I've written before about the Two Factor Theory of Emotion and how it relates to game design. Last year (almost two years ago--crap, I'm asleep at the wheel here) I wrote that the Theory seems directly related to good horror games because it suggests that developers can trick you into feeling stressed about the images on the screen when you're actually stressed over extremely difficult game mechanics. This idea has been bouncing around in the back of my mind for some time, and I'm increasingly interested by it, so I thought I'd expand on it here.
In the 1940s and '50s there were a series of experiments performed by psychologists that attempted to establish a link between aggression and sexual arousal. In 1965 an experiment by Barclay and Haber showed that students in a classroom that were verbally abused by their teacher (and thus became angry) showed significantly higher levels of sexual arousal than students in a classroom with a calm teacher. The conclusion of this study and several others was that there was some sort of link between aggression and feelings of arousal.
In 1962, Stanley Schachter and Jerome Singer performed an experiment to see how emotions and physiological reactions might be related. Perhaps, they postulated, it's not anger that causes an increase in arousal but rather the physiological effects of anger (increased heart rate, release of adrenaline, etc) that influenced the subjects. To test this theory, they gave two groups of college students injections: half received a shot of adrenaline and the other half received a placebo. Then they put the students into rooms in pairs of two: one real subject and one confederate that had been instructed to act a certain way. In some cases they told subjects what they had been injected with, and in some cases they did not. In some cases they told the subject to expect physiological change that is substantially different from what adrenaline actually causes (e.g. dullness, head aches). Generally, the subjects were led to believe that they were testing a new drug.
Once paired with a confederate, the subjects were placed in a room and their behavior was observed. The confederate was instructed to act either angry or high. Schachter and Singer were interested in how the subjects would respond. The three factors here (psychologists call them "dependent variables") were whether or not they received adrenaline, what they were told about the shot they received, and how the confederate acted.
The results are fascinating. First of all, subjects who received a placebo didn't show any dramatic increase in emotion. Those that were given adrenaline and told the truth about it also exhibited no dramatic emotional elevation. However, subjects that were not told that they had received adrenaline, and those that were told that they had received something else, both showed higher levels of emotional response. Those subjects reported feeling significantly increased levels of anger (if the confederate acted angry) or euphoria (if the confederate acted high). The misinformed group in particular exhibited an increase in emotion.
This experiment led Schachter and Singer to an idea they called the Two-Factor Theory of Emotion. The theory suggests that emotion synthesis is the result of two "factors": physiological reaction and cognitive context. In cases when the body becomes aroused and the brain has no obvious way to explain that arousal, it will look for external context in order to find a label for an emotion to explain change. What Schachter and Singer demonstrated was that by causing an unexpected physiological arousal (by injecting adrenaline), and then pairing the subject with some external emotional context (a confederate that displayed emotion), they could cause test subjects to mistake their body's reaction as an emotional response. In short, they were able to trick the brain into creating an emotion in a situation where it would not have normally occurred.
This theory might also explain the results of the earlier tests between aggression and sexual arousal. Per Schachter and Singer's theory, the verbally abusive teacher may cause his students to become angry, leading to an elevated physiological state. Some students may misread their body's reaction as sexual arousal rather than rage, and thus report higher levels of interest in their classmates than normal.
But this theory gets really interesting when we consider Donald Dutton and Arthur Aron's follow-up experiment in 1974 involving fear and sexual arousal. Dutton and Aron arranged an experiment where they interviewed males between the age of 18 and 35 on a high, wobbly bridge in Canada. The subjects were people visiting the Capilano Suspension Bridge, which is apparently a pretty scary bridge (it wobbles, has low hand rails, and spans a 70-foot drop). They placed an interviewer on the bridge and asked subjects to answer a short form. At the end of the survey, the interviewer gave the subject their personal phone number in case they had any questions about the experiment later. They repeated this experiment on another bridge that wasn't scary: it had a solid construction, high hand rails, and was only about ten feet up.
Here's the key to Dutton and Aron's experiment: they used two different interviewers, one male and one attractive female. The theory was that if the Two Factor Theory holds, the fear that the subjects felt by being on a high bridge might be misinterpreted as sexual arousal by their brain when it is juxtaposed with the appearance of an attractive woman. The dependent variables included whether or not the men accepted the woman's phone number, whether or not they called her back later (indicating that they were interested in further liaison), and the sexual content of their answers to the survey (which involved writing a short story).
The subjects who were approached by a male interviewer scored low on sexual content in their survey answers. Very few called the interviewer back for more information. Most didn't even accept the call back phone number. This result was the same regardless of whether the subject was approached on the wobbly bridge or the safe bridge.
Subjects approached by the female interviewer scored higher on sexual content. But on the safe bridge, the value was still fairly low, and only two subjects bothered to call her back. On the wobbly bridge, sexual content scores dramatically increased, almost every subject accepted the interviewer's phone number, and 50% of them called back. There seemed to be a strong link between the scary bridge and feelings of sexual arousal in the subjects.
Dutton and Aron had some concerns about flaws with the study, so they did two other experiments. In one, they moved the control group off the bridge to ensure that the same group of people (e.g. people visiting
the Capilano river, probably tourists) were used for the experiment and for control. The results were similar to the original test, showing that population differences were not a factor in the test results.
Capilano Suspension Bridge
The final experiment took place in a lab. Dutton and Aron wondered if sexual arousal might have been increased by the "lady in distress" environment of the first experiment; it might be possible to explain the rise in arousal on the scary bridge by considering that the interviewer herself may have appeared to be in a dangerous or helpless situation. So for the third experiment, they carefully controlled the situation in a lab. This time, they brought in college students and told them that they would be receiving shock treatment. The subjects were again all men and were again paired with an attractive female confederate, whom they believed to be a fellow subject who would also be receiving the shocks. In some cases, the subjects were told that they would receive a "quite painful" shock. In others, they were told it would be merely a tingle. Before the shocks were administered, the subjects were separated from the female confederate and asked to answer questions privately about their level of sexual arousal and feelings towards the confederate. They were also asked about their level of anxiety about the shocks themselves.
The result was a correlation between the level of anxiety and the level of sexual arousal felt by the subject. When the subject expressed higher anxiety about receiving a strong shock, he was also likely to report high level of sexual attraction to the confederate. It didn't seem to matter if the candidate believed the confederate to be about to receive a strong or weak shock herself, thus the "lady in distress" variable was not a factor (although interestingly, there was some increase in sexual imagery in survey questions when both the subject and the confederate were about to receive strong shocks). When subjects were told that they were going to receive a strong shock and were then paired with an attractive woman, their levels of attraction to that woman increased significantly. Dutton and Aron saw this as further evidence of the Two Factor Theory: the anxiety caused by the threat of a shock lead to a rise physiological state, which was then mislabeled as sexual arousal by the brain when it encountered the context of an attractive test partner.
The implications for game design seem clear. Games can obviously cause a physiological reaction; I know that personally my heart rate goes up and I start to increase the strength of my grip on the controller when playing a particularly difficult game. If context (that is, the images on the screen) can then be used to influence my emotional state, that seems like a strong case for making horror games (or, really, any kind of game that is designed to elicit an emotional response) really hard. I wonder if, for example, the trite sexual content in God Hand is more influential than it should be because God Hand's game mechanics are incredibly punishing. Or if, as I've mentioned here before, the incredible level of stress that Siren is able to induce is a function of it's extremely difficult combat and sneaking mechanics.
But this post is too long as it is, so I'll save further game design discussion for the next one. Stay tuned.