I have to write up my research. And, although I’m capable of writing quite nicely at times, Academic writing paralyzes me. At least the initial writing (I’ve become a decent editor). I figure, maybe I can blog about the papers, where I don’t get fenced in by some implicit belief in that academic writing means obscure writing. (I have to bring in my house-gods of good academic writing, David Hull and Daniel Dennet. Oh, and Joseph LeDoux is not too shabby).
I’m interested in the kind of fluid dynamic that is emotion communication. Or is it emotion processes. Dynamic it is though. We read and react to each other’s’ fleeting expressions, but the interpretation and reaction to any given expression is not a set and static one. The judgment, and perception of expressions are multiply determined. Biased, is possibly a term for it, but that kind of implies that there is something to be biased towards or away from. Perhaps more influenced, corrected, over corrected? I’m starting to have this metaphor of a boat on water, being buffeted around by waves, winds, and its own construction.
The early work I tried to do (and failed at – possibly because, well, we were wrong) tried to see if emotional states influenced the perception of emotional expressions, as in, making it easier to process congruent expressions. Didn’t work. Possibly the expressions were too unambiguous and intense. The prototypical intense expressions you recognize regardless.
What worked was pushing around the judgment where people thought an emotion disappeared. That is, make the emotion change from one to another, and ask people to pin point where it is no longer present. Seem like people use their emotional state as a guide, and are sensitive to a congruent expression longer – if it is morphed into a neutral expression. Attachment style also affected sensitivity to the changing expressions. All of this can be read in the series of Niedenthal papers, where I end up at the very tail end of authors for having done some of the morphs, and running or supervising some of the experiments. The results were not entirely straightforward (it matters whether you morph to neutral, or morph to another expression) but we came up with a plausible explanation for that. Also, sticking a pen in the participants’ mouths, a La Strack seemed to mostly disrupt the sensitivity to any expression, rather than having people behave similar to those that had been induced into happiness. Much has been done about mimicry and its disruption since then, and Paula (Niedenthal) has gone on to do interesting work about simulation and smiles.
But back to the research at hand. And I will go historic. Historic for me, that is, from back before I had any kids outside. At that time, I was involved in a project making emotional avatars. We used the Poser Software, and Ekmans AU’s to make controllable stimulus-faces. Even got that one published. Check out Spencer-Smith et al. These days, much has been done with poser, more professionally by people with more resources (I tried to make an update with my student, and it worked OK, but still, needed more resources. Whaaaaa).
And, why do you want avatars? Well, part of what I did was very low level cognitive/mathematical psychology. The kind where sometimes three lines become too complex stimuli. And, here we wanted to check out the face, and facial expressions. Sender effects were well known even then, at least among the few of us that did emotion before it became big. People are differently good at different emotions. There are even expressive confounds between men and women (check out Kenrick). With a nice avatar, we could have a great deal of control of the stimuli, and avoid sender confounds. At one point, Jesse made up a black and a white avatar with the exact same smiles and asked “so, we could check how people judge the smiles on these”. (This was in the US. Skin-color is very salient as being a marker of other there). I said “I think there’s more to it than just apparent ethnicity. It may depend on the observers implicit attitudes”. We went as far as discussing ideas with Russ Fazio, and then we did nothing.
Hugenberg and Bodenhausen did, though. Perhaps they got at least some of the ideas from us, because we were presenting the poser avatars at a meeting where they were present. But, in some ways it was a question waiting to be pursued, and their work is interesting. Goes in directions I would not have taken it (which perhaps is a positive for me, so I have some directions to explore).
They wedded the poser avatars (not using our morph-targets but making their own), and the morph movie technique. Rather than Fazios priming paradigm, they used Greenberg, Banaji and Nosek’s IAT – which was at that time gaining in popularity. Explicit attitudes was measured with a feeling thermometer. Participants then looked at the morph movies first going from angry to neutral, then from neutral to angry in the second experiment. They should click where the emotion either disappeared (offset) or appeared (onset). Unlike the morphs in the Niedenthal series, these morph movies were running automatically (ballistically she calls it), and were not mouse controlled. Perhaps that is why they got interpretable results both for the on and off-set, which they measured in time rather than in frames. Well, when you run it as a film, those two are interchangeable anyway.
And, yes, those that had a relative preference for white names on the IAT also saw the offset of anger later, and the onset of anger earlier for the black faces. Got that? You seem to prefer whites over blacks on the IAT (associating names and adjectives) you see the anger lingering longer on the black faces (experiment 1), and see the anger coming on earlier in the black faces (experiment 2). Nothing happens for the white faces. There is also no relation anywhere with the thermometer. Not with IAT, not with the responses to the faces. (Incidentally, they turned the thermometer into a relative measure to, by subtracting the black score from the white score – just to make it comparable with the IAT score. Quite a good idea, as that is what comes out in the IAT). They did this with the faces too – by analyzing the difference score, but it is only the black faces that stand for the movement. Twenty four participants in each. Four movies. Quick and simple. In some ways much more quick and simple than mine, although I can’t say that they were very slow or cumbersome. The first one is more… embedded I would think.