To my knowledge, there is no need to add a control condition (you mean neutral words, I guess), as there is little difference to the congruent condition. In any case, I would do a within-subject design, as people differ a lot in their overall reaction time. As another possibility, I have recently done the same study with four colors (two hand response) and three colors (right hand only, arrow keys). The training effect was considerably lower with three colors, which is generally an advantage.

Thank you Martin for you answer.

For the neutral condition, i meant a condition without words but only colors.

I'm doing a within-design with 4 colors.

What about the number of trials? I'm familar with the EEG design and its unbelievable huge number of trials... what about this task? Are 30-40 trials a typical number of trials?

Since I've only been doing a variant of the Stroop task in my experiments (Stroop semantic priming), I can't answer your question on sample size and number of trials. A good starting point is following review article:

Macleod, C. M. (1991). Research on the Stroop Effect : An Integrative Review. Psychological Bulletin, 109(2), 163–203.

30-40 sound a bit on the low side to me.

It is critical to make sure that each word is presented equally often in each colour. Many people use 50:50 congruent:incongruent in, say, a four-choice task, but this is very problematic, because it introduces contingency learning confounds. The words should not be predictive of the colours. See some of my work on this issue for the reasons why (e.g., Schmidt & Besner, 2008; Schmidt & De Houwer, 2011).

Thank you very much, Martin and James. I will follow your advices and read your papers or references!

Hi Nicolas

I realize this topic is now a bit old, but I came across it and I thought I would share some findings from my meta-analysis and review of the Stroop task. While my analysis was on the effect of gender, some general variables may still be relevant.

First, performance varies depending on Stroop version. If you use a version that measures number of items recorded within a timeframe you will get less variability, and thus less of an effect, compared to Stroop versions that record reaction times.

Second, whether you include a congruent condition or not largely depends on what you are after. There are multiple different conditions available in the Stroop task, such as:

- Name the color of objects

- Name the ink color of congruent color words

- Name the ink color of incongruent color words (a.k.a. "Stroop task")

- Name the ink color of incongruent color words, which are surrounded by smaller, incongruent objects

- Read color names in black ink

- Read color names in congruent ink

- Read color names in incongruent ink

The vast majority of studies tend to use a version called Golden or Stroop, named after their respective authors. Typically only three aspects are investigated: word reading, color naming, incongruent color naming. Indeed, even though Stroop himself had an incongruent condition where the words were read rather than the ink color named, this component is rarely used.

Third, the number of colors you use appear to show a minimal effect. Even using just one incongruent item will still produce the effect. However, as some other posters have pointed out, you may want to make sure that the distribution of color names and ink colors are equal, in order to avoid any trouble with different colors.

Fourth, most Stroop experiments involve only one round of testing per component. Usually the incongruent condition is presented on a paper or a computer screen, and this consist of only one or two trials, consisting of X amount of items. I personally conducted a version that had 15 trials, each consisting of 30 items, and this increased reliability tremendously. This is not strictly necessary; in my own case the results appeared very reliable after just 5 trials.

Fifth, a negative priming version of the Stroop task is much harder than the regular version. In a negative priming task, the color to-be-named in one item is the same as the color ignored in the immediately preceding item. So, if item 1 is the word “blue” in red ink, and item 2 is the word “green” in blue ink, then you have negative priming. The answer to item 2 is what you had to inhibit during item 1. This is harder. If you are creating your own Stroop task you should be aware of this, even if you do not intend to measure priming. This is because you may create some negative priming items just by chance, and this will reduce performance.

Hope this helps.

Hi there,

I know this is an old thread but maybe someone can clarify to me. Espen said that 5 trials per component are enough for reliable results. What exactly is meant by component here? The particular incongruent ink and word pair?

Hi Austeja

By "component" I meant condition (e.g. incongruent colour-word condition). Apologies if this was unclear.  I referred to the fact that most Stroop studies only have one trial per participant, per condition. For a standard incongruent Stroop task, a participant will usually be given a single set of 100 items. This is perfectly fine, but this means that there is only one datapoint per participant (e.g. time taken). In my version I had 15 trials, and participants varied in performance from trial to trial. This is likely due to an order effect of the colours.

Multiple trials is not strictly necessary, but gives you a more reliable measure of a participant's performance, in my opinion. This also allows you to assess any effect of practice and within-subject variance. The reason I said that 5 trials is sufficient was due to variation in my own data: I saw the greatest variation between trials during the first 5 trials. There was, of course, variation in the next 10 trials too, but performance became more stable. Therefore, I decided to just use 5 trials for the future, as I feel this is reliable enough, and it reduces the total session length.

Hi Espen

Is it possible to know the steps that I have taken in designing the five attempts, which I considered suitable for finding the effects of the stroop phenomenon through the conditions of incongruent color-word condition and congruent color –word condition?

Hi Nicolas

In my opinion, if you want to evaluate the speed of response, the most common number used in the image is 6 rows and 5 columns, which are displayed in 3 rounds.

The first round is only 30 squares of color that the person must announce the names of the colors. The second round shows 30 different words with different colors, in which the person again has to announce only the names of the colors. And in the last round, 30 names of colors with different colors are displayed so that the person can announce only the name of the observed color, regardless of the color name.