Table 1.

Description of key features of four groups of analyses.

Features of the analysis	Four groups of analyses
	I	II	III	IV
Dependent variable (Yit)	Psychological wellbeinga	Psychological wellbeinga	Psychological wellbeinga	Binary variable representing a specific psychological symptom
Main exposure of interest (Xk,it)	Travel mode binary variable(s)	Commuting time-travel mode interaction terms	Travel mode transition variable(s)	Travel mode binary variable(s)
Description of models used	Four separate models, varying in terms of number of travel mode binary variables and number of covariates	Four separate models, varying in terms of number of interaction terms and number of covariates	Four separate models, varying in terms of number of transition variables and number of covariates	Twelve separate models, with binary dependent variables representing each of the GHQ12 symptoms
Method of regression analysis	Linear fixed effects	Linear fixed effects	Linear fixed effects	Fixed effects logit

Table provides a summary of the four groups of analyses which were conducted using STATA (version 12.1).

^aThe 36-point GHQ12 Likert scale, increasing in psychological wellbeing.