for each educational category, with “no credential” as the omitted group. Column 1 shows the associations for the sample of respondents with European ancestry and non-missing cognition scores and allows a comparison with our analysis sample of siblings. The results are nearly identical and suggest large differences in cognition by educational category. For example, respondents with A-level credentials are over 2.3 points higher (1 SD) than those without credentials. Column 3 adds PGS for educational attainment, Alzheimer’s disease, and cognition. One standard deviation in the AD-PGS is associated with a 0.08 point reduction in cognition. Notably, these genetic controls only reduce the associations between educational attainment and cognition by ~5%, suggesting a modest role of genetic confounding.

Table 2.

Main Results

Outcome Sample Fixed Effects? Cognition Full None Cognition Pairs None Cognition Pairs None Cognition Pairs Sibling Cognition Pairs Sibling
Education = 10 1.392*** 1.394*** 1.352*** 0.753*** 0.737***
(0.016) (0.065) (0.065) (0.091) (0.091)
Education = 13 2.318*** 2.365*** 2.287*** 1.373*** 1.343***
(0.019) (0.082) (0.082) (0.119) (0.118)
Education = 15 1.450*** 1.450*** 1.407*** 0.948*** 0.940***
(0.024) (0.103) (0.103) (0.133) (0.133)
Education = 19 0.823*** 0.820*** 0.780*** 0.461*** 0.461***
(0.022) (0.092) (0.091) (0.117) (0.117)
Education = 20 2.743*** 2.710*** 2.600*** 1.535*** 1.493***
(0.015) (0.065) (0.066) (0.104) (0.104)
EA-PGS (std) 0.138*** 0.136***
(0.023) (0.042)
AD-PGS (std) −0.083*** −0.082**
(0.021) (0.037)
Cognition-PGS (std) 0.065*** 0.055
(0.023) (0.042)
Age −0.003*** −0.000 −0.002 −0.013* −0.012*
(0.001) (0.003) (0.003) (0.007) (0.007)
Female −0.229*** −0.274*** −0.276*** −0.250*** −0.252***
(0.009) (0.042) (0.042) (0.056) (0.056)
Constant 4.805*** 4.616*** 4.815*** 6.845*** 6.894***
(0.055) (0.254) (0.254) (0.532) (0.531)
Observations 162,812 8,276 8,276 8,276 8,276
R-squared 0.207 0.212 0.220 0.689 0.690
Number of Pairs 4,138 4,138
Open in a new tab

Standard errors in parentheses. 20 PCs are controlled but not reported

***

p<0.01,

**

p<0.05,

*

p<0.1

Columns 4 and 5 present our sibling difference specifications. In general, the associations between educational attainment and cognition shrink by ~40–50% compared to the results that do not control for sibling fixed effects; adding sibling differences in PGS to these models does not change these results. Overall, the results suggest a considerable level of confounding by family background and genetics but continue to point to remaining large associations between educational attainments and cognition in older age.

We next consider whether there is evidence for gene-environment interactions (GxE) between educational attainment (EA) and later-life cognition (Cog). Table 3 first explores the possibility of selection into education, which would suggest an important gene-environment correlation, where polygenic scores would predict both the “environment” of educational attainment and the old age cognition outcome of interest. Column 1 presents associations predicting educational attainment based on sibling differences in polygenic scores, with controls for age fixed effects, sex, and genetic principal components. Even between siblings, we do find evidence that EA-PGS predicts educational attainment, as found in other studies (Fletcher et al., 2020; Lee et al., 2018). In contrast, we find no association between Alzheimer’s disease (AD)-PGS and educational attainment, which mirrors earlier work focusing on APOE that suggests these variants do not impact decisions early in the life course (Cook & Fletcher, 2015). Thus, the AD-PGS appears to be a reasonable candidate to consider in GxE analysis.

Table 3.

Selection and GxE

Outcome Sample Fixed Effects? Education Pairs Family Cognition Pairs Family Cognition Pairs Family
Fixed Effects
Age Controlled −0.021*** −0.021***
(0.007) (0.007)
Female −0.675*** −0.215*** −0.213***
(0.145) (0.057) (0.057)
Education 0.064*** 0.064***
(0.006) (0.006)
EA-PGS (std) 0.328*** 0.162*** 0.109
(0.104) (0.042) (0.092)
AD-PGS (std) −0.028 0.058 0.058
(0.095) (0.086) (0.086)
Cog-PGS (std) 0.085 0.066 0.159
(0.105) (0.042) (0.098)
Education X AD-PGS −0.010* −0.010*
(0.006) (0.006)
Education X Cog-PGS −0.007
(0.007)
Education X EA-PGS 0.004
(0.006)
Constant 13.840*** 7.288*** 7.286***
(1.181) (0.531) (0.531)
Observations 8,276 8,276 8,276
R-squared 0.656 0.681 0.681
Number of Pairs 4,138 4,138 4,138
Open in a new tab

Robust standard errors in parentheses. 20 PCs are controlled but not reported

***

p<0.01,

**

p<0.05,

*

p<0.1

With the findings from Column 1 suggesting no evidence of gene-environment correlation for the AD-PGS, Columns 2 and 3 examines gene-environment interactions. The results show evidence that the impacts of educational attainment on later-life cognition are reduced for individuals with a higher genetic risk of Alzheimer’s disease (higher AD-PGS scores) and that this finding is not affected by also controlling for interactions between educational attainment and EA-PGS and Cog-PGS. In order to further explore these findings, Table 1A in the appendix examines gene-environment interactions for each level of educational attainment rather than treating attainment as a continuous measure. The results continue to show evidence that individuals with higher AD-PGS have smaller old age cognitive benefits from educational attainment than do individuals with lower AD-PGS.

Discussion

Previous studies have sought to tie the role that educational attainment plays in cognition later in life. However, some of the literature has found mixed results in this relationship, with both positive and negative associations with education (Glymour et al., 2012; Richards & Hatch, 2011; Yount, 2008; Zahodne, et al., 2011). Additionally, research has focused attention on whether the relationship between education and cognition in later life is causal or correlational as a result of social, economic, behavioral, or genetic factors (Clouston et al., 2020; Davies et al., 2018; Schneewei et al., 2014). This paper applies a new research design to these questions by examining how sibling differences, as well as genetic predictors, influence the effects of educational attainment on cognition in old age. The analyses conducted indicate that there are indeed associations between educational attainment and cognition in older age, even controlling for confounding by family background and genetic factors.

There are important limitations to this study. First, the analysis is done with sibling dyads, which does not allow an analysis of individuals with no siblings. Another limitation is that this study does not include an examination of individuals of different racial and ethnic backgrounds. Future research should seek to examine how sibling and genetic differences impact the association between educational attainment and cognition by racial groups, as research has shown that there are racial differences at the population level in cognition and Alzheimer’s disease in later life (Barnes et al., 2015; Howell et al., 2017). A third limitation is the lack of experimental or quasi-experimental variation in educational attainments. We instead leverage sibling differences in educational attainment (controlling for genetic factors) to examine sibling differences in cognition, which allows the possibility of within-sibling confounders6. A fourth limitation is that only a subsample of the UKB has information on cognition, which reduced the power of our analysis, although we still use one of the largest datasets of siblings that measure genetics and cognition in older age. A fifth limitation is that the PGS measures only account for common variants and not rare genetic variants.

Despite the limitations outlined above, this research is the first to use sibling fixed-effect models and genetic predictors to estimate the effects of education on cognition in older age. Results point to important associations between education and cognition in older age. These findings largely support previous work done on how education can aid in coping with age-related deterioration (Lenehan et al., 2015), and thus act as a vehicle to maintain or improve cognition in later life. We also provide novel evidence of an interaction between educational attainment and the PGS for AD in predicting cognition in later life, suggesting those at higher genetic risks receive smaller benefits (between 6% and 54%) from education than those at lower genetic risks. This highlights the need to examine the intersection of both social and genetic factors when studying outcomes in late life.

Highlights.

Explores associations between educational attainment and later life cognition

Uses large scale sibling comparisons in the UK Biobank to control family background

Examines genetic confounding by controlling for polygenic scores

Finds both large confounding and large remaining associations between education and cognition

Evidence that high Alzheimer’s Disease Polygenic Score reduces cognitive benefits of education

Acknowledgments

This project was supported by funding from NIA (AG062765, AG017266). We thank members of the Social Genomics Working Group at the University of Wisconsin for helpful comments. This research has been conducted using the UK Biobank Resource under Application 57284.

Table 1A:

Exploring GxE across Educational Attainment Categories

Outcome Fixed Effects? Sample Cognition Family Pairs Cognition Family Pairs
Age −0.021*** −0.013*
(0.007) (0.007)
Female −0.213*** −0.252***
(0.057) (0.056)
EA-PGS (std) 0.109 0.074
(0.092) (0.080)
AD-PGS (std) 0.058 0.014
(0.086) (0.071)
Cog-PGS (std) 0.159 0.065
(0.098) (0.081)
Education 0.064***
(0.006)
Education X AD-PGS −0.010*
(0.006)
Education X Cog-PGS −0.007
(0.007)
Education X EA-PGS 0.004
(0.006)
Education = 10 0.758***
(0.092)
Education = 10 X AD-PGS −0.058
(0.082)
Education = 13 1.357***
(0.119)
Education = 13 X AD-PGS −0.171
(0.104)
Education = 15 0.975***
(0.135)
Education = 15 X AD-PGS −0.063
(0.117)
Education = 19 0.486***
(0.118)
Education = 19 X AD-PGS −0.262**
(0.114)
Education = 20 1.510***
(0.105)
Education = 20 X AD-PGS −0.130
(0.087)
Education = 10 X Cog-PGS 0.008
(0.092)
Education = 13 X Cog-PGS 0.112
(0.115)
Education = 15 X Cog-PGS −0.150
(0.143)
Education = 19 X Cog-PGS −0.149
(0.125)
Education = 20 X Cog-PGS −0.032
(0.104)
Education = 10 X EA-PGS 0.020
(0.091)
Education = 13 X EA-PGS 0.079
(0.119)
Education = 15 X EA-PGS 0.327**
(0.145)
Education = 19 X EA-PGS 0.062
(0.126)
Education = 20 X EA-PGS 0.096
(0.097)
Constant 7.286*** 6.902***
(0.531) (0.532)
Observations 8,276 8,276
R-squared 0.047 0.078
Number of famid 4,211 4,211
Open in a new tab

Column 1 repeats results from Table 3 in the text. Robust standard errors in parentheses. 20 PCs are controlled but not reported. Age fixed effects controlled but not reported

***

p<0.01,

**

p<0.05,

*

p<0.1

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

A study done by Fors and colleagues (2009) found that people who grew up with many siblings had less educational attainment later in life, but sibling number was not significantly associated with cognition in later life. The study did not look at differences between the siblings themselves.

2

These data are restricted, but one can gain access by following the procedures described in www.ukbiobank.ac.uk/register-apply/.

4

Education is coded as follows: no qualifications= 7 years of education; CSEs or equivalent = 10 years; O levels/GCSEs or equivalent = 10 years; A levels/AS levels or equivalent = 13 years; other professional qualification = 15 years; NVQ or HNC or equivalent = 19 years; college or university degree = 20 years of education

5
In all three summary statistics files, effect alleles were clearly labeled. For each trait, we calculated PGS as:
PGS = ∑i=1MXiβ^i
where Xi is the count of effective alleles of the i-th SNP and β^i is the marginal effect size of the i-th SNP directly obtained from GWAS summary statistics. We note that these effect alleles are not necessarily minor alleles (effect allele frequency can be greater than 0.5) nor risk alleles (β^i can be positive or negative).
6

One example confounder is conscientiousness. This measure has been shown to be a key predictor of both health and non-health related outcomes across the life course, such as educational success and cognition (Lodi-Smith et al., 2010; Noftle & Robins, 2007). Regarding cognition, it has been found that conscientiousness is associated with slower rates of decline and may help compensate for any age differences in cognition (Curtis, Windsor, & Soubelet, 2015; Soubelet, 2011). Due a lack of data we are unable to assess the importance of within-sibling differences in potential confounding factors, thus we acknowledge this potential confounder as well as others and the role they may potentially have when looking at outcomes relating to cognition.

Contributor Information

Jason Fletcher, La Follette School of Public Affairs and Department of Sociology, Center for Demography of Health and Aging, University of Wisconsin-Madison.

Michael Topping, Department of Sociology, Center for Demography of Health and Aging, University of Wisconsin-Madison.

Fengyi Zheng, Department of Biostatistics and Medical Informatics, University of Wisconsin-Madison.

Qiongshi Lu, Department of Biostatistics and Medical Informatics, Center for Demography of Health and Aging, University of Wisconsin-Madison.

References:

  1. Aartsen MJ, Cheval B, Sieber S, Van der Linden BW, Gabriel R, Courvoisier DS, Guessous I, Burton-Jeangros C, Blane D, Ihle A, Kliegel M, & Cullati S (2019). Advantaged socioeconomic conditions in childhood are associated with higher cognitive functioning but stronger cognitive decline in older age. Proceedings of the National Academy of Sciences of the United States of America, 116(12), 5478–5486. 10.1073/pnas.1807679116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ailshire J, Karraker A, & Clarke P (2017). Neighborhood social stressors, fine particulate matter air pollution, and cognitive function among older U.S. adults. Social Science & Medicine, 172, 56–63. 10.1016/j.socscimed.2016.11.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baier T (2019). Does sibling and twin similarity in cognitive ability differ by parents’ education? Journal of Family Research, 31(1). [Google Scholar]
  4. Banks J, & Mazzonna F (2012). The Effect of Education on Old Age Cognitive Abilities: Evidence from a Regression Discontinuity Design. Economic Journal, 122(560), 418–448. 10.1111/j.1468-0297.2012.02499.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Barnes LL, Leurgans S, Aggarwal NT, Shah RC, Arvanitakis Z, James BD, Buchman AS, Bennett DA, & Schneider JA (2015). Mixed pathology is more likely in black than white decedents with Alzheimer dementia. Neurology, 85(6), 528–534. 10.1212/WNL.0000000000001834 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Boss L, Kang D-H, & Branson S (2015). Loneliness and cognitive function in the older adult: a systematic review. International Psychogeriatrics / IPA, 27(4), 541–553. 10.1017/S1041610214002749 [DOI] [PubMed] [Google Scholar]
  7. Chen Y Lv C, Li X, Zhang J, Chen K, Liu Z,…Zhang Z (2019). The positive impacts of early-life education on cognition, leisure activity, and brain structure in healthy aging. Aging, 11(14), 4923–4942. 10.18632/aging.102088 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Choi SW, & O’Reilly PF (2019). PRSice-2: Polygenic Risk Score software for biobank-scale data. GigaScience, 8(7), giz082. 10.1093/gigascience/giz082 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Clouston SA, Kuh D, Herd P, Elliott J, Richards M, & Hofer SM (2012). Benefits of educational attainment on adult fluid cognition: international evidence from three birth cohorts. International Journal of Epidemiology, 41(6), 1729–1736. 10.1093/ije/dys148 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Clouston SA, Smith DM, Mukherjee S, Zhang Y, Hou MW Link BG, & Richards M (2020). Education and cognitive decline: an integrative analysis of global longitudinal studies of cognitive aging. The Journals of Gerontology: Series B, 75(7), e151–e160. 10.1093/geronb/gbz053 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cook CJ, & Fletcher JM (2015). Can education rescue genetic liability for cognitive decline? Social Science & Medicine, 127, 159–170. 10.1016/j.socscimed.2014.06.049 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Curtis RG, Windsor TD, & Soubelet A (2015). The relationship between Big-5 personality traits and cognitive ability in older adults - a review. Neuropsychology, Development, and Cognition. Section B, Aging, Neuropsychology and Cognition, 22(1), 42–71. 10.1080/13825585.2014.888392 [DOI] [PubMed] [Google Scholar]
  13. Davies NM, Dickson M, Smith GD, van den Berg GB, & Windmeijer F (2018). The causal effects of education on health outcomes in the UK Biobank. Natural Human Behaviour, 2(2), 117–125. 10.1038/s41562-017-0279-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Deary IJ, & Johnson W (2010). Intelligence and education: causal perceptions drive analytic processes and therefore conclusions. International Journal of Epidemiology, 39(5), 1362–1369. 10.1093/ije/dyq072 [DOI] [PubMed] [Google Scholar]
  15. Dhuey E, Figlio D, Karbownik K, & Roth J (2019). School Starting Age and Cognitive Development. Journal of Policy Analysis and Management: [the Journal of the Association for Public Policy Analysis and Management], 38(3), 538–578. 10.1002/pam.22135 [DOI] [Google Scholar]
  16. Evans IEM, Llewellyn DJ, Matthews FE, Woods RT, Brayne C, Clare L, & CFAS-Wales research team. (2018). Social isolation, cognitive reserve, and cognition in healthy older people. PloS One, 13(8), e0201008. 10.1371/journal.pone.0201008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fors S, Lennartsson C, & Lundberg O (2009). Childhood living conditions, socioeconomic position in adulthood, and cognition in later life: exploring the associations. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 64(6), 750–757. 10.1093/geronb/gbp029 [DOI] [PubMed] [Google Scholar]
  18. Fletcher JM, Wu Y, Zhao Z, & Lu Q (2020). The production of within-family inequality: Insights and implications of integrating genetic data. In Genetics (No. biorxiv;2020.06.06.137778v1; p. 1883). bioRxiv. 10.1101/2020.06.06.137778 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Galama T, Lleras-Muney A, & Kippersluis HV (2018). The Effect of Education on Health and Mortality: A Review of Experimental and Quasi-Experimental Evidence. Oxford Research Encyclopedia of Economics and Finance. 10.1093/acrefore/9780190625979.013.7 [DOI] [Google Scholar]
  20. Gill TM, Gahbauer EA, Han L, & Allore HG (2010). Trajectories of Disability in the Last Year of Life. The New England Journal of Medicine, 362(13), 1173–1180. 10.1056/NEJMoa0909087 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Glymour MM, Kawachi I, Jencks CS, & Berkman LF (2008). Does childhood schooling affect old age memory or mental status? Using state schooling laws as natural experiments. Journal of Epidemiology and Community Health, 62(6), 532–537. 10.1136/jech.2006.059469 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Glymour MM, Tzourio C, & Dufouil C (2012). Is cognitive aging predicted by one’s own or one’s parents’ educational level? Results from the three-city study. American Journal of Epidemiology, 175(8), 750–759. 10.1093/aje/kwr509 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gottfredson LS (2005). What if the hereditarian hypothesis is true? Psychology, Public Policy, and Law: An Official Law Review of the University of Arizona College of Law and the University of Miami School of Law, 11(2), 311–319. 10.1037/1076-8971.11.2.311 [DOI] [Google Scholar]
  24. Gottfredson LS (2011). Intelligence and social inequality: Why the biological link? In The Wiley-Blackwell Handbook of Individual Differences (pp. 538–575). Wiley-Blackwell. 10.1002/9781444343120.ch20 [DOI] [Google Scholar]
  25. Gow AJ, Pattie A, Whiteman MC, Whalley LJ, & Deary IJ (2007). Social support and successful aging: Investigating the relationships between lifetime cognitive change and life satisfaction. Journal of Individual Differences, 28(3), 103–115. 10.1027/1614-0001.28.3.103 [DOI] [Google Scholar]
  26. Howell JC, Watts KD, Parker MW, Wu J, Kollhoff A, Wingo TS, Dorbin CD, Qiu D, & Hu WT (2017). Race modifies the relationship between cognition and Alzheimer’s disease cerebrospinal fluid biomarkers. Alzheimer’s Research & Therapy, 9(1), 88. 10.1186/s13195-017-0315-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kuiper JS, Zuidersma M, Zuidema SU, Burgerhof JG, Stolk RP, Oude Voshaar RC, & Smidt N (2016). Social relationships and cognitive decline: a systematic review and meta-analysis of longitudinal cohort studies. International Journal of Epidemiology, 45(4), 1169–1206. 10.1093/ije/dyw089 [DOI] [PubMed] [Google Scholar]
  28. Kunkle BW, Grenier-Boley B, Sims R, Bis JC, Damotte V, Naj AC, Boland A, Vronskaya M, Van Der Lee SJ, Amlie-Wolf A and Bellenguez C, 2019. Genetic meta-analysis of diagnosed Alzheimer’s disease identifies new risk loci and implicates Aβ, tau, immunity and lipid processing. Nature genetics, 51(3), pp.414–430. 10.1038/s41588-019-0358-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lee JJ, Wedow R, Okbay A, Kong E, Maghzian O, Zacher M, Nguyen-Viet TA, Bowers P, Sidorenko J, Karlsson Linnér R, Fontana MA, Kundu T, Lee C, Li H, Li R, Royer R, Timshel PN, Walters RK, Willoughby EA, … Social Science Genetic Association Consortium. (2018). Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. Nature Genetics, 50(8), 1112–1121. 10.1038/s41588-0180147-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lenehan ME, Summers MJ, Saunders NL, Summers JJ, & Vickers JC (2015). Relationship between education and age-related cognitive decline: a review of recent research. Psychogeriatrics: The Official Journal of the Japanese Psychogeriatric Society, 15(2), 154–162. 10.1111/psyg.12083 [DOI] [PubMed] [Google Scholar]
  31. Lodi-Smith J, Jackson J, Bogg T, Walton K, Wood D, Harms P, & Roberts BW (2010). Mechanisms of health: Education and health-related behaviours partially mediate the relationship between conscientiousness and self-reported physical health. Psychology & Health, 25(3), 305–319. 10.1080/08870440902736964 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lövdén M, Fratiglioni L, Glymour MM, Lindenberger U, & Tucker-Drob EM (2020). Education and Cognitive Functioning Across the Life Span. Psychological Science in the Public Interest: A Journal of the American Psychological Society, 21(1), 6–41. 10.1177/1529100620920576 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Luchetti M, Terracciano A, Stephan Y, & Sutin AR (2016). Personality and Cognitive Decline in Older Adults: Data from a Longitudinal Sample and Meta-Analysis. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 71(4), 591–601. 10.1093/geronb/gbu184 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Meng X, & D’Arcy C (2012). Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta-analyses and qualitative analyses. PloS One, 7(6), e38268. 10.1371/journal.pone.0038268 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Moorman SM, Carr K, & Greenfield EA (2018). Childhood socioeconomic status and genetic risk for poorer cognition in later life. Social Science & Medicine, 212, 219–226. 10.1016/j.socscimed.2018.07.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Noftle EE, & Robins RW (2007). Personality predictors of academic outcomes: big five correlates of GPA and SAT scores. Journal of Personality and Social Psychology, 93(1), 116–130. 10.1037/0022-3514.93.1.116 [DOI] [PubMed] [Google Scholar]
  37. Panza F, Lozupone M, Solfrizzi V, Sardone R, Dibello V, Di Lena L, D’Urso F, Stallone R, Petruzzi M, Giannelli G, Quaranta N, Bellomo A, Greco A, Daniele A, Seripa D, & Logroscino G (2018). Different Cognitive Frailty Models and Health- and Cognitive-related Outcomes in Older Age: From Epidemiology to Prevention. Journal of Alzheimer’s Disease, 62(3), 993–1012. 10.3233/JAD-170963 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Polderman TJC, Benyamin B, de Leeuw CA, Sullivan PF, van Bochoven A, Visscher PM, & Posthuma D (2015). Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nature Genetics, 47(7), 702–709. 10.1038/ng.3285 [DOI] [PubMed] [Google Scholar]
  39. Plassman BL, Williams JW Jr, Burke JR, Holsinger T, & Benjamin S (2010). Systematic review: factors associated with risk for and possible prevention of cognitive decline in later life. Annals of Internal Medicine, 153(3), 182–193. 10.7326/0003-4819-153-3-201008030-00258 [DOI] [PubMed] [Google Scholar]
  40. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, De Bakker PI, Daly MJ and Sham PC, 2007. PLINK: a tool set for whole-genome association and population-based linkage analyses. The American journal of human genetics, 81(3), pp.559–575. 10.1086/519795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Reynolds CA, Finkel D, & Zavala C (2014). Gene by Environment Interplay in Cognitive Aging. In Finkel D & Reynolds CA (Eds.), Behavior Genetics of Cognition Across the Lifespan (pp. 169–199). Springer; New York. 10.1007/978-1-4614-7447-0_6 [DOI] [Google Scholar]
  42. Richards M, & Hatch SL (2011). A life course approach to the development of mental skills. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 66 Suppl 1, i26–i35. 10.1093/geronb/gbr013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Richards M & Sacker A (2011). Is education causal? Yes. International Journal of Epidemiology, 40(2), 516–518. 10.1093/ije/dyq166 [DOI] [PubMed] [Google Scholar]
  44. Rietveld CA, Esko T, Davies G, Pers TH, Turley P, Benyamin B, Chabris CF, Emilsson V, Johnson AD, Lee JJ and De Leeuw C, 2014. Common genetic variants associated with cognitive performance identified using the proxy-phenotype method. Proceedings of the National Academy of Sciences, 111(38), pp.13790–13794. 10.1073/pnas.1404623111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Ritchie SJ, Tucker-Drob EM, Cox SR, Corley J, Dykiert D, Redmond P, Pattie A, Taylor AM, Sibbett R, Starr JM, & Deary IJ (2016). Predictors of ageing-related decline across multiple cognitive functions. Intelligence, 59, 115–126. 10.1016/j.intell.2016.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Robertson DA, Savva GM, & Kenny RA (2013). Frailty and cognitive impairment--a review of the evidence and causal mechanisms. Ageing Research Reviews, 12(4), 840–851. 10.1016/j.arr.2013.06.004 [DOI] [PubMed] [Google Scholar]
  47. Schneeweis N, Skirbekk V, & Winter-Ebmer R (2014). Does education improve cognitive performance four decades after school completion? Demography, 51(2), 619–643. 10.1007/s13524-014-0281-1 [DOI] [PubMed] [Google Scholar]
  48. Simandan D (2018). Rethinking the health consequences of social class and social mobility. Social Science & Medicine, 200, 258–261. 10.1016/j.socscimed.2017.11.037 [DOI] [PubMed] [Google Scholar]
  49. Sofi F, Valecchi D, Bacci D, Abbate R, Gensini GF, Casini A, & Macchi C (2011). Physical activity and risk of cognitive decline: a meta-analysis of prospective studies. Journal of Internal Medicine, 269(1), 107–117. 10.1111/j.1365-2796.2010.02281.x [DOI] [PubMed] [Google Scholar]
  50. Soubelet A (2011). Age-cognition relations and the personality trait of Conscientiousness. Journal of Research in Personality, 45(6), 529–534. 10.1016/j.jrp.2011.06.007 [DOI] [Google Scholar]
  51. Stern Y (2009). Cognitive reserve. Neuropsychologia, 47(10), 2015–2028. 10.1016/j.neuropsychologia.2009.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Stern Y (2012). Cognitive reserve in ageing and Alzheimer’s disease. The Lancet Neurology, 11(11), 1006–1012. 10.1016/S1474-4422(12)70191-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. van den Kommer TN, Comijs HC, Aartsen MJ, Huisman M, Deeg DJH, & Beekman ATF (2013). Depression and Cognition: How Do They Interrelate in Old Age? The American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry, 21(4), 398–410. 10.1016/j.jagp.2012.12.015 [DOI] [PubMed] [Google Scholar]
  54. Vossius C, Larsen JP, Janvin C, & Aarsland D (2011). The economic impact of cognitive impairment in Parkinson’s disease. Movement Disorders: Official Journal of the Movement Disorder Society, 26(8), 1541–1544. 10.1002/mds.23661 [DOI] [PubMed] [Google Scholar]
  55. Wilson RS, Yu L, Lamar M, Schneider JA, Boyle PA, & Bennett DA (2019). Education and cognitive reserve in old age. Neurology, 92(10), e1041–e1050. 10.1212/WNL.0000000000007036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Yount KM (2008). Gender, resources across the life course, and cognitive functioning in Egypt. Demography, 45(4), 907–926. 10.1353/dem.0.0034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zahodne LB, Glymour MM, Sparks C, Bontempo D, Dixon RA, MacDonald SWS, & Manly JJ (2011). Education does not slow cognitive decline with aging: 12-year evidence from the Victoria longitudinal study. Journal of the International Neuropsychological Society, 17(6), 1039–1046. 10.1017/S1355617711001044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Zaninotto P, Batty GD, Allerhand M, & Deary IJ (2018). Cognitive function trajectories and their determinants in older people: 8 years of follow-up in the English Longitudinal Study of Ageing. Journal of Epidemiology and Community Health, 72(8), 685–694. 10.1136/jech-2017-210116 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES