order compatible-incompatible blocks

[thomasschubert]

hi guys,

It looks to us like the order of compatible vs. incompatible blocks is not counterbalanced? We tried our test several times, compatible always comes first. And in the output, all codes are for the current trial only, there doesn't seem to be a condition variable (or what is round#?)

Is that correct? Do you plan to change this?

Related, I think in the old days of the IAT, we also counterbalanced assignment of categories to left and right across participants. That is also not done here, right?

I realize that we can influence whether compatible or incompatible comes first by exchanging the categories in the input.txt, but it's of course always the same order, no randomization

best, thomas

[nolach]

Hey Thomas,

You're right on both counts. Neither the order of compatible vs. incompatible blocks nor the assignment of categories to the left/right is counterbalanced. I'm unfamiliar with the latter as common procedure in the literature (it might well be), the former is very important.

My present focus is a database-driven solution for the IAT, so it might be awhile before this counterbalancing is added. If you're interested in adapting the code yourself, it would be fairly straightforward. The initRounds function in /core/js/iat.js would need to be altered to randomize the order. A common method to create 50-50 odds is to call Math.random() and check if the return value is < 0.5 (see https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Math/random).

Regarding the output, each trial is on a separate line. The variables, separated by commas, appear in the following order:

1. Round #
2. Trial #
3. Category Data Label (e.g., in "Science," the male data label is "M")
4. Category Index

5. of Errors

6. Response Time

Hope that helps!
Steven Allon

[thomasschubert]

Steven, thanks for the quick response. I looked into the code, and yes, I can see how to put this in in initRounds(). However, I also suspect it is relevant to other parts (instructionPage(), calculateIAT(), WriteFile()?), not to mention the R file and possibly the experiment manager, if you want to make counterbalancing an option. So I hope I can leave this major change to somebody who is more familiar with the code than I am. Thanks, Thomas

[TTrotter]

Hi,thanks to all who were involved in creating this IAT. Following up on Thomas's post from 9 months ago, does anyone have a solution for counterbalancing (compatible/incompatible)? It would seem that without this the computed effects wouldn't be valid due to ordering effects. thanks.

[jm2209]

Hi Nolach,

I have given these changes a go, but they don't appear to any difference to the order presented. Has anyone had any success with this?

Cheers,

J