Predicting Performance with Letters of Recommendation

Predicting Performance with Letters of Recommendation

What is the stated purpose of the study?

The purpose of the study was to determine if analyzing the content of letters of recommendation through the identification of individual traits would be a valid method of predicting the performance of psychology instructors and graduate students.

What factors are given as possibly affecting the validity of recommendation letters in the past?

The factors that affected the validity of recommendation letters in the past included; knowledge of the applicant whereby it was concluded that the past writing the recommendation letter either did not know the applicant or was not aware of the aspects of the applicant, reliability whereby it was alluded that there was lack of agreement between two people writing different recommendations for one person, extraneous factors like use of examples in the letters, interest of the writer to the applicant and gender. The last factor that affected the validity of recommendation letters in the fast was leniency, which contained factors like sex and race, the confidentiality of the reference, and choice of the reference.

How was this study conducted? (Subjects, procedures, predictors, criteria)

The subjects used in the study were 78 psychology graduate students and 26 graduate teaching fellows, both students from Radford University. The criteria that were used involved the examination of files from 200 former graduate students to ascertain those who had at least one letter of recommendation resulting in 78 who had at least one letter. The criteria used for the graduate teaching fellows were student rating. The procedure used for both groups involved reading every letter, highlighting the traits present in each letter, every identified trait was then placed into one of the five categories, in case two letters were present, the number of traits in each of the five categories was averaged across the two letters, and finally, the number of traits in each category was divided by the total number of traits in order to control the impact of the length of the letters and the number of traits that were used. Letters of recommendation and graduate record exams were amongst the predictors used in the study.

The article indicates that reliability among recommendation writers tends to be lower than among the readers. Why is this?

Reliability among recommendation writers tends to be lower because two people writing the same recommendation letter for a single applicant would state different character traits or give different information for the same applicant. Consequently, there would be contradicting information about the applicant.

Of the individual traits, which tend to have the highest validity in terms of predicting performance?

Mental agility was considered as the individual trait that had the highest validity in predicting performance because it had consistency with already existing data. Data that was already stored by Peres and Garcia held that graduate record exam scores were also negatively correlated with teaching success.

What does the article indicate would be good areas for future research?

The article indicates that good areas for future research would include development and validation of a trait-based checklist, an optimal number of letters that should be written by each applicant, and there should be reliable sources from which letters of recommendation should be obtained.