WRITE AN INTRODUCTION

Paper Project 2 Stroop Background The Stroop effect is a well-known and classic phenomena in experimental psychology. The effect was first reported by J. R. Stroop (1935), and a link to the original paper is available on the class website. Several hundreds of Stroop experiments have been conducted since 1935 and these experiments are summarized in Macleodʼs (1992) review, also available from the website. The original Stroop task involved color naming of word stimuli that are printed in different ink colors. There are two important conditions involving congruent and incongruent item types. Congruent items occur when the ink-color of the word matches the name of the word (e.g., the word red printed in red ink: RED). Incongruent items occur when the ink-color of the word does not match the name of the word (e.g., the word red printed in red ink: BLUE). For each of these stimuli the task is to identify the ink-color of the word and ignore the written meaning of the word. Here are a few more examples of congruent and incongruent Stroop items: Congruent: BLUE, GREEN, YELLOW, RED Incongruent: BLUE, GREEN, YELLOW, RED In the original set of experiments participants were presented with long lists of items and asked to read through the lists naming only the ink-colors and ignoring the word information. The important finding was that people were faster to finish reading lists that were composed of congruent items than incongruent items. This difference in reading time is termed the Stroop effect. For example, letʼs say you were given a list of 60 congruent words and you read all of the ink-color names in 71 seconds. Next, you receive a list of 60 incongruent words and it takes you 94 seconds to read all of the ink- color names. You would compute your Stroop effect by taking the difference between the Congruent and Incongruent reading times. Congruent reading times are subtracted from incongruent reading times to give a positive Stroop effect value: Stroop effect = Incongruent – Congruent = 94 seconds – 71 seconds = 23 seconds The Stroop test can be administered in many different ways. A common modern variant of the task is to present a single Stroop item on a computer screen per trial and have participants identify the ink- color by pressing a key or typing out the response. Several trials could be presented over the course of an experimental session, and the experimenter could vary whether upcoming trials are congruent or incongruent. The trial-based version of the Stroop procedure provides more experimental control and more precise measurement of reaction times. This provides a more fine-grained measure of the time taken to respond congruent and incongruent items. Stroop effects using this procedure are usually measured in the milliseconds (ms). For example, reaction times for congruent items are usually around 500 ms, and reaction times for incongruent items are usually around 600 ms. So, the overall Stroop effect might be around 100 ms. Stroop and Selective Attention The Stroop effect has been used as a tool to study various aspects of learning and attention. In the lab we are focusing on how the Stroop effect relates to questions in the attention literature. In this context, the Stroop effect has been used as a tool to study cognitive control. Cognitive control refers

broadly to the psychological processes that allow people to plan, coordinate, and execute actions necessary to accomplish goals. A central aspect of cognitive control involves the attention processes that are responsible for selecting task-relevant information and ignoring task-irrelevant information when performing a task. The Stroop procedure provides a simple, yet effective, method for presenting people with two sources of information, one that is task-relevant (color) and one that is task-irrelevant (word). For this reason, Stroop items are sometimes referred to as bi-valent stimuli, in that they present two sources of information. Congruent Stroop items (blue in blue) do not present much of a selective attention challenge, both the word and the color information point to the same response. Incongruent items (blue in red) do present a selective attention challenge, the color points to the correct response and the word points to an incorrect response. When faced with incongruent items people must pay attention to the relevant color information and ignore or somehow filter out the irrelevant word information. The fact that people get Strooped, or that the Stroop effect exists at all, tells us something important about selective attention in this task. Specifically, people can not completely ignore the irrelevant word information. If people could completely ignore the irrelevant word information then the Stroop effect would cease to appear. People would be just as fast identifying congruent and incongruent items because they would be able to successfully ignore word information. Using the logic above the Stroop effect is often taken as a measure of selective attention ability. This means that changes in the size of the Stroop effect may represent differences in selective attention ability. People who have very large Stroop effects have poor attentional control over their ability to ignore the word stimulus. People who have very small Stroop effects have excellent attentional control over their ability to ignore the word stimulus. Horse Race Model of Stroop There are several models of the processes involved in Stroop interference. One class of models are known as race models. These models emphasize another important aspect of the Stroop effect, that is the notion of automaticity. The concept of automaticity is that some kinds of information are processed rapidly, effortlessly, and without voluntary control. Basic perceptual processing is a good example. When you open your eyes you automatically see the world. You can control some aspects of how you see the world by opening and closing your eyelids, blurring your vision when your eyes are open, and orienting your gaze toward or away from different locations, otherwise your visual system is automatically processing incoming visual information. For example, if you are looking directly at a tree with clear focused vision, you may find it incredibly difficult, most likely impossible, to make the tree completely disappear from your visual field (of course without closing your eyes or blurring your vision). This is to emphasize that our visual systems automatically process incoming visual information. Some researchers have claimed that word stimuli are also processed automatically. When you look at words, your brain processes the words even when you do not want it to. One of the long-standing arguments in the Stroop literature has been whether or not words are truly processed in an automatic fashion. Regardless of whether words are processed in a truly automatic fashion, it is clear that different kinds of information (words included) are processed at different rates. The fact that different stimuli have different processing rates (e.g., faster or slower) is the big idea behind the horse race model of Stroop. It is well known that people can make responses faster to word information alone than to color information alone. So, for example for the following lists you would be faster to read the list of words than the list of colors.

Word list: red, green, blue, yellow, green, blue Color list: xxx, xxx, xxx, xxx, xxx, xxx The take home message here is that people can make faster responses to identify word information than they can to identify color information. Words have faster processing times than color information. This fact is the central feature of the horse race model. In the model, different kinds of information are assumed to race against each other. Letʼs break this down. Imagine the following stimulus is presented on a computer screen: BLUE Above we have the word BLUE printed in red ink. The horse race model assumes that when this stimulus is presented a race begins. Just like two horses might race against each other to a finishing line, the word information and the color information race against each other to a finishing line. We liken the word information to one horse, and the color information to a second horse. We know that the word information is processed more quickly than color information. In terms of the race, the word information will get to the finish line faster than the color information. In the model, once information gets to the finishing line, that information starts to compete for the control over responding. In this case the word blue gets to the finishing line first. The word blue will then start competing, urging you to make the response BLUE. However, the response blue is the incorrect response. In order to make the correct response, you must somehow suppress the response blue and wait for the color information to provide the correct response. The word information gets to the finish line first, however the color information will eventually get to the finishing line, and at that point will start competing to control the response. In the example, once both sources of information cross the finishing line, they both compete for the final response. In order for you to make the correct response, the color information (red) must win the competition. In this example we are dealing with an incongruent stimulus. The word is competing for an incorrect response and the color is competing for the correct response. It takes some time to resolve this response competition, and for this reason response times for incongruent items are generally slow. Letʼs consider a congruent stimulus: BLUE Here we have the word BLUE printed in blue ink. What does the horse race model say about this kind of stimulus. The beginning of the race starts in the same way. The word information rushes toward the finish line, and the color information rushes toward the finish line. Just as in the above example the word information wins the race, this is because word information is processed faster than color information. As soon as the word BLUE crossed the line it starts competing for the response blue. Even though the word should be ignored, and the task does not involve word identification, for congruent trials the word provides the correct response. Eventually the color information will also cross the finish line and start competing for the response blue. After the word and color information cross the finish line they are both indicating the same response. There are two sources of information pointing toward the same response. As result there is strong evidence for the response blue, there is no response competition, and the final response can be made very quickly.

To summarize, when a Stroop item is presented both the word and color information race to the finish. The word information always wins the race because it is processed faster than color information. When the word wins the race it starts to compete for the final response. For incongruent items there is response competition because the word information points to a different response than the color information. It takes time to resolve the response competition and pick the correct response. As a result, responses to incongruent items are slow. For congruent items there is no response competition, the word and color information point to the same response, and responses to congruent trials are fast. The main difference between incongruent and congruent response times is whether or not response competition occurs. Predictions for eliminating the Stroop effect Why are we learning about the horse-race model of Stroop? There are several purposes to this lab. It is good to know about the Stroop effect, as it is a classic phenomena. More important is that we can use the Stroop effect and the horse-race model to understand how theories can be used to generate new predictions, and how we can run experiments to test predictions made by the theory. The horse- race model has two fundamental assumptions:

1. Information is processed at different rates: word information is processed faster than color information

2. Information competes for control over responding. Responses are slower when response competition occurs than when response competition does not occur

These two assumptions explain the classic Stroop effect. Incongruent items involve response competition and are therefore responded to more slowly than congruent items which do not involve response competition. We would like to go beyond this initial data and see what the model predicts for in a new situation. The big question here is how can we manipulate the Stroop task to produce a change in the size of Stroop effect. Can we think of a manipulation that would make the Stroop effect larger, or can we think of a manipulation that would make the Stroop effect disappear entirely? Second, would the horse-race model of Stroop predict these outcomes? If the model can predict these new outcomes then we would be supporting or corroborating the model. If the model can not predict these new outcome the we would be falsifying the model. The purpose of paper 2 is to run an experiment that tests predictions of the horse-race model. You will have already completed a mini-project where you tried to think of a manipulation that makes the Stroop effect bigger or smaller. In this lab we will use a tried and true manipulation that almost always makes the Stroop effect smaller, and sometimes go away entirely. Your challenge is to determine what the horse-race model predicts will happen in the new situation. What is our manipulation? You will be completing a computerized version of the Stroop task. On each trial you will see a Stroop stimulus. Your task will be to type out the correct response on the computer keyboard as quickly and accurately as you can. Some of the items will be congruent (word matches color) and some of the items will be incongruent (word does not match color). This is the first factor or independent variable in the experiment. We can call this factor congruency. It has two levels: congruent and incongruent. Remember the Stroop effect itself is the difference in reaction time between incongruent and congruent items. The main purpose of our experiment is to include one more factor and see if we can make the Stroop effect smaller. We will call the second factor Task. When you complete the Stroop experiment you will be given two different tasks: typing the color (as in

the normal Stroop experiment), or typing the word. When you complete the experiment, the computer will give you instructions for which task to complete. For half of the experiment the computer will tell you to type out the color name of the Stroop item, for the other half of the experiment the computer will tell you to type out the word name of the Stroop item. This kind of design is also known as a 2x2 factorial design. We will cover this kind of design in depth in the lecture. For now, it is enough to say that the experiment involves two factors. The first factor is congruency, and this factor has two levels: congruent items and incongruent items. The second factor is task, and this factor has two levels: name the color and name the word. This design is factorial, which means that each of levels of one factor are fully crossed with the other factor. In other words, you will receive congruent and incongruent items when you do the name color task, and you will receive congruent and incongruent items when you do the name word task. This way we will be able to measure two separate Stroop effects. We will have the Stroop effect (incongruent – congruent) for the color naming task, and we will have the Stroop effect (incongruent – congruent) for the word naming task. Our main empirical question is whether these two Stroop effects will be different. Will the Stroop effect for the color naming task be larger or smaller than the Stroop effect for the word naming task? The horse-race model actually makes a prediction about what we should find in this experiment. Your challenge is to figure out what the model predicts, and explain in your paper why it makes the prediction. Here is a hint: The model predicts that response competition starts when information passes the finish line. Remember the word information always passes the line first. The model predicts that response competition will occur differently when the task involves word identification than color identification. Why is this, and what does the horse race model predict for the size of the Stroop effect for each task?