729
Mario Trottini and John Robertson
This paper focuses on the analysis of the factors that influenced the
reelinstment decision of the volunteers who fought
in the Union Army during the American Civil War.
We examine 15 of the companies which served
in four infantry regiment recruited in Pennsylvania.
Individual soldiers' covariates
as well
as company covariates
and regiment covariates are considered.
A two-stage hierarchical logistic
regression model with a company's random effect is used.
Based on our analysis the occupation of the
soldier and the date in which
he started his service are significant predictors of the soldier
reenlistment decision. In particular we found that
non-stakeholders are more likely to reenlist than stakeholders,
and soldiers that started their service later are more likely
to reenlist than soldiers that started their service earlier in 1861.
The evidence in our data set suggests that
other variables, such as the size of the company
or the number of battles that the company went through or
the area in which the company was recruited, did not have any
effect on the decision of reenlistment.
Some of these results
confirm results from previous studies,
as for example the non relevance of the number and type of battles for
the decision of reenlistment.
The modest sample size and the presence of potential confounding variables
not included in the study such as the marital
status of the soldier or the economic condition of his family,
suggest caution in interpreting these results.