A model of Air Force enlisted retention: Theoretical background, implementation, and estimates using conditional choice probability estimation

Abstract: Structural models of military retention are useful for evaluating policy alternatives that are outside historical experience. However, estimating the parameter coefficients for these models might require solving stochastic, dynamic programs thousands of times for each individual in the data. Even relatively simple models can require significant compute time, which rises rapidly with each additional explanatory variable. Although advances both in computing power and in software over the past few decades have helped increase the size of models that are feasible to estimate, a sufficiently large number of state variables can still render dynamic programming impractical. In this report, the author documents an alternative approach to estimating the coefficients of dynamic discrete choice retention models that avoids the need to solve dynamic programming problems — and, thus, the substantial compute time required to do so.

Read the full article
Report a problem with this article

Related articles

  • More for Policy & Practice

    Unveiling binge drinking trends and triggers among army personnel: A cross sectional study

    Abstract: BACKGROUND: Military populations are known to have higher prevalence and heavier alcohol use compared to the general population globally. This has serious negative implications to the military. The objective of this study was to describe the prevalence, patterns and associated factors of binge drinking among male military personnel in the Sri Lanka Army. METHODS: A cross sectional study was conducted among 1337 male Army personnel in active service using multistage sampling. A self-administered questionnaire and the interviewer-administered Alcohol Use Disorders Identification Test which is a 10-item screening tool were used. Prevalence of binge drinking was summarised as a proportion with 95% Confidence Intervals (CI). Age specific prevalence rates and the age standardized prevalence rate of binge drinking were calculated. The standard measure of one unit of alcohol being equivalent to 10 g of pure alcohol was used as a reference to calculate the units of alcohol consumption. Binary logistic regression analysis was used to determine the factors associated with binge drinking. RESULTS: The overall prevalence of binge drinking was 51.2% (95% CI 48.5%-54.0%). The age standardized prevalence of binge drinking was 28.3%. The majority binge drank once a month (50.4%). Those engaged in binge drinking used 5.6 median units of alcohol on a typical day, 84% consumed arrack, 69.3% have ever thought or attempted to quit and median age of first alcohol consumption was 18 years. When controlled for confounding, those who had mental distress (AOR 2.46, 95% CI=1.72-3.53), had sex with a commercial sex worker (AOR 1.92, 95% CI=1.21-3.06), ever smoking (AOR 1.69, 95% CI=1.27-2.25), had serious consequences (AOR 1.58, 95% CI=1.13-2.20), currently used cannabis (AOR 1.39, 95% CI=1.02-1.89) and had combat exposure (AOR 1.37, 95% CI 1.00-1.87) had a higher likelihood of binge drinking. CONCLUSIONS: The high prevalence of binge drinking warrants immediate advocacy to the highest level of command of the Sri Lanka Army for support to implement sustainable evidence-based alcohol prevention programmes.