We implemented a correlational study to evaluate the effectiveness of Rappaport's Disability Rating Scale in predicting length of hospital stay and discharge status. We evaluated 128 patients who had had cerebrovascular accidents or head injuries using Rappaport's Disability Rating Scale and documented their lengths of stay and discharge dispositions. Significant correlations were obtained between initial Disability Rating scores and lengths of hospital stay (r = .50, p less than .01), discharge Disability Rating scores (r = .66, p less than .01), and discharge status (r = .40, p less than .01). Higher initial ratings of disability were associated with higher discharge ratings and longer terms of hospitalization. Higher ratings of disability were also associated with discharge to another treatment facility rather than home. Despite this evidence supporting the predictive validity of Rappaport's Disability Rating Scale, accurate prediction of length of hospitalization remains difficult because of extreme sample variability. Areas of further investigation are considered.

This study explored the career progression of police officers in a police department by examining the relation among 13 promotional tests at four ranks administered to 1,960 candidates over two decades. It was hypothesized that performance on promotional tests would be relatively stable during a law enforcement career, consistent with research demonstrating consistency in general mental ability. Overall, 28% of subjects retested one or more times and their retest scores on non-equivalent forms correlated about .40 with their initial scores. On average, subjects increased their score by one-fifth of a standard deviation by retesting. With some exceptions, candidate performance on promotional tests significantly correlated with tests for the immediately superior rank as well as tests for higher ranks as much as 15 years later. However, using multiple regression to predict performance on higher rank tests by combining lower rank tests was unsuccessful. Implications for the management of police promotional testing programs are discussed as well as suggestions for future research.In 1989, a promotional examination for sergeant was conducted, resulting in a rank-order list with ten pages and 747 names. Of interest was the first page consisting of candidates ranked 1 through 78. Many of these individuals later promoted to leadership roles, including 15 lieutenants, 11 captains, four deputy chiefs, two assistant chiefs, and a chief. In essence, a substantial portion of a department's executive leadership in 2010 appeared at the top of an eligible list from a test conducted 20 years before.This outcome is familiar to HR analysts and managers working in civil service testing. Most of us have observed the advancement of talented candidates and the rank and file often predict who will appear at the top of eligible lists from which promotions are made. Some coworkers attribute the success to "good test taking" or "book knowledge" while others acknowledge the intense motivation of these rising stars who may study for a year or more. Not surprisingly, a recent study found selfreported motivation to correlate with test performance (McCarthy et al. 2009).Most law enforcement agencies have a well-defined rank hierarchy and therefore predefined paths for career progression. And to varying degrees, most conduct promotional examinations which result in public lists, often with ranks and scores. These eligible lists provide a rich historical record of employee career progression. Despite the availability of volumes of raw data, no one appears to have explored the relationship among sequential promotional tests administered over an entire career. That was the purpose of this study. Of particular interest was the consistency of candidate performance on tests at different ranks and the prevalence and outcome of retesting.To begin, it is clear that assessment centers (AC) are valid predictors of performance during a career. In a metaanalysis integrating the results of 34 validation studies, Arthur et al. (2003) regressed assessor ratings ...