F. PREVIOUS STUDIES
2. Study by Bernard (2002)
Bernard (2002) analyzes the effect of commissioning sources on promotion and retention for U.S. Navy officers. He also utilizes lifecycle-cost techniques to find out which commissioning source is more cost effective, assuming a steady-state force structure. He examines whether an officer survives until O-4 promotion board and then if he is promoted.
Bernard states that lateral transfer into the SWO community negatively effects O- 4 promotions. That is credited to the absence of ship experience, which is vital for the SWO career path. The data does not allow further investigation because there is no date information on when the lateral transfers occurred. If it does not occur too late to affect a particular career, there might be a chance to acquire experience for officers who transfer into the SWO community.
Bernard also argues that most prior studies use survival rates alone as a performance measure. This is a weakness, because there are innumerable reasons for an officer to leave the Navy, and without understating these underlying reasons, it is difficult to adopt performance measures. Officers who leave before the O-4 promotion board may leave either voluntarily or involuntarily. One of the biggest reasons for involuntary separations was the Navy personnel drawdown policy in the early 1990s. The results would be biased if the reason of separation for these officers were not dealt with explicitly.
Bernard uses data from the Navy Officer Data Card. The data includes restricted and unrestricted line officers from years through 1983 through 1990, which is merged with data from promotion board results from the years 1986 through 2001. Observations with missing information are dropped and the sample is limited to those between twenty and thirty years of age. Although dropping observations with missing values may cause
bias in the results, Bernard deletes 2,934 observations (15% of total sample) due to their omission of academic background. That leaves 17,134 URL and 5,129 restricted line observations.
Referencing Bowman (1995) Bernard suspects that examining commissioning sources as a whole may cause an aggregation bias. This bias implies that not examining each community separately may hide a community’s particular effect by aggregating it with other communities. He thus divides data into three URL communities (surface, submarine, and aviation) and one restricted-line group.
a. Model
Bernard chooses dependent variables based on whether an officer survived to the O-4 promotion board for the retention model and whether he was promoted to O-4 for the promotion model. Independent variables consist of demographics, human capital, college selectivity, community designator, and control variables for years.
The sources of commission are USNA, ROTC-scholarship, ROTC- contract, OCS, and ECP. Three “college selectivity” variables control the academic background of the officers. He uses Barron’s Profiles of American Colleges publication to classify undergraduate educations into three tiers: most selective, moderately selective, and least selective. The study also includes variables for lateral transfers, which was not analyzed in previous studies. He hypothesizes that being transferred into another community may indicate dissatisfaction with the initial community, which should have a positive effect on performance.
b. Results
The retention model consists of two results. The basic model finds that ROTC-scholarship and OCS graduates are less likely to stay up to O-4 board than USNA graduates. On the other hand, ROTC-contract and ECP graduates have the same retention rates of USNA graduates. Having technical degrees, prior enlisted service, or being male are other variables that positively affect retention. There is no significant difference
between black and white officers, but other races are less likely to stay than whites. Also, officers that are married or single with children are more likely to stay than those who are single without children.
The second retention model is similar to the first, except that it includes variables for warfare communities and college selectivity, hoping to control for college quality among accession sources. Bernard finds that non-selective ROTC-scholarship and ROTC-contract graduates are more likely to stay than USNA officers, while non- selective OCS graduates are less likely to stay. Looking at net retention rates for each accession source after grouping same-quality colleges, ROTC-scholarship and OCS graduates have a lower probability of staying by 15.6% and 9.4% respectively. The other effect of controlling for college selectivity makes the ”black” variable significant and shows that black officers are 9.6% more likely to stay.
The promotion model implements the same logit model techniques as the retention model. Bernard finds that ROTC-scholarship graduates are 8.1% less likely to promote than USNA graduates.
The promotion model for the SWO community reveals that ROTC scholarship and OCS graduates are less likely to be promoted than USNA graduates. Having a technical major increases the promotion probability by 3.3%. Having prior service and being black are not significant in the model, whereas other race officers are 8.6% less likely to promote to O-4. Also, surprisingly, neither sex nor marital status is significant. After rerunning the same model and adding the college selectivity variable, he finds that ROTC-scholarship and OCS graduates from highly selective colleges have a higher probability of promotion than USNA graduates.
3. Study by Bowman and Mehay (1999)