Search

Media multitasking during homework completion has reached epidemic proportions in the modern educational environment. There is a crucial need to resolve the paradox of why students engage in these behaviors, even though they are linked to self-control and performance decrements. We evaluate the proposition that student media multitasking decisions are made in the context of inaccurate forecasts regarding the influence of these behaviors on affect, self-control, and performance. After providing forecasts to estimate the effects of media availability and media removal on these outcomes, sixty college students (N?=?60) were randomly assigned to alternative media availability conditions as they completed an actual homework assignment during an in-lab session. Students predicted media use to result in lower negative affect and diminished self-control. The direction of these forecasts was accurate, with media availability resulting in decreased negative affect and diminished self-control during the homework session. Nevertheless, students exhibited moderate to large forecasting errors in predicting the magnitude of these effects. Although no evidence was attained to demonstrate forecasting errors when predicting homework performance, exploratory analyses suggested the presence of individual differences in the nature of these predictions, with 53.3% of participants predicting a performance decrement, 23.3% of participants predicting no difference, and 23.4% of participants predicting a performance gain under conditions of media availability. We discuss the implications of these findings for students and educators. Student expectations and experiences when using media during homework were compared.Students were randomly assigned to work on homework with or without media available.Students predicted media use to yield reduced negative affect and self-control.The nature of students' expectations was accurate, but the effect magnitude was not.Students exhibited moderate to large affective and self-control forecasting errors.