Table 5.

Associations of spherical equivalent refraction with exposure and behavioral characteristics of religious group, time outdoors, near work, age learned to read, and hours spent in school (n = 168).

Exposure and behavioral characteristics	Univariable analysis $\widehat{\beta }(95\%CI)$	P value	aMultivariable analysis $\widehat{\beta }(95\%CI)$	P value	bAdjusted analysis $\widehat{\beta }(95\%CI)$	P value
Group		P < 0.001		0.006		P = 0.049
Religious	0.87 (0.37, 1.38)	P = 0.001	1.24 (0.46, 2.02)	P = 0.003	0.93 (0.19, 1.67)	P = 0.015
Secular	1.06 (0.51, 1.62)	P < 0.001	1.52 (0.58, 2.46)	P = 0.002	0.99 (0.1, 1.89)	P = 0.031
Time outdoors	0.14 (− 0.26, 0.53)	P = 0.493	− 0.09 (− 0.48, 0.3)	P = 0.655	− 0.02 (− 0.38, 0.34)	P = 0.905
Near work overall	− 0.09 (− 0.17, 0)	P = 0.053	− 0.1 (− 0.19, − 0.02)	P = 0.018	− 0.11 (− 0.19, − 0.03)	P = 0.007
Early readers (< 6 years old)	− 0.62 (− 1.07, − 0.16)	P = 0.009	0.62 (− 0.14, 1.39)	P = 0.106	0.61 (− 0.09, 1.32)	P = 0.086
Time in school	− 0.31(− 0.5, − 0.12)	P = 0.002	− 0.17 (− 0.37, 0.03)	P = 0.086	− 0.17 (− 0.35, 0.02)	P = 0.082

Reference categories: Religious group: ultra-Orthodox; Older Readers (> 6 years).

Each model was implemented with numerical and categorical exposure forms included into the model to safeguard model assumptions (not presented).

$\widehat{\beta }$ Estimated beta coefficients from mixed linear regression model, CI confidence intervals.

^aMultivariable Analysis: collective effects of exposure and behavioral characteristics without a priori age and parental myopia covariates; ^bMultivariable models further adjusted for age and parental myopia.

*3 boys with missing parental myopia status.