Using a Joystick to Increase Identification with Math

Nov 18, 08 Using a Joystick to Increase Identification with Math


Article critique of
Kerry Kawakami, Jennifer R. Steele, Claudia Cifa, Curtis E. Phills, and John F. Dovidio (2007). Approaching math increases Math=me and Math= pleasant. Journal of Experimental Social Psychology, 44, 818-825
by Alla Andelman

This study, done in two parts, tested women’s identification with math based on approach behavior training, with two other variables in two separate experiments. In the first experiment, women (44 participants total) with a low identification with math were tested against women with a high identification with math to test their implicit identification attitude and implicit positive regard attitude for math (as tested by the identification IAT and attitude IAT) after going through approach behavior training for approach verses avoidance. The results showed that women with a low initial identification had a higher implicit identification and positive regard for math when in the approach verses avoid training group. However, the women with a high initial identification showed almost equally high implicit positive regard from both training groups, and a much higher implicit identification to math for the avoid training group. In the second experiment, 56 women with a low initial identification to math were tested with an approach behavior training verses neutral behavior training, for its affect on implicit identification to math and for behavior during a math test. The results showed that women who had approach training had a higher identification response rate than women who went through neutral training. Also, women who went through approach training had a slightly higher rate of attempting more math questions on a quiz, though not doing better on this quiz. The authors concluded that if women who originally have a low identification with math are given an intervention for approach behavior, they will have an improved identification and regard for math. Also, they will be more likely to attempt to do math, hopefully improving their proficiency in math. They also theorized that women with an already high identification with math would be threatened with stereotype bias, namely that men are more proficient in math, and that will possibly hinder or destroy their identification with math if pushed to approach behavior.

The study has several very positive aspects which we should consider. Firstly, the undertaking of a very relevant social issue in today’s society is very admirable. The authors postulate that the reasons that job positions associated to the fields of math and science are filled with a large majority of men is that social views and stereotypes influence women to avoid or be disinterested in those fields. They further say that this is not due to proficiency in these fields, as has been claimed to be the case. This theory has been circulated for at least a couple of decades now, but the alternative explanation of men’s superiority is sometimes too potent in the minds of society to overcome. To put such a hypothesis to practical application and testing it would go a long way toward shining a light on this controversy. Another advantage to this study is that authors’ innovative idea that women’s identification with a subject or field would allow them to become more willing to try within that field with which they would not traditionally expect themselves to be involved. Furthermore, they claim that this identification with the field could be facilitated by an approach behavior intervention. Another particularly interesting aspect of this study is the way the authors tested identification and regard for math among women. They used an identification IAT and attitude IAT that tests the implicit attitudes that women have for math, rather than a subjective self-report of the same using reaction time to positive verses negative stimuli when paired with math.

While this is an interesting study, there are some inherent problems with it. These can be considered in two parts, one affecting the next. The first is in the authors’ initial assumptions about women and their attitudes toward math that shaped the study, and the conclusions which they draw based on these assumptions. The other is in the physical set up of the study and the question of construct validity. The initial questionable assumption that the authors discuss is that women who have a high identification with math also have a high vigilance for the stereotypes associated with the field as relating to men verses women. This is saying that women who are invested in math, and identify it as something they are a part of are very aware and wary of the stereotypes of others on their proficiency in math. This is also implying that women who do not associate math as something that is a part of themselves are not as aware of this stereotype, or are not alert to the stereotype when it presents itself. There is no proof to this assumption, nor even any theoretical backing for it in the article, and is a big assumption on which to base conclusions. This theory is used to explain the fact that women with an initially high identification with math are not improved by approach training, but sometimes even hindered by it. The explanation is that women who are pressured by stereotype threat, when pushed to identify with a field that is not stereotypically one they belong in, women de-identify with it as a result. This is quite a leap in logic, and further is not well enough explained to give a well understood and studied reason for this step to de-identify when such pressure exist. The question that many people would have is why would these women, who are wary of stereotype, and so have been familiar with the opinions of their abilities for a long time feel as though they do not belong to this field now, when given approach training? Why would they still show signs are positive regard for it, as that was unchanged, but no longer have an implicit identification to math? Why now, after having gone, presumably, a long time liking and being good at math to suddenly not identify with it because they are pushed to approach something they have already been approaching in their lives? Feeling pressure to perform in a stereotyped field where your gender is not expected to excel can hinder someone who is not already proficient in that field; that makes sense. There is research in sports showing that people who are good at a sport excel under pressure, while those who are bad at it get worse with pressure. This can be applied here. However, this assumption that the authors postulate either is not well fleshed out or is not well explained.

Now, considering the fact that initial identification with math is such an important variable in this study, it seems ill advised how casually this assessment was garnered. Initial identification in the first test is assessed based on each participant merely stating on a seven point scale if they liked math and how good they were at math. If they reported a score of 4 or higher on both measures, they were labeled as highly identified with math. Assuming that these two variables accurately measure identification with math, this is acceptable. However, if the participant reported a score of less than 4 on either measure, they were labeled as not identifying with math. Someone who is good at math, though doesn’t like do it, or doesn’t like school in general, or doesn’t like that math requires homework, whatever the reason, is very different from someone who doesn’t like it because it is hard for them. In the same vein, someone who likes math, but finds it very difficult, and they know it is not something they are inherently good at is also very different from the previous two types. Just not liking it or not being good at it alone does not mean the person is not identified with math. Someone who considers themselves having a logical mind may identify themselves with math without liking algebra or without being good at trigonometry. In the same way, the authors test initial low identification to math based on how much the participant liked five different math related activities. In the first experiment, they made note to differentiate between identification with math and regard for math, having gotten a different pattern of results. In the second experiment, they use a test of regard for identification, contradicting themselves. Clearly this main variable used in the study is not well measured enough to warrant the kinds of conclusions that the authors are claiming. Furthermore, no baseline was measured for the participants in their level of identification for math, regard for math, or test taking habits. Therefore, we can not say that low identification participants improved their identification, but that they merely had a higher identification rate than the alternate group. Since true initial identification rate has not been established, the results are questionable in content validity. Can we say that both low identification groups are the same, or that low verses high identification groups are truly different?

Another issue exists within construct validity. The moderator of ability in math is essentially overlooked in the experiment, even while discussed in the introduction and conclusion sections. Perceived ability is vaguely used as a measure of initial identification rate in experiment one, but is overlooked in whether it affects any of the results afterward. Isn’t actual ability a reasonable variable to consider when testing for identification with a field, regard for the field, and test taking behavior? The results of test taking behavior show that ability was about the same between groups given the approach treatment verses a neutral treatment, but that those who received the approach treatment attempted more questions. If we look at the numbers, they attempted about one more question than the other group, where only 12 questions existed. Furthermore, since only the average of correct answers were used as well as the average number of attempts, we must wonder whether the results would naturally be about the same, since ability is never considered, and we can naturally assume that some participants in each group was reasonable good at math on a “difficult quiz”, though difficult for whom, we do not know.

Lastly, we may consider what this approach behavior treatment technique really entails. This asks participants to move a joystick toward themselves when math symbols or words come up, verses art symbols and words, for which they move the joystick away. This is a technique the authors report has been used for a study of attitudes of prejudice towards people. This author is skeptical to accept the concept that moving a joystick towards oneself is equivalent to approaching a topic we normally do not associate with and making it part of ourselves. Perhaps there is a great deal of research proving this, but the authors of the article did not see fit to share the type of information with us. They certainly did not explain their rational for using it, besides to say it is a similar concept to the study already mentioned.

In conclusion, this study was an ambitious one. The truths we are taught to assume about ourselves based on social norms is an interesting topic. It is also quite relevant in our current environment. However, we must view the results that Kawakami et al. make, and the conclusions they draw with a grain of salt. Hopefully, more studies will follow this one, having altered some of methods of this study to eliminate some of its drawback, and therefore strengthen the conclusions that these authors propose.

Andelman, A. (2008). Using a Joystick to Increase Identification with Math (2). http://psychologyalert.com/2008/11/using-joystick-to-increase.html