Background
The SART task has been previously implemented in a follow-up experiment comparing the influence of online or offline environments on task performance. A custom Jspsych script (de Leeuw et al., 2021) was used to create the online equivalent of the laboratory version of the task in a within-subject control study (young adult students, n=14, mean 22.36 years ± 1.78). In the original experiment, each participant conducted the experiment twice – once “offline” in the laboratory and once “online” remotely, with the order counterbalanced – to discern effects of environment.
Method
The SART was modified to create a duration hypothesised to induce fatigue. The whole experiment contained 6 experimental blocks, each with 90 trials for a total of 540 trials. A practice block of 20 trials was also delivered beforehand, and a brief break given between blocks. Each stimulus flashed on the screen for 500ms before disappearing for 3500ms. The total duration of the experimental blocks was 36 minutes; this was selected to expand on prior research with a SART duration of ~10 minutes.
Results
Overall, little evidence was found for a relationship between state fatigue and SART measures: Welch’s t-tests concluded no difference between mean commission error frequency and go response times of the first and last blocks (p>.05) when fatigue was hypothetically induced. Regressions also found no correlation between changes in VAS-F scores pre- and post-SART and commission error frequency (p>.05). In a follow-up revisit of the findings, log inverse efficiency scores were recomputed in the first and last block of the comparison study for both online and offline conditions to facilitate comparison and serve as a further indication of invariant baseline inverse efficiency levels specific to the implementation of the paradigm. Altogether, log Inverse efficiency mean difference overall were largely unimpacted by age group (young, older), environment (offline, online) or fatigue levels induced by repeated blocks of SART (befor, after) but consistently showed a mean performance mean of 7.63, sd = 1.38. In contrast, the motivational manipulation alone clearly introduced differences in the IE coefficient in the younger participants, indicating a greater impact of motivation, rather than fatigue on the behavioral metric. Strikingly, the IE in the motivated young condition was by far the lowest, indicating least expenditure of time for a correct answer. The following interactive interface can be used to access summaries of the scores, indicative of a stable IE scores across most conditions with the exception of the heightened difference between young motivated and unmotivated participants.