Fig. 3

Associations between indicators of adulthood socioeconomic position and grip strength at age 46 years in the BCS70 (linear regression models stratified by sex with sample restricted to those with valid measures of grip strength at age 46 years (maximum N = 7,617* (3,922 females and 3,695 males))) results are combined from analyses run across 50 imputed datasets. *70 participants unable to complete the grip strength tests for health reasons were included by allocating them grip strength values equivalent to the mean of the bottom sex-specific fifth. Model 1: unadjusted (p-values from formal tests of sex interaction, p = 0.015 for Fathers occupation at age 5y, p = 0.025 for Mother’s highest qualification at age 5y and p = 0.016 for Father’s highest qualification at age 5y); Model 2: adjusted for height at age 46y; Model 3: adjusted for Model 2 + birth weight (kg), BMI at age 10y (kg/m²), leisure-time physical activity at age 10y, sedentary behaviour (TV watching) at age 10y and disability at age 10y; Model 4: adjusted for Model 3 + BMI at age 46y (kg/m²) + smoking status at age 42y, sedentary behaviour (TV watching) at age 42y, leisure-time physical activity (days/week) at age 42y and occupational activity at age 46y