Value added by using personality assessments in tenant screening

Criterion validity is the ability of an assessment to predict important outcomes. The most common measure of criterion validity (also the one used in the graph) is the Pearson correlation (described in Wikipedia here). Values range from 0 (no correlation) to 1 (perfect prediction). The following are some examples that everyone can probably relate to:

• Medications: Imagine that 60% of people with a certain illness die from that disease. A medication drops the mortality rate to 40%. That 20% drop in mortality rate represents a validity coefficient of .2. Note that .2 is a lot less than the validity of personality assessments (where validity is .4), but is still significant. If you had the disease, would you want the life-saving medication? I would!
• High school grades: High school grades are one of the best predictors of college GPA. They correlate with college GPA at .4 (source: College Board Research Report No. 2001-2).

1. Personality & admissions: Data combine results of 665 studies, summarized by Ones, Viswesvaran, & Schmidt (1993). Numbers represent the correlation of the predictor with job performance (as measured by supervisor ratings, production records, sales records, and other methods).
2. References: Data summarize results from 10 studies, summarized by Hunter & Hunter (1984). Numbers represent the correlation of the predictor with job performance.
3. Interviews: Data summarize results from over 200 studies, summarized by Huffcutt & Arthur (1994), as well as Wiesner & Cronshaw (1988). Numbers represent the correlation of the predictor with job performance.
4. Financial history/credit: Data come from Oppler et al. (2008). It is the only published study to examine the relationship between a person's finances and work behavior. Participants were 2519 employees at a US government agency. Employees were identified as at risk if they were more than 180 days delinquent on any loan or had a bankruptcy/lien/legal judgment for debt against them in the last 7 years. This binary variable correlated at .13 with whether or not the person was ever identified as engaging in work misconduct (e.g., failure to pay debt, misuse of credit cards, theft or related behavior, accepting anything prohibited by law).
5. Criminal record: Data come from Roberts et al. (2007). It is the only published study to examine the relationship between a person's criminal record and work behavior. In a 26-year longitudinal study, 930 newborns of diverse socioeconomic backgrounds in New Zealand were assessed every 3 years until they reached the age of 26. The authors found that adolescent criminal convictions did not predict counterproductive behaviors at work at age 26 (average correlation of .005). In fact, people with a criminal record were actually less likely to steal at work or use things without permission. However, the authors note that adult criminal convictions may show different relationships. Interestingly, while criminal record failed to predict counterproductive behavior, several personality traits measured during childhood succeeded at around the .3 level (e.g., constraint, aggression).