Theories, Methods, and Procedures
This study initially recalled research done pertaining to the measure of effect that media (specifically negative or violent media) has on behavior. For example, previous studies used the General Aggression Model (GAM) to determine whether violent video games led to aggression (Anderson & Bushman, 2002). Because correlational data was found to support the theory, using the GAM, a similar model to promote both violent and non-violent measures of effect due to media, called the General Learning Model (GLM), was created (Buckley & Anderson, 2006). At the time of the current study, that theory (GLM being a valid measure of prosocial media’s effects on the internal state of a participant, and whether that state then effects behavior) hadn’t been tested. The current study aimed to do so (Greitemeyer, 2009).
In addition, the current study used an experimental design to attempt to show that prosocial lyrics in music promote prosocial behaviors in participants. They did this in three experiments: one to measure increase in prosocial thoughts (a dependent variable, operationally defined as the number of prosocial words created via word fragments), one for increases in empathy (a dependent variable, operationally defined as the self-reported feelings for the author of two reviewed essays), and one for increases in action/behavior (a dependent variable, operationally defined by whether or not a P made a monetary donation). (Grietemeyer, 2009)
Experiment one sought to determine whether listening to prosocial lyrics, as opposed to neutral lyrics, would increase prosocial thoughts. Ps included 34 students (most of which were women) from a Germany university. Ps were randomly assigned to either the control or experimental conditions. In the control condition, participants listened to a neutral song, then completed a list of word fragments. They then answered two questions to control for the perceived prosocial content of the song they listened to. Ps in the experimental group did the same, with the only difference being the prosocial lyrics of the song. (Grietemeyer, 2009)
For experiment two, 38 students from a German university (again, mostly women) were asked to listen to a prosocial or neutral song, respective to which group they were randomly assigned to, after which they read two essays (which they were told were written by another, missing, participant). After reading these two essays, Ps were asked how they felt about towards the author with regards to sympathy, compassion, and soft-heartedness. The aim of this experiment was to determine the effects of prosocial music, as opposed to neutral, on empathy towards others. (Grietemeyer, 2009)
Experiment three sought to measure to what extent prosocial songs, as opposed to neutral ones, affected prosocial behavior. They did this by randomly assigning Ps (consisting of 90 German university students, most of which were female) to either the control group or the experimental group; differentiated again by whether they listened to prosocial or neutral songs. After listening to respective songs, Ps were offered the option to donate to a non-profit organization. After given two minutes during which they were left alone, participants were questioned about the perception of anything suspicious. (Grietemeyer, 2009)
In all three experiments, researchers controlled for possible confounding variables in a variety of ways. For example, in order to control for whether a song was understood to be neutral or prosocial, researchers used songs in two languages (one English, and one German song each for control and experimental groups in all three experiments), as well as questioned Ps about perceived level of prosocial content. This was done not only during the actual experiments, but also in a pilot study. Additionally, researchers controlled for mood and arousal during the pilot study by asking Ps to rate their level of arousal and how well they liked the song. This led to the song choice. Researchers also controlled for possible effects that liking a song might have on thoughts, feelings, and behavior. They did so by measuring liking via questions submitted to Ps. Finally, because all three experiments had a greater number of female Ps, researchers compared results from both sexes to control for any possible effects thereby. (Grietemeyer, 2009)
Results and Discussion
In experiment one, researchers found, after controlling for possible sex differences, that Ps in the experimental condition (M = 0.21, SD = 0.11) completed word fragments with significantly more prosocial words than did Ps in the control group (M = 0.14, SD = 0.08), t(32) = 2.05, p < .05. This suggested that prosocial songs do have an effect on prosocial thoughts. (Grietemeyer, 2009)
In experiment two, researchers found, via 2×2 ANOVA (song type compared with essay story), that a main effect for type of song had occurred. In other words, Ps in the experimental group rated their feelings about the author as significantly more empathetic, regardless of the essay (F(1, 36) = 6.51, p < .05, n2 = .15). (Grietemeyer, 2009)
In experiment three, researchers found that Ps in the experimental group were significantly more likely to donate money than those in the control group (x2(1, N = 90) = 4.56, p < .05). They reported that 53% of the experimental group donated, while only 31% of the control group donated. This suggested that prosocial songs do have an effect on prosocial behaviors. (Grietemeyer, 2009)
Researchers mentioned that while the hypotheses were supported in the sense that there was a significant difference in prosocial thoughts, feelings, and behaviors between experimental and control groups, the current study did not allow for an understanding of why the changes occurred. There was no way of knowing whether the changes were due to changes in the Ps’ internal states; there was no way to know what the exact cause of the change is cognitively. As such, researchers suggested that a measure of internal processes be taken in addition to the explicit measures used in this study. (Grietemeyer, 2009)
Suggestions for further research include examining whether prosocial songs (and media in general) not only instigate prosocial thoughts, feelings, and behaviors, but whether they also serve to decrease aggressive thoughts, feelings, and behaviors. Another suggested study for future research is that of long-term effects of media on thoughts, feelings, and behaviors; the current study only looked at the short-term effects of prosocial media. (Grietemeyer, 2009)
When reading this article, I was first drawn, naturally, to the claims made at the beginning of the results of the GAM. I am very wary of the aggression by way of violent video games claim. However, as a basis for additional studies, and as long as the measure is, in fact, reliable and valid, I can muscle my way through the irritation. The claims that correlational studies show cause and effect (i.e., violent video games promote aggression based on a correlational study), frustrates me excessively; far more than playing video games does. I found myself overly critical of the steps used to get through the justification of the research, however, knowing that this is not exactly the point, and agreeing that this research is necessary and has to start somewhere, I won’t dwell on these minor criticisms.
I had a few struggles with the actual measures used. As this study is was the beginning of a string of measures on prosocial songs’ effects on thoughts, feelings, and behaviors, I understand that the research has to start somewhere, but in some instances I felt procedure could have been cleaned up a bit. For example, many of the questions asked to rate variables (e.g., the helpful or cooperative content of songs) seemed to prime responses from Ps. Another example of possible priming was the wording used at the end of experiment three, “Participants were told that it would be great if they would donate these 2 € but that it would be also fine if they did not donate. Upon saying this, the experimenter pointed at a box…. (Grietemeyer, 2009, p. 189)” If participants are all hearing the same spiel, regardless of which group they’re in, the priming becomes less of an issue. But I think it somewhat ironic that in a study where they are studying the effects of prosocial songs on prosocial behaviors, they’re using less neutral wording for the experimental procedure.
Another confounding variable may be the stories used in the essays. The subject matter is vague enough that it may have been something similar to an occurrence with a variety of Ps, which may have unknowingly caused the increase in empathy. Relationships and sports injuries are not unusual, after all. A similar confound may be in experiment one, with the use of word fragments. There are those who may not have chosen prosocial words because they don’t have a well-developed lexicon, or aren’t good at word games. Whether a person uses a word that holds prosocial meaning, doesn’t necessarily mean there is not prosocial content to their thoughts.
Typically, it is easier for me to find holes in other researchers’ methods, as I am far less creative than I am critical. That being said, I was unable to think of any other measures of prosocial thoughts, feelings, and behaviors. I feel that there is little external validity in this particular study, though I would intuitively agree with the findings. Empirically, however, students at a German university, and mostly female no less, is not a widely generalizable sample. Researchers could use a more diverse group to collect data, however difficult that may be. Construct validity, for what they claimed to be attempting to measure, seemed to be relatively high, though I think the third experiment would have higher construct validity than the other two experiments. I think that internal validity is also dependent on the interpretation of the person reading the study, as the measures seem to be focusing on very specific definitions of prosocial thoughts, feelings, and behaviors. But this is why we have operational definitions. Certainly, in relation to the research cited in the introduction, this study uses comparable means of measuring results. Results which, I would be hard pressed to come up with a better way of interpreting. That being said, I would have very much liked to have seen a chart of numbers as an appendix; the numbers seemed jumbled and somewhat hard to keep straight.
Overall, I love the concept of this study. I think that it is useful, particularly as a jumping off point for attempting to show the benefits that media can provide. If researchers continue to follow the path set off on in this study, we can continue to further understand the implications of various media and their effects on us mentally, emotionally, and physically.
Anderson, C., & Bushman, B. (2002). Human aggression. Annual Review of Psychology, 53, 27-51.
Buckley, K., & Anderson, C. (2006). A theoretical model of the effects and consequences of playing video games. In P. Vorderer & J. Bryant (Eds.), Playing video games: Motives, responses, and consequences. (pp. 363-378). Mahway NJ: Lawrence Erlbaum.
Grietemeyer, T. (2009). Effects of songs with prosocial lyrics on prosocial thoughts, affect, and behavior. Journal of Experimental Social Psychology, 45, 186-190.