I am currently trying to find information to build a computerized Stroop test, it has come to my attention that many research papers do not specify how many items are in their Stroop tests and what is the ratio of congruent to incongruent items in their test. Some research use 240 items, some use 12 trials, the huge difference from paper to paper is confusing.
I am looking for evidence or a guide as to what is the right amount, and why with evidence. If anyone knows of such a source it would be most appreciated.
The number of trials is generally identical across conditions.
The three things that seem to vary a lot (in my limited experience) are:
(1) Whether there is a "read" condition with the colored words (i.e., where the word red printed in blue would elicit the response red). Usually there is not, because it seems to come out pretty much the same as reading words printed in black & white.
(2) Whether three or four colors are used.
(3) Whether the test involves a complete run-through of the items for each condition (e.g., starting with reading, followed by color-naming, and finally interference) or uses some sort of randomization procedure so that smaller groups of items are presented in each condition, but there are several blocks.
On the whole, these minor technical variations don't seem to have much impact on the effect as such. Of course you'll obtain slightly different results whenever you change anything, so you need experience with a given format (or norms) to know what it the standard/expected type of result.
I would like to add one problem I've observed in computer-based versions of the Stroop. Since the test-taker cannot simply say (for example) "BLUE," the program will establish a set of keys that are to be pressed for each - e.g., a slash for red, and so forth. The problem is that these are arbitrary, not associated with the words/colors in question, so there is an error term introduced when a person hesitates over which key represents the color they wish to give in response.
Check out Kane and Engle (2001). They manipulated the percent of congruent trials (0%, 50%, and 75%) on the stroop. They found individual differences were difficult to detect in the 0% - 50% congruency conditions. In the 75% congruency condition, low-spans made about twice as many errors as high-spans. This may help guide you in determining your congruency ratio.
Kane, M.J., & Engle, R.W. (2001). Individual differences in executive attention and the Stroop.
@Jane Yonehiro where can i find the research you are suggesting? It appears to be an unpublished manuscript
Hi Edbert,
The results are reported in this article: http://englelab.gatech.edu/2002/current-directions-in-psychological-science-2002-engle-19-23.pdf
You may also be able to contact the Engle lab for more information through their website : http://englelab.gatech.edu/
Hi Edbert,
someone asked a similar question here:
https://www.researchgate.net/post/What_is_the_optimal_design_for_a_Stroop_task
I found the answers there quite helpful as well.
- Badges
- Science topic