Tests for autocorrelation need to be conducted when communication researchers use time-series data that include stationary variables. This case is even more pertinent when dealing with various physiological measures. For instance, heart rate (HR) and respiration are more stationary, while other variables can change more quickly (such as electrodermal response). We tested a series of autoregressive models in which HR was monitored over time periods as a function of students watching video clips of their football team's highlights and lowlights. The physiological measure as captured through average beats per minute (bpm) showed four crucial variations during the viewing of the videos representing changes in affect. These variations were divided into four intervals, and latent growth curve models were estimated to assess the extent that average bpm was changing over time. The models were estimated with and without sex as a time-invariant predictor of change. The analysis revealed that HR did change significantly in each of the four intervals and that there were strong interindividual differences in the initial bpm and in their rates of change over the duration of the intervals.