IR Scores Help Predict Success

IR scores provide unique information to help predict success in business school.

Research shows IR can predict success in business school. 

  • IR adds something not measured by GMAT Q, V, A or undergraduate GPA. 
  • Results differ by program—each school should conduct a validity study.

Even though the Integrated Reasoning section is more than two years old, many schools continue to look for data and information that will allow them to pinpoint the value of the scores for their programs. Recently, more than a dozen schools have conducted validity studies to see not only how well the IR scores predict on their own, but also how much unique information they provide over what is already known about each candidate. The research has shown that IR scores provide useful additional information, but, like each of the admission factors, the importance of the scores will differ based on the characteristics of each unique program and its applicants.

Data from thousands of students were collected to assess predictive validity of each of the GMAT scores along with undergraduate grade point average. Information was standardized within each program so that data could be combined, while still accounting for the differences across the schools and the programs within them. Predictive validity is measured as a correlation, with values from 0 to 1. Figure 1 shows the comparison among different admission factors. The results show that the relationship of graduate grades with IR scores is similar to the relationship with UGPA, Verbal, and Quantitative scores. 

Figure 1
Figure 1, IR Validity

Admissions evaluate the combination of factors when considering a candidate, as opposed to just one factor. Therefore, the predictive validity of the factors should be assessed when used in combination. Adding IR scores improves the prediction, regardless of the combination of predictors used. Figure 2 shows the relative weight of each factor in the combination, which indicates the unique contribution made by the individual scores. From the combined dataset, results showed that a significant piece of the prediction pie was allocated to the unique information from IR scores. Considering that IR is a shorter section compared to the UGPA, Verbal, and Quantitative measures, these findings show quite a bit of value can be gleaned from this addition. 

Figure 2
Figure 2, IR Validity

The analyses presented in Figure 2 are based on a combination of programs. Reporting averages or results based on combined data does not show the myriad differences that could be observed for an individual, a program, or a school. In Figure 3, samples of actual program level results show that the importance of each of the scores can vary widely. 

Figure 3
Figure 3, IR Validity

Schools interested in how to use IR scores most effectively with their other admission information can conduct their own validity study through free resources at GMAC. For more information on conducting a validity study, contact Eileen Talento-Miller at talento-miller@gmac.com or 1-703-668-9788.