Professor David R. Segal, University of Maryland

October 1, 1998

I want to talk about some research I’ve been involved in since before the advent of the all-volunteer force. Early in the 1970s I was asked how we can tell if a young man will join an all-volunteer force if we get rid of conscription. I said "ask them." Some military analysts said that that would not work, that high-schoolers don’t know what they will do. A recent RAND study argued that most enlistees had a low propensity to enlist. I still think they are wrong.

Since very early in the all-volunteer force era (in the early 1970s), I have been interested in what kinds of people went into the military and what happened when we made the conversion to a volunteer force. Some of this research took the form of analyses of the National Longitudinal Survey of the High School Class of 1972. In this survey we found that the military was not recruiting from the American underclass, that the military was not creaming off the best young African-Americans and that the propensity to enlist was a good indicator of enlisting. In fact, propensity to enlist was found to be the strongest factor, and proves that seniors in high school knew what they were going to do.

Since 1975, I have been working with colleagues at the University of Michigan on a project called Monitoring the Future. This is a cohort sequential survey of high school seniors who are then followed to age 35. We take an initial survey of approximately 17,000 students per year and then take a sub-sample of about 2,400 students who are tracked till age 35, although we really don’t need to track them that far after high school to understand their pattern of military service.

The result of this study is that of the 6-8% of males who say they definitely will enter the services, 60% join by two years after high school, and 70% by six years after, showing that this is a good predictor, especially for white males. The percentage of males with a propensity to enlist who actually enlist is higher among white males than among Hispanic males and considerably lower for African-American males.

For women, very few say they definitely will serve (less than 6%), and most say they definitely won’t. The likelihood of joining among high propensity youth is significantly lower (by a factor of about two) for women than for men and the relationship between propensity to join and actual enlistment is much less strong for women than for men.

Monitoring the Future has looked at 12th grade students since 1975. Since 1991, we have also looked at 8th and 10th graders. These, however, are artificial cohorts, and not the same students. The patterns are not good news for the military manpower community. The percentage who say they will definitely enlist goes down from 8th to 10th to 12th grade, and down for 8th, 10th, and 12th graders across the years 1991 to 1997. The percentage who probably will enlist and those who probably won’t enlist also goes down and the percentage who definitely won’t join goes up from 8th to 10th to 12th grades and up for 8th, 10th, and 12th graders across the years 1991 to 1997.

Trends in propensity to enter the armed forces among high school seniors from 1976 to 1997 show a long-term decline of those who say they definitely will enter the armed forces, and a more marked decline of those who say they probably will and probably won’t enter, but an increase in those who say they definitely won’t enter. When recruitment rates are high, recruiter resources get cut which leads to a one year lag in reduced enlistment propensity (resulting in wobbles in propensity).

We do not have parents’ military experience in this data set, although we do in other research at the center. There is a very high level of intergenerational inheritability of occupation in the military. People who go into the military are affected by their parents’ experience. If the parents had a good experience, their children might want to join the military, but if the parents had a bad experience, then they probably won’t want to join. We do ask those who say they will join if they plan to be officers or enlisted and which service they prefer to join.

A comparison of trends between high school seniors who are "likely to enter" and those who "want to enter" the armed forces between 1976 and 1991 showed that more males said they probably or definitely would serve than wanted to (a fact which is highly attributable to African-Americans, who continue to show this difference), while females were more likely to say they wanted to enter than that they probably or definitely would enter. This shows that something is deterring girls from entering the military, and that although we don’t know all the reasons why, that there is a promising recruiting ground there. One result we noticed is that in the year following the problems at Aberdeen Proving Ground, the propensity to enter the military did not go down for females, but instead showed a slight upsurge. It may be that they felt that all this has come out now and that it will all be fixed, so it is safe to go into the military now. Furthermore, sexual harassment was found to be low on the list of dissatisfiers for women in the military. However, an "Aberdeen effect" may appear in subsequent data.

OSD keeps asking me: "What impact on recruiting is a decision on gender integrated basic training going to have?" I feel that whatever decision gets made, it is not going to make a difference on enlistment propensity or enlistment (for women or men).

Finally, around the early 1990s, there was a marked drop in enlistment propensity. Why? The mean propensity had been dropping, mainly for African-Americans, and especially after 1990. We can’t explain the marked drop from 1990 to 1991, although one factor that may be operating is the representation of minorities in reserve units and the use of reserves in the Gulf War. We fought Vietnam with conscripts and volunteers, not the lily-white reserves. Since then, we have used reservists for overseas deployments, and more reservists are not white.

Rapporteur: Sara D. Berman

Back to Seminar Schedule, Fall 1998