On Internet Experiments

The entry on aggression caused certain turmoil among those present on the blog. My dear followers even decided to share the video in the comments. I checked it out and decided to share my thoughts on the social experiments on the Internet. I taught experimental psychology at the university back in the day. And when I come across all sorts of Internet studies of people’s reactions, I often had just one question: “Have you at least read something on the subject?”. I’m not talking about the students’ research papers, the university methodology is OK. Also, I hold nothing against the sociologists. But market research is often seems amateurish. Especially strong issues I have not with specialized offices but with respective departments of large and medium-sized companies. Let’s proceed to the analysis of video and special aspects of experiments on the Internet. The first point where you can screw up is while picking a sample group. The video summary says that the experiment investigated “aggressive society.” But I would – considering the procedure – describe it as “studying the reaction of passers-by in the evening time.” Not quite the required society, in my opinion. On the Internet we often get the situation where a group of users, used in the experiment upon which the conclusions were drawn, does not reflect the group of users of the service. For example, researchers often make study of the comments on the forums. Sometimes that will do, but we must bear in mind that a group of commenters, as a rule, is different from the general group by at least extraversion, narcissism and, in the case of critical comments, negativity. You may also factor in the differences in education, place of residence, the average age. I’d rather see the forum as a place where ideas are essentially set out, without projecting their propagation in the general group. The second point – the way the experiment is built. In the video I clearly see the provocation, most notably the sexual one. I do not see the aggression research there. Nowadays, one very solid research method is spreading over the Internet, the AB testing. The method of definitely satisfactory and minimizes the number of possible errors. And almost always, when the survey is conducted by another procedure, its correctness is doubtful to me. For example, as a test procedure, we gave our American business partner a list of psychological characteristics of his staff based on their performance on the informational web site. We offered to issue protocols recording various gradations (eg: “none-slightly-medium-developed”) of available respondents’ qualities. Offered to take into account the selfconsciousness and opinion of colleagues and, if possible, the test data. He just gave those lists of characteristics to his staff and watched them gleefully poke at each other and make fun of each other. He was satisfied. We then had to carry the whole procedure again with other partners to ensure the quality of output and arrange the conclusions we’d not be ashamed to show the experts 🙁 And finally, the conclusions. In the above-mentioned video I saw fairly adequate response: men reacted positively, women – sometimes aggressively. But the overall range of reactions was very wide. The actions looked mostly like artifacts. The conclusions drawn were not based on real results, but on a desire to prove the initial idea. I may refer to the words I’ve written earlier – the conclusions are best left to a person not involved, someone not depending on the outcome of the study. Surely only if you are interested in objectivity :).