The Thrill Is Gone, but You Might Not Know: Habituation and Generalization of Biophysiological and Self-reported Arousal Responses to Video Games

Abstract

Past research on consequences of video game play have conflated two distinct psychological mechanisms, habituation and generalization, into a unified process dubbed “desensitization.” The current paper reports the results of two studies, a repeated exposure study and a single exposure study, which examine habituation and generalization of biophysiological and self-reported arousal responses to violent video games. The findings indicate that repeated play leads to habituation in both biophysiological and self-report responses. Notably, evidence of generalization is more apparent in the biophysiological data. The results are discussed in terms of: (1) implications for game developers and players, (2) implications for game researchers, (3) current conceptual ambiguity between cross-sectional and longitudinal models, and (4) the appropriateness of utilizing self-report measures in longitudinal studies examining arousal.

Keywords:

• Video Games
• Arousal
• Blood Pressure
• Cardiovascular Activity
• Media Effects

Notes

[1] Desensitization is a term often used interchangeably with habituation. Both processes describe a reduction in response related to repeated exposure. We use the terms habituation and generalization exclusively, as desensitization often conflates these two distinct processes.

[2] The power law states “the probability of perceiving similarity or analogy between two items, a and b, is a negative exponential function of the distance, d(a, b), between them in an internal psychological space” (Chater & Vitányi, 2003, p. 346). Put more simply, the more similar stimuli are perceived to be to one another, the more likely they are to be processed similarly. As such, generalization from a psychological standpoint is “a failure of discrimination” (i.e., a failure to distinguish between two stimuli that are different; Shepard, 1987, p. 1322).

[3] Processes related to misattribution of arousal seem particularly important to the current study, as initial experiences are important for determining expectations about future experiences (e.g., first impression effects; Willis & Todorov, 2006). As such, when individuals in the current study experience the novel stimulus, they may attribute higher levels of arousal through self-report to this novel stimulus based on their initial experiences of the original stimulus.

[4] The only difference in gameplay occurred in the introduction to the game, which informed participants of their role in the game.

[5] It is important to point out that not even the controls varied between the original game and the novel game. Both games used the common first-person shooter WASD control scheme, where the keys W, A, S, and D control a character's movement; movement of the mouse controls the character's view; the left mouse button fires the character's weapon; and R is used to reload the weapon (Clarke & Duimering, 2006). The similarity between all features of the games excepting the visual graphics further reinforces our tests of generalization. Observing the expected pattern for habituation and generalization prevents an attribution of the effect to any feature of the procedure or game play, as these features were the same across time and across games.

[6] Researchers (Razminia et al., 2004) have proposed a new formula to calculate MAP. We also examined our data utilizing this newer formula in addition to the traditional formula reported in the paper. Findings were nearly identical regardless of the formula. We report the data from the traditional formula in this paper for consistency with prior research. Analyses utilizing the new formula are available upon request from the corresponding author.

[7] We examined the correlations across all five days of the study for the two biophysiological measures and the two self-report measures. The correlation between the two biophysiological measures indicated that there was a significant correlation with a narrow confidence interval (r = .24, 95% confidence interval LB = .18, UB = .30). Similarly, the two self-report measures were significantly correlated with a narrow confidence interval (r = .35, 95% confidence interval LB = .30, UB = .40). Though these correlations were modest, the fact that their confidence intervals are so narrow (only varying by .06) and the fact that analyses did not functionally differ using the individual scores lend weight to a decision to collapse the biophysiological measures into a single indicator and the self-report scores into another single indicator.

[8] Effect size is generally examined with regard to an absolute value. However, in the current situation, the size and direction of effect are important. As such, our effect sizes demonstrate the difference between CODMW2 and OF (i.e., CODMW2 minus OF). Negative effect sizes here indicate that OF had a higher value compared to CODMW2, while positive effect sizes indicate that CODMW2 had a higher value compared to OF.

[9] It should be noted that increases in the quality of visual graphics or increases in arousal-producing content might overcome the habituation response. It is also important to note here that we are not arguing that violent game content would not lead to aggressive thoughts through priming or aggressive behavior through learning. In fact, repeated exposure to aggressive scripts as primed through media exposure would seem to strengthen perceptual links between such concepts in line with GAM's predictions. Nor are we arguing that habituation of arousal would lead to a reduction in the audience to seek out arousing content. In fact, based on research on film and television (Zillmann, 1991), we would argue that individuals who habituate (particularly those who enjoy arousal-inducing content) would seek out ever-stronger content to overcome habituation. Our discussion here points to the fact that one motivating mechanism for aggressive responses, that is arousal stimulated by violent media exposure, is actually reduced with repeated exposure. Questions still remain as to what is a more important determinant for aggressive responses, arousal (the response energizing mechanism) or aggressive thoughts and cognitions (the response directing mechanism). So, although the importance of aggressive scripts might increase after repeated exposure leading habitual violent media viewers to think about violence more, it seems possible that decreases associated with repeated exposure in the energizing mechanism for determining the strength of aggressive responses could lead to an overall decrease in aggressive retribution (see Zillmann, 1983, for a discussion of arousal as an energizing mechanism for violent retribution after media exposure). These questions are obviously beyond the scope of the current data, but we discuss them here as we believe they are indicative of the theoretical ambiguity between cross-sectional models of media violence and how these models would play out over repeated exposure.