Biases influencing recommendation letter contents: Physical attractiveness and gender
Nicklin and Roch (2008) present the results of a psychological experiments that looks at the effects of inflated letters of recommendation (LORs) and the effect of two well known types of bias (gender, and physical appearance) on their evaluation. The experiment involved 244 participants who evaluated candidates with a gender neutral name and identical resumes based on letters of recommendation that, through a series of pre-tests, they established were either inflated (i.e., they included words like "excellent" and "the best" and other superlatives) or non-inflated. The candidates were, in theory, applying for a job and the study participants were asked to rate the persons hirability and their likelihood of further success. In different treatments, they included photographs of the application which were either male or female and either attractive or unattractive (as measured in previous tests).
The paper tested two hypotheses:
- Hypothesis 1. Applicants with inflated LORs will (a) have a better chance of being hired; and (b) be predicted to become more successful in the company than will applicants with less inflated LORs.
- Hypothesis 2. Attractive male applicants with inflated LORs will (a) have the best chance of being hired; and (b) be predicted to become more successful in the company than all other applicants.
They found support for Hypothesis 1 as inflated letters were associated with higher evaluations along both tested criteria.
Hypothesis 2, which predicted that attractive men who with inflated letters would be considered the most hirable and and likely to succeed. was not supported, was not supported. However, there was support for an unaticipated interaction in their models. Attractive women, and only that condition, benefited relative to others when they were paired with a non-inflated letter.