In the coming months, I’ll be taking results from the solo response project to several conferences, and reviewer feedback has me worried about people dismissing this data because I collected the data from myself. I keep getting distracted by these imaginary confrontations with suspicious researchers so it’s time I lay down some concisely-expressed arguments to appease the hypothetical skeptics.
Problem 1: One subject = bad empirical research
I don’t like this one because the premise is wrong, but I’ve gotten it a lot already, so here goes.
One subject does not a rule make… I am the first to admit that it is a bad idea to generalise too much from a single person’s response to music. In fact, a lot of my work to date has been focused on the problem of how differently people respond to the same music. This experiment is an opportunity to document consistency and variability within a single subject. Eventually it would be great to get dozens of responses from each of hundreds of people, but I’d like to know more about how these signals behave before undertaking such an intense collections process. Looking at data from one subject at a time makes it possible to consider: how to evaluate different manifestations of emotion across all of these different measures, how to quantify/parametrise consistency of response, and how to improve the methodologies available to get cleaner data in the future.
…and one counterexample can a rule break. We can’t know where on the spectrums of listener characteristics this subject lies, but that is not to say we can learn nothing from the specific responses she exhibited. There are theories of emotion in music which can not explain the responses collected for the solo response project. Unless the skeptics dismiss these data as fictional (see Problem 3), these theories need to be revisited.
Problem 2: The researcher is the wrong subject
This one is a little harder to handle because my initial reasoning sounds more like lazy excuses than serious justification for an abnormal practice (in “science”). I get that it is kind of bad form to put yourself under the microscope and then ask everyone to come see what is exposed. There are conclusions I could (and have) drawn from this data which would be a little too personal to share at an academic conference. People will just have to trust me to manage that, I guess.
I used myself as a subject for several practical reasons. I only had a short time to prepare the experiment and I don’t know how to get some random subject to participate in this kind of study. How much should you pay someone so that they come back day after day for 3ish hours of set up and listening? I wasn’t sure how many sessions someone could go through before starting to hate the music; by being the participant, I could pull the plug before things got unbearable. I wanted to be sure the stimuli would cover a range in familiarity, emotional expression, and genres; tailoring an 80+ minute play list for myself to a good chunk of time and it would take longer to do the same with someone else. So basically, getting another subject for this study would have been time consuming and difficult. Not great excuses, I know.
But there were particular advantages for me in being the participant. 1) I like music, and I know I can engage deeply with it, so the chance of measuring strong responses were high. 2) Personally invested in the quality of the data, I was highly focused on both the music and the reporting task. While self assessment is hard and awkward, I’m experienced with the paradim and wanted to be accurate. 3) I had a lot of control over when and where the sessions were being recorded, so I could make the environment safe for musical and emotional engagement. It’s hard to make people comfortable in a lab and I’m sure that my responses would have been dulled if someone in a lab coat had been standing behind me. 4) I was able to introspect on the experience during and following the experimental sessions, deeply enriching my own theories of emotion and music through these hours of dedicated listening.
My academic interest and my personal details (high empathy, excessive musical training) are factors which should be kept in mind when interpreting the responses collected, but they are not, I hope, grounds for ignoring the results. This experiment does not preclude other similar studies from taking place using other people. Research programs need to start somewhere, and I’m confident this data set will be useful for directing future work on music and emotion.
Problem 3: Dishonest emotional and musical engagement
Some people have trouble believing that I cried nearly every session. That kind of emotional expression seems excessive, perverse, and (in some minds) signal that I was trying to hard to get results. In all honestly, I was surprised too (though I would add that my partner was not.) To this accusation I would present two arguments:
a) The emotions collected were more intense than have been measured in many experiments, and context has a lot to do with it. I would compare my responses to those you might experience as a teenager listening to music alone behind closed doors, rather that to formal (concerts) or distracted (while doing the dishes) listening situations. Like the hypothetical teenager, I was willing to get emotional and follow those feelings as far as the music and I could take them. Some days that meant sobbing to Glen Gould and J. S. Bach, other days it meant waiting for the angsty rock to be over. I didn’t get to choose how much I would feel; maybe I could have held back more, but it would have been hard to maintain while still focusing on the music. Anectodatlly, I think I can avoid strong emotions with distraction, but that wasn’t part of the experimental design.
b) It is not uncommon for emotion research, particularly physiological models of emotion research, to use actors for collecting data, rather than “normal” people. Emotional experience and emotional expression are strongly tied together, making it reasonable to work with people who are already emotionally expressive for feeling out the range of bodily consequences to different states. In such cases, the participants are not experiencing emotion in relation to real life events, but the results are still considered valid. For the purpose of looking at how we measure emotions to music in time, starting with a strongly emotional person isn’t a bad idea. Once we have a better handle on how stuff can unfold, we can turn to more subtle responses.
Hoof. I think that is the main set of issues. Hopefully my analyses and conclusions will be of sufficient clarity and quality that people can see past the awkwardness of studying myself.