Michael Wilson
Westat, Inc.

Jerry Lehnus
Defense Manpower Data Center

This paper evaluates the stated future plans of the Youth Attitude Tracking Study (YATS) respondents in light of subsequent behaviors. The 1991, 1992, and 1993 Fall YATS administrations included respondents from prior administrations. These data, which allow comparison of initial educational plans with subsequent behavior, reflect a strong interest in education, a correlation between plans and subsequent behavior, and the influence of circumstances on behavior. The same data also show that propensity for military service is influenced by educational prospects, and that one’s "propensity" -- the self-appraised likelihood of military service, changes with time.

Matching YATS interviews, using SSNs provided by 70% of respondents, with military records shows the relationship of propensity with enlistment behavior. The data show a strong correlation between propensity and enlistment behavior within a year of the YATS interview, and a much weaker correlation a year later.

In-depth interviews with selected Youth Attitude Tracking Study (YATS) respondents suggest that the accuracy of their future plans or expectations will vary from one individual to another, depending on the circumstances of the individual, and his or her own maturity in planning for the future. These interviews suggest some "future plans" may be more wishful thinking than realistic expectations. Future plans regarding military service, generally labeled "propensity" for military service, are a key barometer of youths’ attitudes toward military service. Thus, understanding the relationship between future plans and behavior, particularly as these relate to military service, is of interest.

From 1991 through 1993, the Youth Attitude Tracking Study included a sample of youth who had been previously interviewed. These panels provide the opportunity to relate plans to subsequent behavior. For this study, we assembled data from three cohorts:

· young men initially interviewed in the 1990 administration and reinterviewed in 1991;
· young men initially interviewed in the 1991 administration and reinterviewed in 1992; and
· young men initially interviewed in the 1992 administration and reinterviewed in 1993.

We included only young men who, at the time of the initial interview, were 16-21 years-old and had not completed more than two years of post secondary education. We also excluded high school dropouts and high school students below their senior year. A total of 4,258 young men were included.

To examine the relationship between plans and behavior, we compare respondents answers to the question "Now let’s talk about your plans for the next few years; what do you think you might be doing?" to their status the following Fall. Those who mentioned going to school are classified as having an expected status (left column of Table 1) of "Student." Those who mentioned going to work, but did not also say they expected to go to school are shown as having an expected status of "Worker." Table 1 shows the percent of respondents (weighted to reflect the population) who, at the time of the second interview, were full-time students and the percent who were non-student workers. The "Other" status includes part-time students as well as nonstudents who are not employed.

These data reflect the very strong desire of young men to get more education. Regardless of initial status, the majority of respondents indicated they planned to go to school, even those who had dropped out of college (H.S. Grad, Some College). The data also show that a young man's plans are an imperfect but significant predictor of future status. Those who said they expected to go to school became full time students more often than those who didn't. Only among full time college students does there appear to be no relationship between expected status and actual second-year status. (This might be due to a logical disconnect between the predictor and the criterion: we asked what one was likely to be doing "in the next few years," but examined behavior in the following year. College students who failed to mention school may have assumed we were asking about behavior after college graduation.)

While plans or intentions are predictive of future behavior, the data in Table 1 reflect the strong influence on one's circumstances on one's future. For example, among the high school seniors who planned to go to school, 53 percent were full time students the following Fall; among the college students, 81 percent were full time students; but among the college dropouts, only 20 percent.

Our measure of propensity is based on responses to questions in the form:

How likely is it you will be serving on active duty in the {Army, Navy, Marine Corps, Air Force}; would you say definitely, probably, probably not, or definitely not?

The question is asked with respect to each of the Services. Those who say they will "definitely" or "probably" for at least one of the four Services are said to have expressed "positive propensity" for military service; those who say they will "probably not" or "definitely not" join, along with those who say they "don't know" or decline to answer are said to have "negative propensity." Generally, we define "propensity" as the percent who express positive propensity.

It is well established that propensity is inversely related to education -- persons with more education have a lower propensity for military service (see, for example, Hintze et al., 1996). The in-depth interviews, and focus groups conducted earlier in the year, suggested a strong relationship between educational prospects and propensity. College bound youth, those with the intent, academic abilities, and funding, appear to have very little interest in enlistment. Some, however, see military service as a means to gaining a college education. Either the military would provide the necessary funding, or the military teach the discipline required to turn a poor student into a successful student. The interviews also suggested that a young man's "propensity" -- his appraisal of the likelihood he will enter the military -- changes as his circumstances change. One young man's interest in military service may evaporate as college funding materializes. Another's propensity may increase as he realizes the need for more discipline to become a successful student.

Table 2A shows the propensity of the 814 high school seniors in our sample, at the time of their initial interview ("First Year Propensity"), and at the time of the second interview the following Fall. Of the 34 percent who expressed a propensity for military service at the initial interview, only 15 percent were positively propensed the following Fall. Of the remaining 66 percent who were initially negative, 6 percent were positively propensed the following Fall. Table 2A also show that initial propensity depends on educational prospects. Of those who were to become full-time students at a four-year college or university, only 19 percent expressed a propensity for military service when they were high school seniors. In contrast, the propensity of the high school seniors who were not to go to school the following Fall was 39 percent. The propensity of those who became "other" postsecondary students -- part time students and students at two year colleges or technical schools -- was midway between those who did not become students and those who entered four-year colleges.

While the second year propensity estimates shown in Table 2A accurately reflect the respondents answers, they reflect an unrealistic drop in propensity. We know that propensity expressed in reinterviews is less than that expressed in initial interviews (Nieva, et al., 1996). The data in Table 2A suggest, for example, that the propensity of high school graduates who never attended college is 11 percent (16 percent of 39 + 8 percent of 100 - 39). In fact, the initial propensity for high school grads with no college in this sample was 30 (this statistic doesn't appear in the tables of this document). Table 2B shows second year propensities inflated to match initial propensities observed for high school grads with no college, college students, and other postsecondary students. In each case, the propensities for initially negative and initially positive respondents were multiplied by the same corrective factor, so the relative proportions of increasing to decreasing propensities remained constant.

The adjusted figures still show dramatic shifts in propensity from initial to second interview. Of the seniors who did not go to school, over half of those initially expressing positive propensity have become negative; among the seniors who did go to college, three-quarters of the positives have become negative. On the other hand, nearly a quarter of the negatively propensed seniors who did not go to college reappraised their likelihood of entering the military.

The in-depth interviews suggested two premises. First, propensity for military service is a function of educational prospects. Second, an individual's propensity changes from year to year. The data support both premises.

The first section of this paper showed plans, as well as circumstances, were related to subsequent behavior. The second showed an individual's propensity changes from one year to the next. In this section, we will examine the relationship, over time, between propensity for military service, as expressed in the YATS interview, and enlistment behavior, as recorded in administrative records.

To examine this relationship, we need a different database. YATS respondents are asked to provide a social security number. Over the years, about 70 percent of those asked have provided a number. These are matched with administrative files maintained at DMDC to identify those who have taken an Armed Services Vocational Aptitude Battery (ASVAB), signed an enlistment contract, or "shipped" to basic training. Many high school students take an ASVAB in school, without the intent of applying to enlist in the military. Those tests are not reflected here. Thus, taking an ASVAB, as shown here, is the first step in applying to enlist. For this paper, we matched 1985 through 1994 YATS files to administrative files from prior to 1985 through 1995, and examine results for 4,784 young men who, at the time of the YATS interview, were high school seniors.

Figure 1 shows the percent of high school seniors who took an initial ASVAB, according to their propensity at the time of the interview, and the months before or after the YATS interview that they took the ASVAB. For example, in the same month as the YATS interview, approximately 1.5 percent of those positively propensed took a production ASVAB; perhaps a third of a percent of those negatively propensed took an ASVAB. One to three months after the YATS interview, about 2.5 percent of those expressing positive propensity took an ASVAB, while about a half of a percent of those negatively propensed took an ASVAB. At this point in time, positively propensed high school seniors are applying to enter the military about 5 times as frequently as those who indicated negative propensity. Propensity is an imperfect but very significant predictor of enlistment behavior.

Between 12 and 24 months after the YATS interview, the difference between "positively propensed" and "negatively propensed" individuals is relatively small. Those who were positively propensed at the time of the interview take an ASVAB perhaps twice as frequently as those who were negatively propensed. This is consistent with the premise that an individual's propensity changes. At the twelfth month following the YATS interview, the propensity label applied a year earlier would, in many cases, be inaccurate.

Over half of those who initially expressed positive propensity would, if reinterviewed, express negative propensity. Of those who went to college, we'd expect three quarters to express negative propensity.

We should note that Figure 1 reflects more than differences between positively and negatively propensed individuals. As most YATS interviews are conducted in September and October, the sharp dip in ASVABs taken at month 4 would correspond to winter holidays. Applications for enlistment decrease with age, so we would expect fewer ASVABs to be taken over a year after high school graduation, regardless of YATS interviews.
____________________________________________________________
*Many individuals take the ASVAB more than once. We use the date of the initial test.

This brief examination of YATS data indicate (1) that a young men's plans, as reported in the YATS interview, whether related to education or enlistment, are imperfect but significant predictors of behavior; (2) that educational prospects affect propensity for military service; and (3) that an individual's propensity changes; sometimes young men's expectations of enlistment increase; more often they decrease. We should research these issues more.

Hintze, D.W., Lehnus, J. D., Wilson, M. J, Norris, D., & Martens, S. (1996). Youth Attitude Tracking Study 1995: Propensity and Advertising Report. Arlington: Defense Manpower Data Center.

Nieva, V., Morgenstein, D., Wilson, M., Cantor, D, Chu, A., Hintze, W., Lehnus, J. & Nathan, G. (1996). Youth Attitude Tracking Study: Panel Effects in Enlistment Propensity Estimates. Arlington, Defense Manpower Data Center.

Back to Table of Contents