Figure 3.

Bifactor measurement model of the cognitive variables. All paths are statistically significant at p < .05. The circles represent the latent variables for Common Executive Failures [Common Exec (Fail)], the “residual” variance shared only among the WMC measures (WMC^Res), and the “residual” variance shared only among the mind wandering measures (TUT^Res). The boxes represent the observed variables loaded onto each latent factor. The arrows represent the modeled direction of the pathway between variables (double-headed arrows indicate correlation). The rightmost column of numbers next to boxes indicates factor loadings onto the Common Executive factor and the leftmost column of numbers next to the boxes indicates factor loadings on the WMC-specific or TUT-specific factors. For the observed variables, OPERSPAN = operation span, READSPAN = reading span, SYMMSPAN = symmetry span, ROTASPAN = rotation span, RUNNSPAN = running span, COUNTERS = updating counters, ANTI-LET = antisaccade with letters, ANTI-ARO = antisaccade with arrows, SEM-SART d′ = d′ score from semantic SART, SEM-SART rtsd = intrasubject standard deviation in RT from semantic SART, N-Stroop = number Stroop, S-Stroop = spatial Stroop, ARROFLNK-SR = Stimulus-response (S-R) conflict effect in arrow flanker, ARROFLNK-SO = Stimulus-stimulus (S-S) conflict effect in arrow flanker, LETTFLNK-SR = S-R conflict effect in letter flanker, LETTFLNK-SO = S-S conflict effect in letter flanker, ACCYFLNK-SR = S-R conflict effect in conditional accuracy flanker, ACCYFLNK-SO = S-S conflict effect in conditional accuracy flanker, MASKFLNK-SR = S-R conflict effect in masked flanker, MASKFLNK-SS = S-S conflict effect in masked flanker, CIRCFLNK = circle flanker, SART-TUT = proportion of TUTs reported in the semantic SART, LETT-TUT = proportion of TUTs reported in letter flanker, ARRO-TUT = proportion of TUTS reported in arrow flanker, NUMS-TUT = proportion of TUTs reported in number Stroop, 2BAC-TUT = proportion of TUTs reported in two-back task.