formats

The Stroop Task

The Stroop Task is a choice reaction time task where there is dimensional overlap between all three dimensions of the task: the relevant stimulus, the irrelevant stimulus, and the response.  In the dimensional overlap taxonomy, it is considered a Type 8 task.

Stroop (1935; see also Dyer, 1973; Lu & Proctor, 1995; MacLeod, 1991) studied the effects of an irrelevant color word in a color-naming task. In the simplest and most well-known version of the Stroop task, subjects are told to name the color of a stimulus (e.g. say the word “blue” when they see a blue stimulus). Along with the color, however, a color word (e.g. “blue” or “green”) is also presented. The color word can be either presented in the colors, or super-imposed on a colored background (see MacLeod, 1998). The color word is irrelevant, but it can be either consistent with the stimulus color and the response name (e.g. the word “blue” written in blue ink, requiring a response of “blue”) or inconsistent with the stimulus color and the response name (e.g. the word “green” written in blue ink, requiring a response of “blue”).

 

A simple Stroop task

 

Performance is slower and less accurate when the irrelevant stimulus word does not match the relevant stimulus and response, and faster and more accurate when it matches the relevant stimulus and response. The difference in reaction time is usually called the Stroop effect.

Stroop tasks permit a great deal of variation while still yielding the same basic consistency effect. This basic Stroop effect has been found using shapes and shape names (e.g. Irwin, 1978; Redding & Gerjets, 1977; Shor, 1971), pictures and picture names (e.g. Babbit, 1982; Dunbar, 1986; Rayner & Springer, 1986; Toma & Tsao, 1985) and spatial locations and their labels (e.g. Baldo, Shimamura, & Prinzmetal, 1998; Lu & Proctor, 1995; Palef, 1978; Seymour, 1973; Virzi & Egeth, 1985). Auditory Stroop tasks have also been studied, with typical stimuli being the words “left” and “right” spoken in the left and right ear (e.g. Green & Barber, 1981; McClain, 1983; Proctor & Pick, 1998; Ragot & Piori, 1994; Simon & Rudell, 1967). Although Stroop tasks have most commonly used conceptual dimensional overlap between dimensions (e.g. the learned associations between colors and their names) they have also used perceptual overlap (e.g. flanker stimuli) (e.g. Glaser & Glaser, 1982, 1989; Zhang & Kornblum, 1998).

In this standard version of the task, however, there is a problem: there is no way to determine whether this overall difference in reaction time is due to the S-S consistency or the irrelevant S-R consistency, or a combination of both.

Because there are three kinds of dimensional overlap in a Type 8 task, there are in principle three different compatibility effects that could appear: a mapping (relevant S-R) effect, an irrelevant S-R consistency effect, and an S-S consistency effect.

Zhang and Kornblum (1998) were able to measure each of these effects independently by constructing a four-choice Stroop task that included both a congruent and incongruent mapping. In this task, subjects responded to one of four words (“red”, “green”, “blue”, “yellow”) by saying one of four words from the same list. Above and below the target word appeared another irrelevant word from the same list. In the incongruent mapping condition, the words that appeared above and below the target could be either S-S inconsistent and S-R consistent (e.g. “blue” flanked by “blue” requiring a response of “green”), S-S inconsistent and S-R consistent (e.g. “blue” flanked by “green” requiring a response of “green”), or both S-S and S-R inconsistent (e.g. “blue” flanked by “red” requiring a response of “green”).

 

complete Stroop task and effects

 

When the mapping was incongruent, the reaction times were fastest for S-S consistent and S-R inconsistent flankers, intermediate for S-S inconsistent and S-R consistent flankers, and slowest for S-S and S-R inconsistent flankers. This demonstrates that both S-S consistency and S-R consistency contribute independent and separately measurable effects in the Stroop task, as predicted by the dimensional overlap model.

 

 

 
 Share on Facebook Share on Twitter Share on Reddit Share on LinkedIn
formats

The Stimulus-Response Compatibility Task

The Stimulus-Response Compatibility Task, also called the SRC Task  is the type of choice reaction time task where there is dimensional overlap between the relevant stimulus set and the response set.  In the dimensional overlap taxonomy, it is considered a  Type 2 task.

The results of performance with SRC tasks were first reported by Paul Fitts (Fitts & Seeger, 1953; Fitts & Deininger, 1954), who called them “Stimulus Response Compatibility Effects”.   In both papers the tasks  were eight-choice  reaction time tasks in which subjects had to move a stylus from a home position to one of eight target positions.  Since that time, many experiments have been run that have confirmed Fitts’ original results, and have used non-spatial stimulus and response attributes.  Thus was  the field of stimulus response compatibility born.

An example of this type of task is described next.  A subject is sitting at a table, on which  is a screen showing  four small,  horizontally arranged squares.  Inside each sqare is a small circle that functions as a light.   A stimulus consists of one of these lights going on.  In front of the screen are four keys.  The subject’s index and middle fingers of each hand are resting on these keys.  When the mapping instructions assign a response to its corresponding stimulus light (congruent mapping), people are much faster and more accurate than when the instructions assign a response to a non-corresponding stimulus light (incongruent mapping).

This effect is  extremely robust, and happens for a wide variety of stimuli and responses. People are slower to say “blue” in response to a red square than they are when saying “red” in response to a red square, even though there is no perceptual similarity between the stimulus and the response. The effect even shows up when the relationship is very abstract: for example, people can be told to respond to lights of different brightness by pressing buttons with different levels of force, and they are faster when they are told to press a button hard for a bright light than when they are told to press a button hard for a dim light.

The key factor that all mapping tasks have in common is the dimensional overlap between the relevant stimulus and the response.

 
 Share on Facebook Share on Twitter Share on Reddit Share on LinkedIn