Date of Award:


Document Type:


Degree Name:

Doctor of Philosophy (PhD)


Instructional Technology and Learning Sciences


Byron R. Burnham


The purpose of this study was to measure the influence of an ambient peppermint aroma on participants' time-on-task and performance while using FACTOR, an open-source e-Learning application. I proposed time-on-task was moderating between olfactory stimulation and performance.

A 2x2 research design measured interaction of group (nonscented, scented) and gender of participants (N = 65). The learning content consisted of 28 African countries. Two methods for measuring time-on-task were employed: participant self-report at six learning intervals, and second, video recordings captured by, and stored on each participant's computer.

Independent samples t tests were used to measure group and gender differences in time-on-task and performance. Relationships between time-on-task and performance were assessed using bivariate correlation and were reported as r values.

Time-on-task differences between groups were not significant but garnered ES =.53. After 24 minutes of learning time, control females spent more time-on-task than control males (ES =.71), which was a statistically significant result.

There was a weak to almost neutral correlation between time-on-task of all participants and performance (r =.1) where controls showed a weaker relationship (r =.05) than treatments (r =.26). The correlation between observed time-on-task and posttest performance was neutral for controls (r = .008) but moderate for treatments (r =.38). During the 40-minute learning session, the relationship between observed time-on-task and performance was r =.04 (females) and r =.55 (males), which was statistically significant.

When examining time-on-task at the six measured intervals, the relationship with performance was strongest for treatments after 16 (r =.39) and 24 (r =.39) minutes of learning time. The direct influence of olfactory stimulation on performance was weak as the peppermint scent had a greater influence on time-on-task. Significant differences and notable effect sizes were not achieved by examining these variables.

Analysis of the entire model showed the variables (condition, time-on-task, performance) were weakly correlated (r =.19) and that only 4% of the variance in the model was explained by its variables. As such, I failed to reject the null hypothesis, which was that time-on-task did not act as a moderator between condition and performance.

Included in

Education Commons