Leaning to the Left Makes the Eiffel Tower Seem Smaller

Many researchers investigate embodied cognition by looking for effects of body posture on mental activities, such as estimating the size of things. In this post, I critique a high profile example of this literature that has many of what I believe are the key problems in this kind of research.

Eerland, A., Guadalupe, T., & Zwaan, R. (2011). Leaning to the Left Makes the Eiffel Tower Seem Smaller: Posture-Modulated Estimation. Psychological Science, 22 (12), 1511-1514.


About Andrew D Wilson

I am a Senior Lecturer in Psychology at Leeds Beckett University. My research is in perception, action and embodied cognition.
This entry was posted in Had I Been A Reviewer and tagged , , . Bookmark the permalink.

One Response to Leaning to the Left Makes the Eiffel Tower Seem Smaller

  1. You say: “If I had reviewed the paper, it would have gone something like this.”

    1. The hypothesis was clearly written once the data were in

    If I had been the editor, I would have told you: “You cannot make unfounded accusations like this. Please remove it from your review; it is not professional.” As a co-author of this article, I can assure you that we DID have this hypothesis a priori. See my post on the genesis of this idea (http://rolfzwaan.blogspot.nl/2013/02/behind-eiffel-tower.html). It follows from the SNARC (http://en.wikipedia.org/wiki/Spatial-numerical_association_of_response_codes) idea that a bias to the left (vs. the right) activates smaller numbers. This has been shown with hands, eye movements and whole body movements. We included the upright condition only because we wanted to hide our manipulation (preventing subjects from becoming aware they were leaning one way or the other) rather than for theoretical reasons (given that I recommended adding this condition, I am responsible for this choice). Because we had the data for the upright condition, we reported them (we always report all of our conditions) and included them in the analysis. Strictly speaking, the comparison left-vs. right is the only comparison that matters theoretically. The mental number line experiments that I know of typically only compare left vs. right. Because including the upright condition cost us power, it was not an ideal solution from that perspective either. However, we reasoned that not including it would have been the worse option. Running experiments is about making choices and these were the ones we made.

    2. The data don’t actually support the mental number line theory

    Again, the key comparison is between left and right, see all the work on the SNARC effect.

    3. Handedness doesn’t explain this

    This is indeed a post-hoc explanation. We reasoned that a slight shift to the left would be more salient to the right-handed subjects than a shift to the right and found some support for this in the literature. Like I said, it is not theoretically key.

    4. There is no postural data in this paper

    Subjects saw a dot representing their center of pressure (COP). When this dot moved out of the pre-defined area, they received a warning signal (on the screen that they were attending). This was our way of ensuring that they maintained their center of pressure in the right location. In other words, rather than measuring (and analyzing after the fact) whether subjects were following instructions, we tried to “enforce” them. This still seems the better option.

    5. Many of the questions do not show the effect

    True, but note that the analyses were performed across subjects, as is the rule in these types of counterbalanced studies (http://www.raaijmakers.edu.fmg.uva.nl/PDFs/Raaijmakers%20et%20al%20MinF%20paper.pdf). Because items were manipulated between subjects (i.e., each subject only saw the same question only in one COP condition) we did not expect that every question would show the effect. That not every item shows the effect is extremely common in many experiments, which is why there has been so much discussion about item analyses, starting with this classic paper (http://www.personal.psu.edu/kxz134/blogs/ist590_sp08/clark73.pdf).

    6. The effect sizes are tiny

    True. And we wouldn’t have expected it any other way. To give you some background, I need to tell you something about other experiments we had planned but have not yet been able to conduct because (1) both the first and second author are at different universities doing different things and (2) this line of research is not my line of research, which focuses on language processing.

    As I describe in my post on this (http://rolfzwaan.blogspot.nl/2013/02/behind-eiffel-tower.html), the idea for these experiments arose out of a lab discussion of a mental number line paper (I cannot recall which one). As I wrote: “A key conclusion from our discussion of the target article was that, yeah sure, you can make people think of smaller numbers when the number is presented on the left but the subjects are aware that the number is presented on the left. So maybe the whole process is mediated via lexical associations. The location of the target number activates left, which activates small numbers, simply put. A stronger test of the idea would be one that could show this effect without people being aware of left or right.”

    This led to the experiments under discussion here. However, we also planned to conduct further experiments. In one experiment (or series of experiments), we would instruct the subjects to lean “left” or “right”. However, they would actually be standing upright but we would manipulate the dot representing the COP such that it would suggest to the subjects that they were leaning left or right. In the final experiment(s), we would tell subjects to lean left or right and the dot would accurately show this.

    Across the three sets of experiments we would then have:

    Subconscious cues Conscious cues
    Experiments 1-2 (published) + –
    Experiments 3-4 (planned) – +
    Experiments 5-6 (planned + +

    And the corresponding effect sizes we were expecting were: small, larger, largest. And this is why the small effect size was no surprise to us. We expect lexical mediation to play a much more important role than subconscious cues.

    As I said, because of personnel issues, these experiments have not (yet) been conducted.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s