Table 1.

Testing with a single-trial measure of time-on-task

Variable	Coefficient (as % difference)
	Model 1	Model 2	Model 3†	Model 4
In terms of trial types
Foreperiod
=1.0 s	5.24***	4.91***	7.28***	7.46**
=1.5 s	−2.21
≥1.5 s		−5.21
≥2.0 s	−4.46
Hazard Rate
=high	25.30*	24.17*	28.38***	45.59***
Beyond trial types
Time-on-task
T				−16.52***
Interactions
T × {HR = high}	−3.93*	−4.01*	−3.73*
T × {HR = low}	−26.46***	−26.64***	−26.68***

Estimated differences were given by GLIRT coefficients represented as percentage change in v value (perceptual processing speed) per explanatory variable unit increase on average across subjects and sessions. From Model 1 onwards, the foreperiod (FP) coefficients were not significant beyond the FP of 1.0 s. Model 2 was designed as an alternative to simply eliminating the nonsignificant FP coefficients beyond 1.0 s. The step from Model 1 to Model 2 could not be rejected, −2lnΛ = 66.7, p [> χ ²(64)] = .383. Model 3 was designed to test elimination of FP coefficients beyond 1.0 s, and the step from Model 2 to Model 3 could not be rejected, −2lnΛ = 58.3, p[> χ ²(64)] = .677. Model 4 was designed to test whether the time-on-task (T) effects were independent of the hazard rate (HR) conditions, but this model was rejected in favor of Model 3, −2lnΛ = 92.1, p[> χ ²(64)] = .012. Model 3 won the model selection as further nesting to Model 4 was rejected

HR = hazard rate; T = time-on-task

^†Model 3 wins the model selection. Further nesting to Model 4 was rejected, p < .05

*p < 0.05, ** p < 0.01, *** p < 0.005