Feeling bias in the measurement of happiness (by guest blogger Dan Haybron)

For starters, I want to thank Eric for letting me guest on his blog. This has been a lot of fun, with great comments, and definitely converted me to the value of blogging! Thanks to all. Now...

Suppose you think of happiness as a matter of a person’s emotional condition, or something along those lines. If you don’t like to think of happiness that way, then imagine you’re wanting to assess the emotional aspects of well-being: how well people are doing in terms of their emotional states. What, exactly, would you look to measure?

An obvious thought is feelings of joy and sadness, but of course there’s more to it than that: cheerfulness, anger, fear, and worry also come to mind, as well as feelings of being stressed out or anxious. So if you’re developing a self-report-based instrument, say, you’ll want to ask people about feelings like these, and doubtless others.

Here’s what Kahneman et al. (2004) use in one of the better measures, the Day Reconstruction Method (DRM): “Positive affect is the average of happy, warm/friendly, enjoying myself. Negative affect is the average of frustrated/annoyed, depressed/blue, hassled/pushed around, angry/hostile, worried/anxious, criticized/put down.” Also measured, but not placed under the positive/negative affect heading, were feelings of impatience, tiredness, and competence. (I’d be inclined to put the former two under negative—detracting from happiness, and the latter under positive—adding to happiness.) Another question asks, “Thinking only about yesterday, what percentage of the time were you: in a bad mood, a little low or irritable, in a mildly pleasant mood, in a very good mood.”

I think these are reasonable questions, but doubtless they can be improved. First, are these the right feelings to ask about? Second, should each of these feelings get the same weight, as the averaging method assumes? But third, should we only be looking at feelings?

Exercise: think about the most clearly, indisputably happy people you know. (Hopefully someone comes to mind!) Good measures of happiness should pick those individuals out, and for the right reasons. So what are the most salient facts about their emotional conditions? How do you know they are happy? Did you guess the integral of the feelings listed above over time? I doubt it! In my case, the first thing that comes to mind is not feelings at all, but a palpable confidence, centeredness, or settledness of stance. (BTW, the most blatantly cheerful people I know don’t strike me as very happy at all; their good cheer seems a way of compensating for a basically unsettled psyche.) For the people I’m thinking of, I’m guessing they’re happy because of what seems to me to be their basic psychic orientation, disposition, or stance. They are utterly at home in their skin, and their lives.

If this is even part of the story, then affect measures like those above appear to exhibit a “feeling bias,” putting too much weight on feeling episodes rather than matters of basic psychic orientation. How to fix? I don’t know, but one possibility is to use “mood induction” techniques, e.g. subjecting people to computer crashes and seeing how they respond. A happy person shouldn’t easily fly into a rage. But this won’t work well for some large surveys. And what about occurrent states like a constant, low level stress that doesn’t quite amount to a “feeling” of being stressed, or at least not enough to turn up in reports of feeling episodes, yet which may have a large impact on well-being? And how do you tell if someone is truly centered emotionally?

I believe that in the psychoanalytic tradition little stock is put in sums of occurrent feelings, much less reports of those feelings, since so much of unhappiness (and by extension happiness) in their view is a matter of the unconscious--deep-down stuff that only came through indirectly, in dreams, reactions to situations, etc. I think this is roughly right. But how do we measure that?

Basically, my question is, how should the sorts of affect measures used in the DRM be changed or supplemented to better assess happiness, or the quality of people’s emotional conditions?