Several people, knowing I’m the poster blog guy, asked me what I thought.
The authors themselves write:
Our study had several limitations.
This is an understatement. This study, by Keegan and Bannister, is almost nothing but limitations. It’s as lightweight as this:
The hypothesis is that having the presenter’s clothes match the colour of the poster will result in more visits to the poster. The design of the experiment is actually not bad. The authors show a statistically significant decrease in poster visits when the presenter was wearing clothing that did not match the colour of the poster. This is also supported by ad hoc observations of poster visitors:
5 people were overheard by the observer during the clashing-attire phase to say that the presenter’s blouse did not match her poster, and none visited the poster.
Let’s run through some interpretive issues.
1. The test poster had not one, but four colours on it: blue, lavender, green and yellow. This makes it tricky to say the lavender blouse “matched” the poster colour. It matched a colour, but not all.
I took the colour with the largest surface showing in the picture, and also closest to the presenter’s eye level - green - and placed it into Kuler. The picture below shows the suggested complementary colours using the “triad” model.
The colours at the end, a purple and a gold, are not too different from the other colours in the poster. This suggests that coordination might be a more appropriate description than simple matching.
If I take the same base green and select a “complementary” colour scheme:
I get a suggestion for an orange-brown that is not too far off from the rust worn as the “clashing” colour. Designers often use such contrast colours to make subjects “pop”. For example, red rose petals look redder next to the green leaves of the rose.
2. There is no control for the clothing colour itself, isolated from the matching to the poster. The authors acknowledge this, noting:
People may have decided to not visit the poster... because they did not like the rust blouse regardless of whether it coordinated with the poster(.)
There is no accounting for what colours look good on this presenter. Some people look great in particular colours and terrible in others (I happen to think I look stunning in purple).
This is particularly an issue given that the clashing colour is nearly red. There have been many suggestions that red signals all sorts of things in humans (Changizi et al. 2006, Elliot et al. 2007, Elliot & Niesta 2008), including aggression (Hill & Barton 2005), though the latter has been contentious. Just wearing something reddish may drive away visitors. To test this, you’d have to do an experiment with a red poster; you’d predict lower numbers of visits with the presenter wearing lavender blouse if Keegan and Bannister’s paper is correct.
3. Joshua Drew asks:
(W)hat about if it was a dude who was clashing?
Indeed. You could fill entire libraries have been written about gender expectations, particularly with regards to appearances. It’s an open question if this effect would persist if the presenter were male.
4. The definition of a “visitor” at the a poster includes people looking at it. This would include mere glances of people walking by. It would be interesting to see the data broken down by those looking versus the number talking to the presenter. The numbers would be smaller, but might be a more meaningful measure of poster popularity.
Especially given that we don’t know how the details of the conference. If it’s a busy conference in a small space, with tables in the middle of the walkway (I’ve seen this), you almost can’t look at anything else, because you’re stuck in foot traffic.
5. Two presenters at two posters at one meeting. The sample is tiny. This is barely even a preliminary study. The authors themselves admit:
It would have been ideal to have conducted this study during several poster sessions; however, funding limited us to one medical education conference, which had only one poster session.
Given the large number of people who go to multiple conferences a year, perhaps the authors could have rounded up some more volunteers rather than going it alone.
This paper has been cited three times according to Web of Knowledge, and eight times according to Google Scholar. But nobody ought to take this research too seriously yet. This needs replication and a more robust study design. If anyone wants to collaborate on replicating and extending this experiment, let me know.
In the meantime, remember that black goes with everything.
References
Changizi MA, Zhang Q, Shimojo S. 2006. Bare skin, blood and the evolution of primate colour vision. Biology Letters 2(2): 217-221. http://dx.doi.org/10.1098/rsbl.2006.0440
Elliot AJ, Maier MA, Moller AC, Friedman R, Meinhardt J. 2007. Color and psychological functioning: The effect of red on performance attainment. Journal of Experimental Psychology: General 136(1): 154-168.
Elliot AJ, Niesta D. 2008. Romantic red: Red enhances men’s attraction to women. Journal of Personality and Social Psychology 95(5): 1150-1164.
Hill RA, Barton RA. 2005. Psychology: Red enhances human performance in contests. Nature 435(7040): 293-293. http://dx.doi.org/10.1038/435293a
Keegan DA, Bannister SL. 2003. Effect of colour coordination of attire with poster presentation on poster popularity. Canadian Medical Association Journal 169(12): 1291-1292. PMID: 14662667. http://www.cmaj.ca/content/169/12/1291.full
Related links
Dress sense
Google Plus discussions here and here.
Hat tip to Liz Neeley for this lead.
Photo by Neal. on Flickr; used under a Creative Commons license.
I really enjoyed this post! Here's some andecdotal evidence for you: my labmate, an undergrad at UW-Madison, wore our school colors (red shirt with a UW Madison tie, black pants) next to his poster at ESA last year and won "best undergraduate poster." His poster was mostly in greens.
ReplyDelete