“Blue, brown, red, gr-white, blue, white, eh…red!”
Colour naming like the above is what I listened to for weeks during the testing phase of my dissertation experiments. I calculated that I heard at least 68.000 colour words being spoken during testing. Such is the fate of the Stroop experimenter. In this article I will outline some of the key findings of my dissertation, which included a Stroop colour-word test. Currently the dissertation is being rewritten into at least two publications, and therefore details surrounding the paper will be revealed later.
The Stroop task: a meta-analysis
In my dissertation I looked at gender differences in inhibition, primarily the Stroop (1935) task. If you are new to the Stroop task, let me briefly explain how the colour-word (CW) subtask works: the participant is presented with colour words (usually 100 words) printed in incongruous colours, such as the word “blue” written in red ink (blue). The participant must name the ink colour rather than reading the word. This is quite difficult because the colours conflict (you want to say “blue”, but the answer is “red”). The task has been used in a variety of psychology experiments for decades, but gender effects had never been systematically investigated outside single experiments.
Despite over 80 years of research, no systematic analysis of gender effects had ever been conducted on the Stroop task. Some reviews of the task exist (e.g. MacLeod, 1991), but these were highly subjective and concluded that no gender differences existed. I decided to take a statistical approach to the question of gender differences on the Stroop task and conducted a meta-analysis. This is a complicated statistical process where all the available studies on the Stroop tasks were reviewed and analysed in order to find out what the overall gender effect was on the task. This involved a very large literature review and dozens of authors were contacted for additional data, as it was fairly uncommon to report statistics on gender.
Interestingly, my analysis found a small to moderate female advantage in the Stroop task, which depended on which version of the task was used. For instance, some Stroop versions measured how many colours you could name within a timeframe (e.g. Golden), and such versions reported no gender differences (d = 0.05). This is likely because such measurements have little variance. However, tasks that recorded reaction time (how long it took to name X amount of colours) had a lot more variance, and this version of the task found a small female advantage (e.g. Comalli, d = 0.23). If more detailed measurements were used (e.g. milliseconds rather than seconds), or more colour items were used, the difference increased further (d = 0.5).
Why the female advantage?
However, my analysis did not give insight into why this female advantage occurred. In the field there were primarily two main theories proposing an explanation: 1) women have better verbal abilities and can name colours faster (Golden, 1974; Lee et al., 2004), or 2) women have better inhibition abilities and can ignore the colour words more effectively (Bjorklund & Kipp, 1996).
To investigate this, a Stroop task was created with 450 colour items rather than the traditional 100. Results from the meta-analysis suggested that increasing the number of items should also increase the gender difference, and indeed this is exactly what was found (d = 0.5).
My version of the Stroop task also included a negative priming component, which allowed me to investigate if the gender effect was due to inhibition. This works in the following way: in one trial the colour you must name is identical to the colour word you ignored in the trial before. For instance, in trial 1 you might have the word “blue” in red ink (blue), and in trial 2 you have the word “green” in blue ink (green). It now becomes extra difficult to say the colour “blue” in trial 2 because you had just ignored this colour in trial 1. In other words, this pattern adds a systematic interference between trials. Thus, the differences in performance on negative priming trials and ordinary trials can be considered a measure of inhibition (because the only difference between them is the interference system in negative priming). Let us call this difference NPI (negative priming interference).
Interestingly, I found that even though a female advantage existed on the task, this was not due to inhibition because men and women did not differ in NPI. In other words, men and women were equally good at inhibiting irrelevant colours, but women still outperformed men overall. Therefore, the female advantage is not due to superior inhibition skills in women.
Limitations and final comments
Of course, there is one limitation in this study: I cannot say with certainty that the female advantage is due to superior verbal abilities in women. The data suggests that this is so, but I have only found evidence showing that the female advantage is not due to inhibition, and no direct evidence showing that the advantage is due to verbal abilities.
Nevertheless, my analysis found a significant female advantage in a task where it was previously believed that no gender difference existed. Currently we are preparing these results for publications, and more details will be posted at a later date. I also did a third experiment involving motor inhibition, but this experiment will also be outlined in greater detail in a later post as I am currently looking into follow-up experiments that may be used online and posted at this blog.
* Bjorklund, D. F., & Kipp, K. (1996). Parental investment theory and gender differences in the evolution of inhibition mechanisms. Psychological Bulletin, 120(2), 163-188.
* Golden, C. J. (1974). Sex differences in performance on the Stroop color and word test. Perceptual and Motor Skills, 39, 1067-1070.
* Lee, D. Y. et al. (2004). A normative study of the CERAD neuropsychological assessment battery in the Korean elderly. Journal of the International Neuropsychological Society, 10, 72-81.
* MacLeod, C. M. (1991). Half a century of research on the Stroop effect: an integative review. Psychological Bulletin, 109, 163-203.
* Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 6, 643-662.