Creswell et al. identified a physical reaction to stereotype threat and found that values affirmation reduces this reaction.

Cohen et al. performed a large-scale examination of the efficacy of values affirmation in an authentic classroom environment.

They found that having students complete a short writing exercise, in which they wrote about a value they found important (values affirmation), reduced the gap in performance between African American students and other students.

A full PDF copy of this research can be found in the Related content tab.

Good, Aronson, and Inzlicht assessed whether specific kinds of mentoring could reduce the effects of stereotype threat in female, minority, or low-income 7th graders.

They found that students paired with a college-student mentor who either 1) encouraged them to view intelligence as malleable, or 2) told them to attribute academic difficulties to the novelty of the educational setting, performed much better than students who did not get this framing.

A graphical representation (usually lines through a point on the graph that run parallel to one of the axes) showing the amount of uncertainty there is in the location of that point. All scientific data collection includes uncertainty; error bars allow researchers to show how confident they are in their results. Large error bars mean there is a lot of uncertainty (lower confidence), whereas smaller error bars mean there is less uncertainty.

A statistical term, the Greek letter beta, refers to the probability that you can accept the null hypothesis (which states that values affirmation has no effect) when in fact the null hypothesis is wrong.

The main way the authors assessed student performance was to compare scores on multiple-choice exams. These scores are referred to as the main outcome measure.

The authors used a double-blind study design, meaning that neither the students nor the teaching assistants working with the students knew the purpose of the study, or to which group each student was assigned. Double-blind studies are meant to reduce unintentional bias on the part of the participant or the researcher interpreting the data.

An initial values affirmation intervention can have long-lasting positive effects.

In one study, a group of students received a values affirmation intervention at the start of 7th grade (initial findings reported in reference 23), and their academic performance was monitored through 7th and 8th grade. African American students that received the intervention had larger and more persistent increases in grade point average than those who did not receive the intervention. This suggests that a single intervention could have long lasting positive effects.

The difference in performance (for example, on standardized tests) between minority students and white students.

Any form of an exam that requires all people taking the test to answer the same questions and is scored the same way in all cases, so that comparisons can be made between all people who take the test. These are typically multiple choice tests taken by large populations of students (for example: all 8th grade students in the United States).

Cohen et al. found that an initial values affirmation intervention could have long-lasting positive effects.

Cohen and colleagues tested whether a simple intervention could reduce the effects of stereotype threat for African American students.

They found that having students complete a short writing exercise, in which they wrote about a value they found important (values affirmation), reduced the gap in performance between African American students and other students.

The full PDF download of this paper is available in the Related content tab.

Spencer, Steele, and Quinn performed several experiments to determine whether the stereotype that men are better at math than women affected women's performance on math tests.

They found that when the gender stereotype was reinforced before a hard math test women performed even worse compared to men on the test than usual. However, describing the test as not producing gender differences eliminated the difference in performance between men and women.

Any outside force (real or perceived) that challenges a person's values, beliefs, or sense of self.

Stereotype (or identity) threat is a subset of psychological threat in which a person feels they will be judged according to common prejudices about some aspect of their identity (for example: race, ethnicity, or gender).

An experiment in which neither the participants nor the experimenters know which group each participant is assigned to until after the data are analyzed.

An intervention in which people reflect on and write about the beliefs and values (e.g., family, integrity) that are important in their lives.

A meeting of a subset of students from a larger lecture course in which students can ask questions, get clarification on lecture topics, and may solve additional problems or take quizzes (typically required with very large college courses).

Lorenzo, Crouch, and Mazur found that using more interactive teaching methods improved conceptual understanding for all students in a college-level physics class and reduced the gap in performance between women and men, compared to classes taught in the traditional lecture-based way. Classes taught with the highest level of interaction eliminated the gender gap.

Schmader, Johns, and Barquissau found that the level of endorsement of the stereotype that men are better than women at math was a moderating variable for gender performance.

Women who more strongly endorsed the stereotype performed worse.

Read more in PBS NewsHour.

The authors discuss their results and caution that values affirmation is not a perfect solution. Other education reforms are necessary.

Read more in AAAS.

Statistical measure of the strength of the relationship between two variables.

Another term for the "gender achievement gap" in which men outperform women in the same field.

High-paying.

Understanding the main ideas that make up the field.

Activities in which the student participates (as opposed to passively listening to a lecture).

Educational resources that can be incorporated into a teacher's lessons.

Short scenarios that give the student a real-world situation in which to apply their knowledge.

Stereotype threat or identity threat.

Fear and anxiety caused by the thought of having to take an exam.

Related to teaching.

The subjects that do not receive treatment.

Related to an individual's sense of identity.

Test on all material covered during the course.

The frequency of occurrence of some measure (for example: how many students got As, Bs, Cs, Ds, and Fs).

Supporting Online Materials (typically located at the end of the article).

The result deviates from that expected to arise by chance (from random variation or errors in sampling).

There are many factors that affect student performance in college, including intrinsic values which can be supported by values affirmation.

Read more in The National Academies Press.

Any activity used to modify behavior, emotional state, or feelings.

The difference in test scores, course performance, and job prospects between people of different genders.

How course material is taught, such as through context-rich problems or curricular materials.

Perception, attention, learning, memory, and problem solving.

Based on a hypothesis that may have been supported in laboratory experiments, but has not yet been shown to work in practice.

Repeating a scientific experiment and finding the same results.

Martens et al. showed that values affirmation was effective under laboratory conditions.

Schmader, Johns, and Barquissau found that women who identified more strongly with the stereotype that men are better than women at math were likely to be less confident about their own abilities and less likely to pursue math careers. They also found that women who believed the gender stereotype were more susceptible to its negative effects on math test performance.

Using more interactive teaching methods has been shown to improve conceptual understanding for students in college-level physics compared to traditional lecture-based teaching.

Interactive teaching has also been shown to reduce the gap in performance between women and men.

Kost, Pollock, and Finkelstein performed the initial assessment of the gender achievement gap in college students taking introductory physics at the University of Colorado at Boulder.

They found that most (but not all) of the gender achievement gap, could be attributed to previous preparation in the field.

A specific branch of knowledge, such as physics or biology.

Values affirmation improved performance for women, especially those who endorsed the gender stereotype.

A z-score is a measure of the number of standard deviations above or below the average score a raw, individual score is. The higher the z-score, the more different a data point is from the average.

Continuous variables have an infinite number of possible values. This is in contrast to categorical variables, which have a discrete number of defined values (for example, in this study "man" or "woman" for gender).

Men's performance on exams was not affected by the degree to which they endorsed the gender stereotype.

Standard deviation, a measure of the amount of variation in data.

It can be used to quantify how far an individual's data is from the average of a data set.

A condition in which when the value of one variable goes up (endorsement of the gender stereotype), the value of the other variable goes down (exam scores).

A function defines one variable in terms of another. Here, the more strongly a woman in the control group endorsed the gender stereotype, the lower her exam scores were.

Defining y "as a function of" x means that y varies based on the level of x.

Values affirmation was more effective for women who believed more strongly that the gender stereotype (i.e., that men are better at physics than women) was true.

Because women did better on all measures when they performed values affirmation, and men only performed worse according to one measure, it is likely that the reduction in the gender gap observed after the values affirmation intervention is due to women's scores improving rather than men's scores becoming worse. 

Men who performed values affirmation did, on average, slightly worse on course exams than men who didn't perform values affirmation. However, this was not seen on all measures of success (for example, the effect was not seen on the FMCE score).

Chi-squared, a test which tells whether there is a statistically significant difference between the distribution of two categorical variables (for example, gender).

Values affirmation resulted in a shift in the distribution of overall grades for women (but not men). In the values affirmation group, more women earned Bs (relative to Cs) than in the control group.

Values affirmation reduced the gender gap in final exam scores even though the intervention occurred several weeks before the final exam. This suggests that the effects may be relatively long-lasting.

Refers to standard error, which is a measure of how far away the mean of your data is likely to be from the true mean of the population.

Values affirmation reduced the gender gap in overall course grades between men and women.

The P value is a measure of how likely it is that your null hypothesis (that values affirmation has no effect) is true.

The authors found that the values affirmation intervention reduced the gender gap between men and women on average exam score.

In the control condition, men scored higher than women. However, in the values affirmation condition, there was no significant difference in scores between men and women.

The authors predicted that the effect of the values affirmation intervention would be greater for women who more strongly endorse the gender stereotypes.

The authors' main predication was that women who completed values affirmation would have a smaller gender gap than women who did not complete values affirmation.

Tools used to assess a subject's performance.

Assigned group, either the intervention group or the control group. In this case, the control group contained students who did not participate in the values affirmation intervention.

The authors tested whether using a values affirmation intervention in their college physics course could reduce the performance gap between men and women.

The key variable in this experiment is whether a student experiences the values affirmation intervention.

In the values affirmation intervention, students briefly write about a value they find personally important.

Schmader, Johns, and Forbes suggest that stereotype threat might affect women's performance through three interconnected mechanisms:

1) It causes a physical stress response that can impair some aspects of brain function, reducing the baseline level of mental resources.

2) It causes women to actively monitor their performance, taking valuable mental resources away from the assigned task.

3) It causes women to actively repress negative feelings and thoughts, which also consumes mental resources.

Spencer, Steele, and Quinn performed some of the original experiments showing that the stereotype that men are better at math than women affected women's performance on math tests.

Effects that result from interactions within groups, and between both groups and individuals.

Previous work has assessed the gender achievement gap in college students taking introductory physics at the University of Colorado at Boulder.

Despite the use of teaching methods thought to reduce the gender achievement gap, there remained a difference between men's and women's test scores. Most of this gap could be attributed to differences in prior preparation (for example, men were more likely to take high school physics), but not all of the gap could be attributed to these factors.

Participants are randomly assigned to different test conditions.

In this case, participants are equally likely to end up in the control group and the values affirmation test group.

Taking into account pre-existing differences in students' preparation for the class, values affirmation was highly effective in reducing or eliminating the gender gap.

Another measure of statistical difference between the means of different groups.

A statistical measure of effect size.

The t statistic is used with the T test, which determines if there is a difference between the average result of two different groups.

A secondary outcome measure was student scores on the Force and Motion Conceptual Evaluation. Because this test is administered throughout the country, the authors can compare their findings to the normal results for the population.

Students in the control group spent time writing about a value that was not important to their identity. This ensures that any difference between the values affirmation group and the control group is due to the act of self-reflection and affirmation, and not simply a result of general writing.

Level of endorsement of the stereotype that men perform better than women in physics is a moderating variable. Its value influences how much the values affirmation intervention affects course performance.

One of the key features of this study is that it takes places in an authentic college classroom, rather than being an artificial, one-time laboratory experiment.

Classes that students are required to take before taking more advanced courses in the field.

A process that may start small, but build on itself becoming larger and more serious.

Teaching strategies in which students are challenged to think about the material on their own and with other students, while the instructor acts as a guide.

Amount left unexplained by other variables.

The mean of the mean of several sub-samples.

Moderating variables are variables that affect the direction of strength of a relationship between dependent and independent variables (in this case, the level of endorsement of stereotype threat is the moderating variable).

A moderating effect is the result of the influence of a moderating variable.

A main effect is the effect of one independent variable (in this case gender) on the dependent variable (in this case beginning-of-semester FMCE scores), and ignores the effects of other independent variables.

A standardized test that is given across the country so that an average performance can be determined for the country.

The "mode" refers to the number that occurs most frequently in a set.

The "modal grade" is the course grade that occurred most frequently.