Skip to main content
. 2021 Aug 17;16(8):e0255348. doi: 10.1371/journal.pone.0255348

Table 2. Standardized regression coefficients from smoking and education polygenic scores (PGS) to outcome variables.

Smoking PGS Education PGS
Age 11 Age 14 Age 17 RI RI Age 11 Age 14 Age 17 RI RI
1 PGS 2 PGS 1 PGS 2 PGS
Grade Point Average -.13 -18 -.15 -.20 -.13 .20 .25 .25 .35 .32
[-.18, -.07] [-.23, -.13] [-.19, -.11] [-.27, -.14] [-.19, -.07] [.15, .26] [.21, .30] [.20, .29] [.29, .42] [.26, .40]
Academic Motivation -.10 -.17 -.16 -.22 -.19 .06 .14 .15 .17 .12
 Self/Parent Report [-.15, -.05] [-.22, -.13] [-.20, -.12] [-.28, -.16] [-.26, -.13] [.01, .10] [.10, .19] [.10, .18] [.11, .23] [.07, .19]
Academic Motivation -.14 -.19 -.15 -.21 -.15 .19 .25 .18 .28 .25
 Teacher Report [-.19, -.07] [-.24, -.14] [-.20, -.10] [-.27, -.15] [-.21, -.09] [.13, .25] [.20, .30] [.13, .23] [.23, .34] [.19, .31]
Disciplinary Problems .08 .17 .16 .28 .26 -.03 -.12 -.08 -.15 -.09
[.03, .13] [.12, .21] [.12, .20] [.20, .36] [.18, .34] [-.08, .02] [-.17, -.07] [-.04, -.12] [-.22, -.07] [-.16, -.01]
Cigarettes per day -- -- -- .20 .18 -- -- -- -.14 .10
[.16, .24] [.14, .21] [-.18, -.10] [-.15, -.06]
Educational Attainment -- -- -- -.19 -.14 -- -- -- .26 .23
[-.23, -.15] [-.18, -.10] [.22, .30] [.19, .27]

Note. Age 11 = regression paths from PGS to outcome variable at age 11; Age 14 = regression paths from PGS to outcome variable at age 14; Age 17 = regression paths from PGS to outcome variable at age 17; RI = random intercept factor; 1 PGS = coefficients from one PGS predictor model; 2 PGS = coefficients from two PGS predictor model; Bold = 95% confidence interval does not include 0. Only one PGS was entered as a predictor in each one PGS predictor model; both PGSs were entered as predictors simultaneously in the two PGS predictor model. Smoking is a random intercept factor score for nicotine quantity assessed at ages 14, 17, 21, and 24 years old. Educational attainment was adjusted for age (mean and median age 29.4 years, SD = 3.9 years, range 19.7 to 39.9 years). All models include participant sex and first five genetic principal components as covariates (covariate regression paths not included). 95% confidence intervals presented under estimates; estimates for which confidence intervals do not include 0 are presented in bold. Confidence intervals derived via clustered, non-parametric percentile bootstrapping (with 1,000 random draws).