Bias in Gene-by-Environment Interaction Effects with Sum Scores; An Application to Well-being Phenotypes

Abstract

In the current study, we investigated the influence of using skewed sum scores on estimated gene-by-environment interaction effects (GxE) for life satisfaction and happiness with perceived social support. To this end, we analyzed item-level data from a large adult twin sample (Ns between 3610 and 11,305) of the Netherlands Twin Register. Item response theory (IRT) models were incorporated in unmeasured (univariate) GxE models, and measured GxE models (with social support as moderator). We found that skewness introduced spurious GxE effects, with the largest effect for the most skewed variable (social support). Finally, in the IRT model for life satisfaction, but not for happiness, heritability estimates decreased with higher social support, while this was not observed when analyzing sum scores. Together, our results indicate that IRT can be used to address psychometric issues related to the use of sum scores, especially in the context of GxE, for complex traits like well-being.

Keywords: Genotype by environment interaction; Item Response Theory; Measurement error; Sum scores; Twin studies; Well-being.

Fig. 1

Histogram and density plots of the Satisfaction With Life Scale (A), Subjective Happiness Scale (B), and social support (C) scores. Dashed line indicates the mean

Fig. 2

Unmeasured GxE model for a single twin, for sum scores (left) and item-level data (right), where θ represents the latent phenotype. Above the dashed line represents the twin model, below the dashed line the IRT model. For clarity, only the item parameters of Item 1 are given (with discrimination/slope parameter α₁ and threshold parameters β₁ through β_k with k being equal to the number of response options M − 1. The models without GxE are estimated by replacing β_0e + β_1eA with e

Fig. 3

Measured GxE model with social support as moderator for both twins, for sum scores (top) and item-level data (bottom), where θ represents the latent phenotype. Above the dashed line represents the twin model, below the dashed line the IRT model (bottom panel). See Fig. 2 for explanation of the item parameters

Fig. 4

95% credibility regions of the AxE interaction effects across the range of estimated genotypic values based on sum scores (top) and item-level data (bottom) for satisfaction with life, happiness, and social support

Fig. 5

95% credibility regions of the GxE interaction effects across the range of social support based on sum scores (top) and item-level data (bottom) for satisfaction with life (SWL). Standardized values in right panel

Fig. 6

95% credibility regions of the GxE interaction effects across the range of social support based on sum scores (top) and item-level data (bottom) for subjective happiness (SH). Standardized values in right panel