Does Ceiling Height Affect the Way You Think?

File this one in the annals of "huh?" There's been a lot of research over the last decade or so on what only be described as the bizarre implicit priming of social concepts. In a typical experiment, participants are given lists or scrambled sentences that contain words associated with a particular stereotype or attitude and people will subsequently behave in a way that's consistent with that stereotype/attitude. For example, Bargh et al.1 gave participants scrambled sentences with words associated with the elderly (e.g., worried, old, lonely, and Florida... no seriously, Florida), told them they were done, and then timed their walk to the elevator. These participants walked slower ('cause old people are supposed to be slow) than participants who'd unscrambled sentences with neutral words. No one knows exactly how this sort of thing works, or why, but it's clear that something's going on, and it's going on without people being aware of it (participants generally report that they had no clue anything had been primed, much less slow-walkingness). So we know that it's possible to get people to think about things in a way that affects their behavior while they remain totally clueless. But a paper by Meyers-Levy and Rui (Juliet) Zhu in press in the Journal of Consumer Research takes the implicit priming weirdness to a whole... 'notha... level2.

It's difficult to know exactly where to start in on this paper. If you read the introduction (oh, I almost forgot, here's the paper so you can read the introduction yourself), you're told that, in marketing or something related to it, there's a "wide-spread belief that ceiling height can affect the quality of indoor consumption experiences." Seriously? But the paper quickly goes from this rather vague common sentiment to the hypothesis that high ceilings activate the concept "FREEDOM," and low ceilings "CONFINEMENT" (for those of you not in the know, words in all caps represent concepts). That doesn't seem too far fetched, I suppose, but it's their next step that loses me. They next propose that priming the concept FREEDOM should promote "relational processing," while the priming of CONFINEMENT should promote item-specific priming. Like I said before, "huh?" Here's their reasoning:

[I]ndividuals in a relatively high ceiling environment, which presumably primes the concept of freedom, should rely predominately on relational processing. This follows because relational elaboration entails elaborating freely or uninhibitedly on multiple pieces of data so as to discern commonalities or higher-order abstract points of intersection that they share . On the other hand, individuals who are in a low ceiling room that primes the concept of confinement may engage predominately in item-specific processing. This should occur because item-specific elaboration involves confining or restricting one's focus to each item by itself and concentrating on its precise, context specific (i.e., relatively concrete) attributes. (p. 8)

I know I've already said this, but "huh?" Anyway, who cares how they arrived at their hypothesis (my guess is, by throwing darts at a board with random English nouns)? They've got data.

The first experiment took place in rooms with ceiling heights of 8 (low) or 10 (high) feet, and involved two tasks. In the first, participants were asked to rate (on a scale of 1 to 7) how much six items "reflected their current body state." The six items related to freedom ("being free, unrestricted, and open") or confinement ("being encumbered, inhibited, and confined"). Naturally, if the concepts FREEDOM and CONFINEMENT are primed by high and low ceilings, respectively, we would expect participants in the 10' room to rate the freedom items higher than the confinement items, and vice versa for the poor folks in the 8' room. The second task involved solving anagrams, three of which involved words related to freedom, and three of which involved words related to confinement. Since priming facilitates the solving of conceptually related anagrams, Meyers-Levy and Zhu predict that participants in the 10' room will solve either the freedom or confinement anagrams faster, depending on the height of the ceiling.

Sure enough, folks in the 10' room rated the freedom-related items higher, in the first task, than participants in the 8' room (5.11 vs. 4.11), and gave the confinement-related items lower ratings (1.89 vs. 3.0). The 10-footers also solved the freedom-related anagrams faster (more than 6 seconds faster), and the confinement-related anagrams slower (about 5 seconds slower). So maybe there's something to this freedom-confinement priming thing. At this point, I'm wondering, "Why ceiling height and not room size?" But never mind what I'm wondering.

Next up, in experiment two, Meyers-Levy and Zhu set their sights on the relational vs. item-specific processing distinction. Recall that they've hypothesized causal relationships between the concept FREEDOM and relational processing on the one hand, and CONFINEMENT and item-specific processing on the other. This leads them to further hypothesize that high ceilings will lead to relational processing, while low ceilings will lead to item-specific processing, when ceiling height is made reasonably salient (read this last bit as saying, "Our first attempt at experiment 2 didn't work out). So experiment 2 involved 4 conditions. Participants were either in the high (10') or low (8') ceilinged-rooms, and the ceiling height was either salient or not. How did they make the ceiling-height salient, you ask? With pretty decorations (from Appendix A, p. 35):

That's a photo of the room for the high-ceiling, height-salient condition. The room for the non-salient condition lacked the pretty decorations.

After being placed in one of the rooms, participants were given two tasks. In the first, they were asked to compare several items that all belonged to the same superordinate category (they give sports as an example). For this task, Meyers-Levy and Zhu predict that participants in the high-ceiling, height-salient condition will produce more common dimensions (they give "require equipment" as an example of a common dimension for the sports category), and to list more abstract rather than concrete properties, since identifying common dimensions and abstract properties both require relational processing. High ceiling, height-salient participants should also classify the different items into fewer subcategories, because relational processing will cause them to notice more relational similarities, and thus be more likely perceive the items as belonging to the same categories. The second task involved evaluating the sleekness of two products. Both products were mostly sleek, with a few "crude" features. Meyers-Levy and Zhu predicted that relational processing would cause people to notice the sleekness common to both products, and thus rate them as more sleek, while item-specific processing would cause participants to notice the "crude" features, and thus rate them as less sleek. So high-ceiling, height-salient participants would produce higher sleekness ratings than low-ceiling, height-salient participants.

Do I need to tell you that the results supported their predictions? I mean, I'm writing a post about it, aren't I? In case you feel like I do need to tell you, the results did support their predictions. In the first task, participants in the height non-salient condition produced about the same number of dimensions and subgroups, and the features they described had about the same level of abstractness, regardless of whether they were in the high or low ceiling rooms. Participants in the height-salient condition, however, produced more dimensions, fewer subgroups, and more abstract features when the ceiling was high than when it was low. The pattern was the same for the sleekness ratings in the second task: similar ratings in the non-salient height condition, but higher ratings in the high ceiling room than the low room when height was salient.

For their third experiment, Meyers-Levy and Zhu went with a memory task. Previous research has shown that relational processing produces "recall clustering." That is, when you give people a list of words to recall, some of which are associated with the same concepts (e.g., six words all related to basketball), participants who process the words relationally will tend to recall the semantically related words together in free recall tasks (that is, tasks in which they're not given any recall cues). Item-specific processing, on the other hand, produces better item recall in cued recall tasks (tasks in which the category-labels are used as recall cues). This time, all participants were placed in height-salient rooms (you know, with the decorations), with either high or low ceilings, and given lists of 36 words representing 6 different categories (6 words per category, for the arithmetically-challenged). After a short delay, participants were given either the free recall or cued recall task. Meyers-Levy and Zhu predicted that high-ceiling participants would produce clustering in the free recall task, and that low-ceiling participants would perform better than high-ceiling participants in the cued recall task.

Sure enough, that's what they found. Furthermore, participants' clustering in the free recall task was mediated by their "freedom-related state" (measured the same way it was in the first task of Experiment 1), and their cued recall performance was mediated by their "confinement-related state." This is further evidence that it is the priming of FREEDOM or CONFINEMENT that is causing the differences in processing styles, and thus the differences in memory performance.

As I've been writing this post, I've also been trying to think of a clever and/or insightful way to close it out. The gist of the findings is pretty simple: ceiling height is related to your feeling of freedom or confinement, and this feeling is in turn associated with whether you tend to process information relationally or in an item-specific fashion. What does this mean? Well, the theoretically interesting lesson is that priming specific concepts can influence the way we process information. That's both cool and a bit frightening. How does priming prejudices, as in Bargh's work described above, affect processing, I wonder? I suppose businesses will want to take note, too. If ceiling height affects the way people think about your products, then you're going to want to make sure the height of your ceilings fits with what you want people to notice about them. Or at least, make sure ceiling height isn't salient, so that you don't have to worry about it (i.e., if you don't want to deal with the effects of ceiling height, get rid of your colorful paper lamps). Other than that, this is just another example of how little we know about ourselves. I mean, how many people out there were aware that ceiling height affects the processing of information? I mean, come on!

¹Bargh, J. A., Chen, M. & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action. Journal of Personality and Social Psychology, 71, 230-244.
²Meyers-Levy, J., & Zhu, R. (In Press). The influence of ceiling height: Theeffect of priming on the typoe of processing people use. Journal of Consumer Research.