IDTIC FIL

ECO

NAVAL POSTGRADUATE SCHOOL

00
0o
Monterey, California

< --
0i

THESIS
PROFILING MARKET POTENTIAL FOR NAVY
RECRUITING AT THE LOCAL
GEOGRAPHICAL LEVEL

by

Christine Elizabeth Huzar

June 1988

Thesis Advisor: Stephen L. Mehay

Approved for public release; distribution is unlimited

DTIC
' jNOV 0 1 1988

!8 1031 I 4
FW ,UNI -

UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE

REPORT DOCUMENTATION PAGE

is REPORT SECURITY CLASSIFICAT,ON 1b RESTRICTIVE MARKINGS
UNCLASSIFIED
2a SECURITY CLASSIFICATIC-. AUTHORITY 3 DtSTRIBUTION AVAILABILITY OF REPOR-
Approved for public release;
2b DECLASSIFICATIONDOWNGRADING SCHEDULE distribution is unlimited
4 PERFORMING ORGANIZATION REPORT NUMBER(S) S MONITORING ORGANIZATION REPORT NUMBER(S)

6a NAME OF PERFORMING ORGANIZATION 6b OFFICE SYMBOL 7a NAME OF MONITORING ORGANIZATION

(If applicable)
Naval Postgraduate School Code 54 Naval Postgraduate School
6c. ADDRESS (City, State, and ZIPCode) 7b ADDRESS (City, State, and ZIP Code)

Monterey, California 93943-5000 Monterey, California 93943-5000

8a. NAME OF FUNDING, SPONSORING 8b OFFICE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
ORGANIZATION (If applicable)

8c. ADDRESS (City, State, and ZIP Code) '0 SOURCE OF FUNDING NUMBEDS
PROGRAM PROJECT TASK WORK UN(T
ELEMENT NO NO NO ACCESSiON NO

11 TITLE (Include Security Classification)

PROFILING MARKET POTENTIAL FOR NAVY RECRUITING AT THE LOCAL GEOGRAPHICAL LEVEL

12 PERSONAL AUTHOR(S)
Huzar, Christine E.
13a TYPE OF REPORT
Master's Thesis
16 SUPPLEMENTARY NOTATION
I,3b
IROM
TIME COVERED
TO
14 DATE OF REPORT (Year, Month, Day)
1988, June T
115 PAGE COUN T
80

The views expressed in this thesis are those of the author and do not reflect the official
policy or position of-the De t of Defese or the
17 COSATI CODES . 18 SUBJECT TERMS (Continue on reverse if necessary and identify by block number)
FIELD GROUP SUB-GROUP Recruiting; Enlistment; Intention;
Estimates of Enlistment Market Potential)
V9 AB TRACT (Continue on reverse if necessary and identify by block number)

-This thesis investigates several alternative methods for estimating intentions to join
the United States Navy. The Youth Attitude Tracking Study (YATS) is used to obtain the
intentions of young male respondents to join the military, and specifically the Navy.
Intention propensity indexes are calculated for Navy recruiting areas and districts.
The main conclusions of the study are:
a. intention propensity can be forecasted at the Navy recruiting district level;

b. a propensity index could be used to allocate the number of recruiters and recruiter
goals at the Navy recruiting area and district level;
c. probit and logit regression models should be tested by predicting enlistment intentions
for 1985-1987, then comparing the prediction against observed out-of-saple years.

20 DISTRIBUTION /AVAILABILITY OF ABSTRACT 21 ABSTRACT SEIURITY CLASSIFICATION

, UNCLASSIFIED/UNLIMITED 0 SAME AS RPT 0 DTIC USERS Unclassified
22a NAME OF RESPONSIBLE INDIVIDUAL 2Zb TELEPHONE (Include Area Code) ,2 OFFICE SYMBOL
Prof. Stephen L. Mehay (408) 646-2643 Code 54Mp
DD FORM 1473, 84 MAR 83 APR ed,tion may be used until ehausted SECURITY CLASSIFICATION OF THIS PAGE
All other editions are obsolete , u $ Coe, 11t P-ti.9 Ofl, 1,66-60O,243
i UNCLASSIFIED
 Approved for public release; distribution is unlimited

Profiling Market Potential for Navy Recruiting

at the Local Geographical Level

by

Christine E. Huzar
Lieutenant, United States Navy
B.A., State University of New York College at Buffalo, 1974

Submitted in partial fulfillment of the

requirements for the degree of

MASTER OF SCIENCE IN MANAGEMENT

from the

NAVAL POSTGRADUATE SCHOOL

June 1988

Author:
Christine E. Huza

Approved by:
Stepherf L. Mehay, Thesis isor

les I. Borack, Se nd Reader

David A W'airman

Department of ministra ive Sciences

Jam~s Fe
A7/tffi
Dean !f7. nforma ion !nd Policy Sciences

ii

47
 0w

ABSTRACT

This thesis investigates several alternative methods for

estimating intentions to join the United States Navy. The

Youth Attitude Tracking Study (YATS) is used to obtain the

intentions of young male respondents to join the military,

and specifically the Navy. Intention propensity indexes are

calculated for Navy recruiting areas and districts.

The main conclvisions of the study are:

a. intention propensity can be forecasted at the Navy

recruiting district level;
b. a propensity index could be used to allocate the
number of recruiters and recruiter goals at the Navy
recruiting area and district level;

c. probit and logit regression models should be tested by

predicting enlistment intentions for 1985-1987, then
comparing the prediction against observed out-of-
sample years.

Accession For
NTIS GRA&I
DTIC TAB
Unannounced El
Justificatio ncr1

Byop

Distribution/
Availability Codes
AvaiA-an-d/or
Dist Special
 TABLE OF CONTENTS

I. INTRODUCTION AND LITERATURE REVIEW 11--------------

A. PROBLEM AND BACKGROUND 11----------------

B. U.S. NAVY GOAL ALLOCATION MODEL 44-------------

C. U.S. MARINE CORPS GOAL ALLOCATION MODEL 8

D. U.S. ARMY ENLISTMENT PROJECTION MODEL ------- 10

E. PREVIOUS PROPENSITY STUDIES AND

THEIR FINDINGS ------------------------------- 11

F. OBJECTIVE ----------------------------------- 14

II. DESCRIPTION OF DATA FILE ------------------------ 16

III. METHODOLOGY ------------------------------------- 18

A. METHODS USED -------------------------------- 18

B. ESTIMATION PROBLEMS ------------------------- 20

IV. RESULTS ----------------------------------------- 21

A. INTENTION TO JOIN THE MILITARY -------------- 21

B. SPECIFIC INTENTION TO JOIN THE NAVY --------- 32

C. SPECIFIC INTENTION TO JOIN THE ARMY --------- 40

D. INTENTION TO JOIN THE MILITARY BY STATE ----- 45

E. LOGIT/PROBIT ANALYSIS ----------------------- 45

V. CONCLUSIONS AND RECOMMENDATIONS ----------------- 58

A. CONCLUSIONS --------------------------------- 58

B. RECOMMENDATIONS --- -------------------------- 58

APPENDIX A: NAVY RECRUITING AREAS AND DISTRICTS ------ 60

APPENDIX B: NAVY RECRUITING DISTRICT SAMPLE SIZE ----- 62

i
iv
 APPENDIX C: CONTINGENCY TABLES, COMPARISON OF
ACTUAL AND PREDICTED INTENTION TO
JOIN THE MILITARY BY YEAR GROUP ---------- 64

APPENDIX D: CONTINGENCY TABLES, COMPARISON OF

ACTUAL AND PREDICTED INTENTION TO
JOIN THE MILITARY BY AREA ---------------- 66

APPENDIX E: COMPARISON OF ACTUAL AND PREDICTED

INTENTION TO JOIN THE MILITARY ----------- 69

LIST OF REFERENCES ------------------------------------ 70

INITIAL DISTRIBUTION LIST -- ------------------ 71

6v

6 rT[ -%' -' i " z" , .' i . "' 2. ''<

 LIST OF TABLES

3.1 DESCRIPTION OF INDEPENDENT VARIABLES ------------ 19

4.1 INTENTION TO JOIN THE MILITARY BY AGE ----------- 22

4.2 INTENTION TO JOIN THE MILITARY BY HIGH

SCHOOL GRADUATE STATUS -------------------------- 22

4.3 INTENTION TO JOIN THE MILITARY BY RACE ---------- 22

4.4 INTENTION TO JOIN THE MILITARY BY CURRENT

SCHOOL STATUS ----------------------------------- 23

4.5 INTENTION TO JOIN THE MILITARY BY SELF-

REPORTED AVERAGE GRADE IN HIGH SCHOOL ----------- 23

4.6 INTENTION TO JOIN THE MILITARY BY NUMBER

* OF MATH COURSES COMPLETED ----------------------- 24

4.7 INTENTION TO JOIN THE MILITARY BY FATHER'S

HIGHEST LEVEL OF EDUCATION ---------------------- 24

4.8 INTENTION TO JOIN THE MILITARY FOR NAVY

RECRUITING DISTRICTS BY YEAR -------------------- 25

4.9 PROPENSITY INDEX OF GENERAL INTENTION TO

JOIN THE MILITARY BY NAVY RECRUITING AREA
AND DISTRICT ------------------------------------ 30

4.10 PROPENSITY INDEX OF GENERAL INTENTION TO

JOIN THE MILITARY BY ARMY RECRUITING
BATTALION AND BRIGADE --------------------------- 33

4.11 PROPENSITY INDEX OF INTENTION TO JOIN THE

NAVY BY NAVY RECRUITING AREA AND DISTRICT ------- 35

4.12 COMPARISON OF MILITARY PROPENSITY INDEX AND

NAVY PROPENSITY INDEX BY NAVY RECRUITING
AREA AND DISTRICT ------------------------------- 38

4.13 COEFFICIENT OF VARIATION OF PROPENSITY INDEX ---- 40

4.14 PROPENSITY INDEX OF INTENTION TO JOIN THE

ARMY BY ARMY RECRUITING BATTALION AND BRIGADE --- 41

vi

0n
 4.15 COMPARISON OF MILITARY PROPENSITY INDEX AND
ARMY PROPENSITY INDEX BY ARMY RECRUITING
BATTALION AND BRIGADE --------------------------- 43

4.16 PROPENSITY INDEX OF GENERAL INTENTIONS TO

JOIN THE MILITARY BY STATE ---------------------- 46

4.17 PROBIT COEFFICIENTS BY YEAR --------------------- 48

4.18 PROBIT COEFFICIENTS BY NAVY RECRUITING AREA ----- 51

4.19 LOGIT REGRESSION COEFFICIENTS ------------------- 57

I-

, vii
 I. INTRODUCTION AND LITERATURE REVIEW

A. PROBLEM AND BACKGROUND

When President Nixon abolished the draft in 1973, he

opened a Pandora's Box of questions for the military. One

critical question the military had to address was how to

attract qualified youths into the all-volunteer military

(AVF). Advertising techniques and recruiting methods all

assumed greater inportance in the AVF environment.

Recruiting issues included recruiter goal allocation and

determining market share of the "high quality" male youth

population.

Most recruiting efforts today are concentrated on highly

qlfi:d non-p-Xor ze-vice male youths. These are 17-21

year old males who have graduated from high school and are

classified as I-IIIA by the Armed Forces Qualification Test

(AFQT). Those individuals who do not sss at least a

high-school diploma are approximately twice as likely not to

P complete their initial enlistment contract whereas those

that score at or above the median Armed Forces Qualification

Test score are more likely to complete technical training.

[Ref. l:p. 225] This group of individuals is considered

"supply constrained" and substantial effort is required to

recruit the quantity necessary to maintain combat

readiness. Females, non-high school graduates, and

 F FU
' ,- , VWW nX*V WrV 1;1 WA"YWMW C%. 7L,7 ,sv

individuals classified as IIIB and below by the Armed Forces

Qualification Test, are considered to be "demand

constrained." The requirements for these groups are such

that the supply is in excess of the services' goals.

In recent years, only the Army has occasionally failed

to meet its goals for highly qualified non-prior service

male youths; and this failure was a small miss occurring

several years ago. All the other services have been able to

recruit the quantity of highly qualified non-prior service

male youths that are desired to meet mission readiness.

There has been some concern in recent years that the

services will have problems reaching their recruiting goals

for highly qualified non-prior male youths. The United

States Bureau of the Census has projected that there will be

a steady decrease in the male youth population through the

mid-1990s [Ref. 2]. This means the military will be

competing with the private sector for its share of a

decreasing supply of 17-21 year old males.

The decline of the 17-21 year old male population has

not been the only factor potentially affecting military

manpower supply. Acquired Immune Deficiency Syndrome (AIDS)

appears to be a growing problem. Current military policy

excludes potential recruits who test positive for the

Acquired Immune Deficiency Syndrome virus during their

initial enlistment physical. Many military leaders are

concerned that Acquired Immune Deficiency Syndrome will also

% V '
 greatly reduce the number of qualified male youth in the

1990s and possibly into the next century. In the face of

these concerns, one response would be to abolish the all-

volunteer force and reinstate the draft. However,

reinstitution of the draft itself presents numerous

problems. It is far more productive to have a military

force that is willing to serve, than a military force of

conscripted men who may hdve attitude problems [Ref. 3:p.

64]. Attitudes in the work place contribute a great deal to

productivity. Low morale and dissatisfaction could reduce

the combat readiness of the force [Ref. 3:p. 23]. Of course

the draft may be necessary at some time in the future,

especially if significant hostilities occur.

Another solution to the declining male youth population

is to place more women in jobs currently filled by men.

This solution, however workable, does not appear to be

popular with the public--at lease in terms of placing women

in combat positions [Ref. 4]. If the services cannot meet

their recruiting goals for highly qualified male youths,

.4.

public opinion may change when faced with the options--

drafting men or placing women in combat positions.

If the services are not treeting their recruiting goals,

this does not necessarily mean that these goals are

unattainable. It is perhaps the way recruiting goals are

distributed to the various recruiting commands that is at

fault and not a shortage of supply. Under current goal

3
 allocation models, it is quite possible that one command's

goals are set too high while another command's goals are too

low, relative to the potential supply in the area. If the

goals are set too high for the area in which the command is

located the recruiters will be unable to reach those goals,

which will indicate a shortage of supply. If the goals are

set too low for the area, the recruiters will not have the

incentive to recruit much above the established quota. As a

result, there may be an untapped supply of possible

enlistees. In order to avoid a false perception of
available supply, it is critical for the services to develop

the best model possible to allocate recruiting goals down to

their smallest components. A current solution to the

declining male youth population, and one that could be

implemented almost immediately, is to enhance the efficiency

of fixed recruiting resources by improving recruiter goal

allocation and recruitment methods.

B. U.S. NAVY GOAL ALLOCATION MODEL

The U.S. Navy currently uses different enlisted goal

allocation models for various subpopulations based upon

ethnic group, gender and other factors. A different model

is used for prior service members, for non-prior service

females, and for non-prior service males. The non-prior

service male group is further sub-divided into four

different models: high school diploma graduate/Armed Forces

Qualification Test score 50-99; high school diploma

a' 4
 II

graduate/Armed Forces Qualification Test score 30-49; black

upper mental groups; and Hispanic upper mental groups. A

score of 50-99 on the Armed Forces Qualification Test would

be equivalent to I-IIIA, also known as "A-cell group" or

"upper mental groups." An Armed Forces Qualification Test

score of 30-49 would equate to IIIB, also known as "Cu-cell

group."

For purposes of this analysis, I will be concerned only

with the non-prior service males in the upper mental groups.

The regression model currently used by the Navy Recruiting

Command [Ref. 5:p. 5] to forecast the number of new

contracts for non-prior service males in the A-cell group is

as follows:

log C = A + r log R + u log U + p log P + n log N

where:

r = recruiter elasticity,

u = unemployment elasticity,

p = respective 'A' cell population elasticity,

n = respective non 'A' cell population elasticity,

C = forecast of new contracts,

A = constant,

R = number of projected on-board recruiters,

U = projected unemployment,

5
 P = projected 17-21 year old A-cell population,

N = projected 17-21 year old male non A-cell

population.

This regression model is used to forecast new contracts

on the national level which then is also used to distribute

quotas to the Navy recruiting areas/districts (except blacks

and Hispanics in upper mental groups). The Navy divides the

nation into 6 areas which are further divided into 41 Navy

recruiting districts. Appendix A lists the Navy recruiting

areas and districts.

The regression model [Ref. 5:p. 12] used for forecasting

black new contracts for non-prior service males in the A-

cell group, is as follows:

log CB = A + r log R + u log U + n log BP + log B

where:

r = recruiter elasticity,

u = unemployment elasticity,

n = 'A' cell black population elasticity,

CB = forecast of new contracts for blacks,

A = constant,

R = number of projected on-board recruiters,

U = projected unemployment,

BP = projected 17-21 year old male A-cell black

population,

B = percent black.

6
 The regression model [Ref. 5:p. 13] used for forecasting

Hispanic new contracts for non-prior service males in the A-

cell group, is as follows:

log CH = A + r log R + u log U + s log S + log H

where:

r = recruiter elasticity,

u = unemployment elasticity,

s = 'A' cell hispanic population elasticity,

CH = forecast of new contracts for Hispanics,

* A = constant,

R = number of projected on-board recruiters,

U = projected unemployment,

S = projected 17-21 year old male A-cell Hispanic

population,

H = percent Hispanic.

According to the Navy Recruiting Command the above

models are extremely accurate at the national level and

accurate at the area level, but not very accurate at the

district level. [Ref. 6]

One dependent variable that may be significant for

predicting new enlistment contracts is the employment plans

or military enlistment intentions of male youth within a

specific local area, such as a Navy recruiting district. If

this variable is significant it may help to increase the

7
 accuracy for forecasting new enlistment contracts at the

Navy recruiting area and district level. The purpose of

this thesis is to investigate the use of military enlistment

intention data at the local level.

C. U.S. MARINE CORPS GOAL ALLOCATION MODEL

The U.S. Marine Corps currently uses intentions to join

the Marine Corps to calculate the percent of national quota

to assign to a given recruiting area, the percent of the

"interested" market in the area and the percent recruiter

distribution for the region. The "special market" is the

estimation of Qualified Military Available (QMA) taking into

account mental category accession goals. The QMA is defined

as the population of 17-21 year old male high school

graduates available for service in the military. The

equation used is:

% of National Quota
or % of Special Propensity
% Recruiter Distribution = Market X Index
or
% of the Interested Market

The propensity index is measured using four variables:

1. Propensity from the Youth Attitude Tracking Study

(YATS).

2. Priority Prospect Card (PPC) return rate.

3. Unemployment rate.

4. Productivity rate.

L Vo....-,."

AAi
The YATS survey is used to produce a YATS propensity index.

District Positive USMC Propensity Rate

YATS Index = National Positive Propensity Rate

The PPC is used to develop a PPC index. First, the

percent quality leads are calculated by dividing the number

of quality leads by the volume of mailings. The PPC index

is then determined using the following:

District % Quality Leads

PPC Index =National % Quality Leads

The unemployment index is determined by dividing the

district unemployment rate by the nationwide unemployment

rate.

Recruiter productivity is determined by dividing the

number of new contracts from prior years by the table of

organization of recruiters. The district productivity

divided by the national productivity produces the

productivity index.

There are three QMA categories for 17-21 year old male

high school graduates: I-IIIA, IIIB and IV. To determine a

district's share of the special market for I-IIIAs, the

district's QMA is multiplied by .63 percent. The result is

the total I-IIIA QMA for the district. The following

equation then is used to determine their percent of the

special market.

9
 Total I-IIIA for District
% I-IIIA Market = Total I-IIIA Nationally

The same process can be applied to the IIIB and IV

mental category groups. The district's QMA is multiplied by

.36 and .01 respectively to determine the district's share

of IIIBs and IVs. [Ref. 7]

D. U.S. ARMY ENLISTMENT PROJECTION MODEL

The U.S. Army uses a linear regression procedure to

develop a production forecasting equation for their five

Army recruiting brigades and three mission categories. The

equations are used to produce a forecast for the next four

quarters.

The dependent variables are:

1. I-IIIA Male Market (Seniors + 2 yrs' grads),

2. IIIB Male Market (Seniors + 2 yrs' grads),

3. I-IIIA Female Market (Seniors + 2 yrs' grads).

The independent variables used in the regression

forecasting equation are:

1. Army on-production recruiters,

2. Other-service recruiters,

3. Unemployment,

4. Army recruiter experience factor,

5. Army enlistment propensity,

6. Market data,

10
 NAN

7. Dummy variable for quarters (seasonality),

8. Dummy variables for battalions within brigades.

The regression procedure produces 15 estimating

equations (five brigades times three dependent variables).

[Ref. 8] Although the regression model provides a

technique for forecasting likely enlistment levels, there is

some subjective judgement included in the final decision of

goal allocation.

E. PREVIOUS PROPENSITY STUDIES AND THEIR FINDINGS

The Department of Defense sponsors an annual national

Youth Attitude Tracking Study (YATS) survey. The purpose of

the survey is to gain knowledge about the impact of

recruiting and advertising programs, and to estimate current

interest in the military service. [Ref. 1:p. 225]

There are two types of questions asked to determine

interest in military service. One question could be

referred to as "unaided" mention. The question is asked,

"What do you think you will be doing in the next few years?"

If the respondent states that he intends to join the

military service, he is considered to have an unaided

mention. Another question directly asks, "How likely is it

that you will join the military service in the next few

years?" The respondent can answer: definitely will join,

probably will join, probably will not join, or definitely

will not join. If the respondent answers definitely or

probably will join, this is considered to be an "aided"

11
mention but if he responds with probably will not or

definitely will not join, this is considered a negative

intention. [Ref. 9:p. 8]

Bruce R. Orvis conducted several studies for The Office

of the Assistant Secretary of Defense/Manpower, Installa-

tions and Logistics. In his research he attempted to show

the extent to which stated intentions on the YATS survey

relate to actual enlistments. In Orvis' 1982 study, the

data suggested that enlistment intentions measured in the

Youth Attitude Tracking Study do a good job of indicating

the probability that the respondent actually will apply to

enlist. The data also indicated that many Youth Attitude

Tracking Study respondents make their enlistment decisions

several years after the survey. The predictive power of the

Youth Attitude Tracking Study intention measures continues

up to 4 years after the respondent answers the survey. The

most accurate predictions, however, are obtained within the

first 12 to 18 months following the survey. [Ref. 9:p. 8]

In 1985, Orvis continued his research on the relation-

ship between intentions and actual enlistments by evaluating

whether this intention information conveys more about an

individual's likelihood of enlisting than demographic

characteristics. The data suggested that intentions do

provide information about an individual's probability to

enlist beyond that available from demograpnic factors.

Respondents that have the strongest enlistment intentions

12
(unaided mention and aided mention) have a 37 percent
enlistment rate, while respondents that had only positive
enlistment intentions (aided mention only) enlisted at a
rate of only 15 percent. The negative intention group had
only a 5 percent enlistment rate. Though a higher percent
of the strongest and positive intention groups enlisted, 46
percent of all enlistees from the sample group were from
the negative intention group. Orvis states that a small
increase in the enlistment rate of the negative intention
group will provide a significant increase in the number of
actual enlistees. [Ref. 10]
Based on the findings of Orvis, Gregory D. Citizen V.

conducted research to determine local area estimates of

market potential for the Army, Navy, Air Force and Marine
Corps, using intention data from the Youth Attitude Tracking
Study survey. The local areas used by Citizen equate only
to the recruiting area level in the Navy. Therefore, his
findings provide no new tool for goal allocation at the
district level. His results indicated that the Air Force

received highest positive propensity for all ages and for

areas, except the southwest and mideast, where the Navy was
preferred. In general he found that the propensity to

enlist was highest in the southeast followed by the

northwest, northeast, west and southwest respectively. The
propensity to enlist in the Navy and Marine Corps
specifically, was highest in the southeast and southwest.

13

.i
 F
U1K~.MJP MJIAiTAR An VW W" F 7 X~WW1

The Army and Air Force had the highest propensity to enlist

in the northwest and southeast. One recommendation made is

that further study should be conducted for smaller areas.

[Ref. 11]

Jules Borack used Youth Attitude Tracking Study data to

develop a profile for the high-quality youth market. The

independent variables used followed some of Orvis' early

work on "high quality." In his logit analysis of the "high-

quality" military market, Borack included as independent

variables educational status, number of math and science

courses completed in high school, self-reported grade point

average, father's education, race and region.

The analysis was conducted in two stages. First, Borack

determined the probability that an individual is a member of

the "high-quality" group. Second, lie estimated the mean

value of a trait of "high-AFQT" individuals that intend to

join the military. He concluded that his two-stage

regression-based technique would be useful for estimating
population parameters when group membership is unknown and

the services require descriptions of different market

sectors. [Ref. l:p. 226]

F. OBJECTIVE

As stated previously, it is critical that recruiter

allocation goals be distributed based on market potential to

ensure the greatest possible market penetration. If current

recruiter allocation goals are not appropriate across

14

S
recruiting districts, it would be beneficial to identify

additional variables that will aid in establishing recruiter

goal allocation. From previous studies there is a strong

indication that a person's stated intentions on the YATS

survey provide a strong indication of propensity to enlist.

There have been relatively few studies that used intentions

to predict enlistment, and of those that have explored this

area, none have provided a useful model for predicting new

contracts at the Navy recruiting district level. The first

step in using intention as a predicting variable for new

contracts is to be able to predict intentions for recruiting

districts. From this, an intentions "index" could be

developed and included in the regression-based forecast of

new contracts in the recruiting districts.

The objective is to provide the Navy Recruiting Command

with a better tool to distribute goals to the Navy

recruiting areas based on forecasts of new contracts in each

Navy recruiting district. The districts would be assigned

goals that are challenging yet attainable based on the

underlying military propensity or "taste" in the area and

other demographic characteristics. Recruiting manpower

would be less likely to be wasted in areas that have a low

potential for enlistment while understaffing areas with

higher potential would also be avoided.

15
 II. DESCRIPTION OF DATA FILE

Currently the Youth Attitude Tracking Study survey is

administered annually to approximately 5,000 males, 16-to-21

years old, nationally to determine their educational

background, work experience, and future plans, including

attitudes and intentions about military service. During

certain periods the Youth Attitude Tracking Study was

administered semi-annually. To obtain a sample size large

enough to be representative of the population in the 41
0 Naval recruiting districts, all the Youth Attitude Tracking

Study waves from 1976 to 1984 have been merged into a single

file.

The total sample size for the nine year period is 82,013

cases. Blacks represent 12.5 percent of the sample. The

question on ethnic background varied from year to year

making it impossible to consistently identify ethnic groups

other than blacks and whites. Sixty-eight percent of the

samplc were still in high school at the time they answered

the survey. Of those not in high school, 77 percent are

high school graduates.

The social security numbers of respondents were matched

against the Military Entrance Processing Station (MEP) files

to include actual accession data. From the MEP files

information was obtained on: Armed Forces Qualification

16

0 <. r .t' _ '_ - o,-- , - - - - .-. .. . .

 I

Test (AFQT), Delayed Entry Program (DEP), Delayed Entry

Program discharges and actual accessions.

All waves of the Youth Attitude Tracking Study survey do

not contain precisely the same questions. Therefore, it was

necessary to recode like questions that could be answered

with a similar response and eliminate questions that were

not used in all waves of the survey. Fortunately all Youth

Attitude Tracking Study surveys used similar phrasing for

the intention questions. Most of the questions on

education, work, and father's education were the same for

most years. Therefore, a large overall sample size could be

analyzed on these variables in smaller geographic locations.

Appendix B contains the sample sizes for each Navv

Recruiting District. The data from the MEPS files are

consistent for all years and required no recoding for
matches with the Youth Attitude Tracking Study.

J.

17

,0-o, ,.o. .. ,,--;. <<' '. '. < :i< < :':'.-O o.
 III. METHODOLOGY

A. METHODS USED

Several methodologies will be used to analyze the data.

First, probabilities of intentions for each district must be

determined for: definitely will join, probably will join,

probably will not join and definitely will not join the

military. The dependent variable, intention, was divided

into two groups. The responses "definitely will join" and

"probably will join" were combined into "will join,"
representing the positive propensity group. The responses

"probably will not join" and "definitely will not join" were
combined into "will not join," representing the negative
propensity group. An intention dummy variable was created

and set equal to 0 for the "will not join" group and equal

to 1 for the "will join" group. The independent variables

used to predict intentions will be age, education status,

number of math courses, self-reported grade point average,

father's education, race, and geographic region. Table 3.1

describes the independent variables and the coding used.

These are the explanatory variables identified by Orvis,

Borack and Citizen to be significant in predicting

enlistments. Each Navy recruiting district was recoded as a

dummy variable. This produced 40 independent variables for

Navy recruiting districts.

18
SM

TABLE 3.1
DESCRIPTION OF INDEPENDENT VARIABLES

Variable Description Code

AGE 16-21 Year Old Males AGE IN YEARS

Non-Prior Service

RACE Ethnic Group 0 WHITE

1 BLACK
CURSCH Currently in 0 YES
High School 1 NO

GRADHS High School 0 YES

Graduate 1 NO
AVGGRD Average Grade 1 LESS THAN D
in High School 2 Cs AND Ds
3 Bs AND Cs
4 As AND Bs
MATH Number of Math 1 ONE
Courses Completed 2 TWO
3 THREE
4 FOUR
FATHER Father's Highest 1 LESS THAN H.S.
Education 2 HIGH SCHOOL
3 MORE THAN H.S.

A NRDi Navy Recruiting 0 'LL OTHERS

District i 1 NRDi

Source: Derived from data provided by the Youth Attitude

Tracking Study, 1976-1984.

A probit regression was conducted on the independent

variables excluding the Navy recruiting districts dummy

variables, in three separate year groups. The year groups

were 1976-1978, 1979-1980 and 1981-1984. The groups were

K 19

Se
determined by the sample size rather than by years, due to

the limited capability of the computer package used.

Next, a probit regression was conducted on the

independent variables listed previously, for each Navy

recruiting area. This produced six separate regression

equations. From this an attempt is made to identify an

individual's positive or negative intentions to enlist given

the information on the independent variables.

Finally, a logit regression was conducted on the

independent variables including the Navy recruiting district

dummy variables. This model contains 47 independent

variables.

B. ESTIMATION PROBLEMS
The SPSSx package was used for all statistical analysis

except the final logit regression equation. Due to the

large sample sizes and the large number of independent

variables a logit or probit regression could not be

performed on the full sample with SPSSx. Thus, SAS was

used to perform the final logit regression on the full data

set (N = 82,013).

20
 IV. RESULTS

A. INTENTION TO JOIN THE MILITARY

The first step in the analysis involved developing a

simple cross-tabulation of intention by various demographic

attributes. Table 4.1 clearly demonstrates that as age

increases intention to join the military decreases. Table

4.2 indicates that non-high school graduates are twice as

likely to be interested in the military compared to high

school graduates. This table does not include those

individuals that were currently in high school.

Table 4.3 indicates that the positive propensity of

black males is approximately twice that of white males.

Table 4.4 indicates that respondents currently in high

school have a significantly higher intention to join the

military than those not in high school. This may be

attributed to the fact that those not in high school are

likely to already have a job, while those still in high

school are somewhat uncertain about there future employment.

Table 4.5 presents the propensity to join the military

by self-reported average grade in high school. The table

shows a general increase in intention to join the military

as average grade decreases. Intention of the less than D

group is slightly lower than the C's and D's group.

21
 TABLE 4.1

INTENTION TO JOIN THE MILITARY

BY AGE

Age
16 17 18 19 20 21

Will Join 33.7% 29.1% 22.0% 17.0% 14.6% 13.0%

Source: Derived from data provided by the Youth

Attitude Trackin! Study, 1976-1984.

TABLE 4.2
INTENTION TO JOIN THE MILITARY
BY HIGH SCHOOL GRADUATE STATUS

High School Non-High School

Graduates Graduates
Will Join 14.8% 31.2%

Source: Derived for data provided by the Youth Attitude

Tracking Study, 1976-1984.

TABLE 4.3

INTENTION TO JOIN THE MILITARY

BY RACE

Race
White Black

Will Join 21.3 % 40.2%

Source: Derived from data provided by the Youth

Attitude Tracking Study, 1976-1984.

22
 TABLE 4.4

INTENTION TO JOIN THE MILITARY

BY CURRENT SCHOOL STATUS

School Status

Currently in Currently Not in

High School High School

Will Join 26.4% 19.7%

Source: Derived from data provided by the Youth

Attitude Tracking Study, 1976-1984.
C

TABLE 4.5

INTENTION TO JOIN THE MILITARY

BY SELF-REPORTED AVERAGE GRADE
IN HIGH SCHOOL

Average Grade
A's & B's B's & C's C's & D's Less than D

Will Join 16.6% 26.2% 32.4% 31.3%

Source: Derived from data provided by the Youth

Attitude Tracking Study, 1976-1984.

Table 4.6 indicates that as the number of math courses

completed increase, the lower the intention to join the

military. (There may be some correlation between the number

of math courses completed and the self-reported average

grade.)

Table 4.7 indicates that father's education level may

have some influence on intention to join the military. The

23

~ .* ~4
table shows an increase in intention to join the military as

the father's education level decreases.

TABLE 4.6
INTENTION TO JOIN THE MILITARY
BY NUMBER OF MATH COURSES COMPLETED

Number of Math Courses

One Two Three Four

Will :oin 29.5% 24.5% 18.8% 13.7%

Source: Derived from data provided by the Youth

Attitude Tracking Study, 1976-1984.

TABLE 4.7

INTENTION TO JOIN THE MILITARY

BY FATHER'S HIGHEST LEVEL OF EDUCATION

Father's Level of Education

Less than Greater than
High School High School High School
Will Join 31.3% 24.8% 18.7%

Source: Derived from data provided by the Youth

Attitude Tracking Study, 1976-1984.

Table 4.8 presents the percentage of respondents with a

general interest in joining the military (for all services)

by Navy district. The table also presents the propensity

for each Navy recruiting district by different year group.

Years are grouped together to provide a sufficient sample

24
I

TABLE 4.8
INTENTION TO JOIN THE MILITARY
FOR NAVY RECRUITING DISTRICTS BY YEAR
(Number Will Join)
(Percent Will Join)

Years
1976-1978 1979-1980 1981-1984 ALL YEARS
Navy Recruiting
Districts
Albany 175 154 111 440
26.7% 21.4% 22.8% 23.6%
Boston 346 348 156 850
30.1% 24.5% 22.9% 26.2%
Buffalo 170 256 102 528
24.2% 24.1% 27.5% 24.7%
New York 197 206 102 491
21.2% 16.7% 21.7% 19.2%
Philadelphia 239 264 64 567
24.1% 20.4% 17.1% 21.3%
New Jersey 157 211 70 438
23.5% 20.7% 21.8% 21.8%
Montgomery 129 104 100 333
34.1% 27.3% 36.0% 32.1%
Columbia 168 211 36 469
36.6% 31.1% 30.5% 32.8%
Jacksonville 254 247
35.6% 28.8% 105 606
26.9% 30.9%
Atlanta 168 224 73 465
30.8% 29.1% 24.6% 26.8%
Nashville 118 107
29.6% 25.2% 109 334
28.6% 27.7%
Raleigh 194 216 187 597
33.6% 32.0% 36.1% 33.7%

25
 TABLE 4.8 (CONTINUED)

Years

1976-1978 1979-1980 1980-1984 ALL YEARS

Navy Recruiting
Districts

Richmond 66 103 49 218

24.0% 27.6% 23.7% 25.5%

Miami 121 119 81 321

32.5% 31.2% 27.4% 30.6%

Harrisburg 171 116 90 377

27.3% 21.2% 22.2% 23.9%

District of 300 361 107 768

Columbia 27.6% 26.3% 29.2% 27.1%

Cleveland 116 117 119 352

22.5% 18.8% 31.5% 23.2%

Columbus 186 167 126 479

26.7% 20.1% 26.1% 23.8%

Pittsburgh 183 203 72 458

25.2% 20.1% 27.0% 22.9%

Michigan 188 183 93 464

24.7%. 20.7% 19.2% 21.8%

Glenview 227 310 100 637

19.6% 20.3% 21.9% 20.3%

St Louis 144 146 101 391

25.7% 22.8% 27.2% 24.9%

Louisville 202 299 119 620

23.1% 25.2% 27.5% 24.9%

Kansas City 11162 51 324

20.5% 20.1% 22.8% 20.6%

Minneapolis 211 211 101 523

22.6% 18.0% 20.4% 20.1%

Omaha 286 283 107 676

23.0% 18.0% 22.2% 20.5%

26
 7Uvvw.
PW Fy 1W uJ-W.-

TABLE 4.8 (CONTINUED)

Years
1976-1978 1979-1980 1981-1984 ALL YEARS
Navy Recruiting
District
Indianapolis 90 75 87 252
23.4%.1 18.6% 24.8%0- 22.1%
Milwaukee 187 257 40 484
19.0% 19.5% 17.2% 19.1%
Denver 146 135 51 332
23.6% 21.4% 19.9% 22.1%
Albuquerque 159 234 114 507
34.0% 27.2% 27.5% 29.1%
*Dallas 115 92 139 346
27.0% 21.2% 20.4% 22.4%
Houston 100 134 72 306
29.2% 23.5% 26.5% 28.0%
Little Rock 149 161 89 399
25.8% 23.5% 22.3% 24.0%
New Orleans 315 441 116 872
28.0% 26.3%, 26.2% 26.9%
Scdn Antonio 95 113 67 275
32.6% 31.5% 27.3% 30.7%
Memphis 135 151 136 422
34.9%9 31.1% 34.5% 33.3%
*Los Angeles 147 149 73 369
22.3% 18.9% 19.9% 20.4%
Portland 148 149 186 483
23.2% 19.7% 26.7% 23.1%
4 San Francisco 209 209 132 550
21.8% 18.5% 20.9% 20.2%
Seattle 136 162 91 389
23.9% 19.7% 19.2% 20.99%
4

27

4
 I- .
-V 1

TABLE 4.8 (CONTINUED

Years

1976-1978 1979-1980 1981-1984 ALL YEARS

Navy Recruiting
Districts

San Diego 130 97 92 319

28.8% 20.7% 21.1% 23.5%
Missing or 61 137 15 213
Unidentified
Total 7,149 8,024 4,067 19,204
25.9% 22.8% 24.7% 24.2%

Source: Based on data provided by the Youth Attitude

Tracking Study, 1976-1984.

size to determine propensity at the district level. An

analysis of variance indicates that there is a significant

difference between year groups. There appears to be no

consistent pattern of change for all Navy recruiting

districts. For some districts there has been a slight

continuous increase, for others a slight continuous decrease

or a slight up and down shift in propensity.

The last column of Table 4.8 provides the propensity for

all nine years for each Navy recruiting district. Across

all Navy recruiting districts the propensity to join the

military ranges from a low of 19.2 percent in New York City,

New York to a high of 33.7 percent in Raleigh, North

Carolina.

28

I
 After missing cases or unidentified cases were removed

there were 79,354 respondents in the sample that could be

matched with a specific Navy recruiting district. Of the

total respondents, 19,240 indicated a positive intention to

join the military. The national average of propensity to

join the military was calculated to be 24.2 percent.

Table 4.9 contains the propensity index of general

intention to join the military by Navy recruiting areas and

by districts. The national average of intention to join the

military is 24.2 percent. The second colurn provides the

total sample size for the indicated area or district.

Column three, provides the propensity to join the military

for each area and district. The last column gives the

calculated propensity index for each area and district. The

ratio is calculated by dividing the percent intend to join

by the percent national average. The ratio is multiplied by

100 tc obtain the index for each area and district.

The highest propensity index for Navy recruiting areas

is in the southeast (126.5) and the southwest (110.3). As

stated previously the same results were found by Citizen.

* However, considerable variation in propensity is observed

within Navy recruiting areas. For example, in the New

England area positive propensity averages 94.628, but varies

from a low of 79.339, 16 percent below the average, to a

high of 108.264, 14 percent above the average.

29

U,X i-W
 TABLE 4.9

PROPENSITY INDEX OF GENERAL INTENTION TO JOIN THE MILITARY

BY NAVY RECRUITING AREA AND DISTRICT

Percent Percent
Navy Recruiting Intend National Index
District/Area N to Join Average Ratio (Ratio X 100)

Area 1
New England 14,472 0.229 0.242 0.946 94.62

Albany 1,862 0.236 0.242 0.975 97.52

Boston 3,250 0.262 0.242 1.083 108.26
Buffalo 2,137 0.247 0.242 1.021 102.07
New York 2,559 0.192 0.242 0.793 79.34
Philadelphia 2,658 0.213 0.242 0.880 88.01
New Jersey 2,006 0.218 0.242 0.901 90.08

Area 3 126.45
Southeast 10,910 0.306 0.242 1.264

Montgomery 1,037 0.321 0.242 1.326 132.65

Columbia 1,419 0.328 0.242 1.355 135.54
Jacksonville 1,963 0.309 0.242 1.277 127.69
Atlanta 1,613 0.288 0.242 1.190 119.01
Nashville 1,204 0.277 0.242 1.145 114.46
Raleigh 1,770 0.337 ?.:!2 1 193 139.26
Richmond 855 0.255 0.242 1.054 105.37
Miami 1,049 0.106 0.242 1.264 126.45

Area 4
Northeast 12,068 0.306 0.242 0.992 99.174

Harrisburg 1,580 0.239 0.242 0.988 98.76

Wash. D.C. 2,829 0.271 0.242 1.119 111.98
Cleveland 1,516 0.232 0.242 0.959 95.89
Columbus 2,010 0.238 0.242 0.983 98.35
Pittsburgh 2,004 0.229 0.242 0.946 94.63
Michigan 2,129 0.218 0.242 0.901 90.08

30

SO%*I
WWWOMMJNUVW77.'iJF7

TABLE 4.9 (CONTINUED)

Percent Percent
Navy Recruiting Intend National Index
District/Area N to Join Average Ratio (Ratio X 100)

Area 5
Midwest 18,348 0.213 0.242 0.880 88.02

Glenview 3,143 0.203 0.242 0.839 83.88

St. Louis 1,572 0.249 0.242 1.029 102.89
Louisville 2,492 0.249 0.242 1.029 102.89
Kansas City 1,573 0.206 0.242 0.851 85.12
Minneapolis 2,602 0.201 0.242 0.831 83.06
Omaha 3,295 0.205 0.242 0.847 84.71
Indianapolis 1,139 0.221 0.242 0.913 91.32
Milwaukee 2,532 0.191 0.242 0.789 78.93

Area 7
Southwest 12,948 0.267 0.242 1.103 110.331

Denver 1,504 0.221 0.242 0.913 91.32

Albuquerque 1,742 0.291 0.242 1.202 120.25
Dallas 1,543 0.224 0.242 0.926 92.56
Houston 1,091 0.280 0.242 1.157 115.70
Little Rock 1,663 0.240 0.242 0.992 99.16
New Orleans 3,244 0.269 0.242 1.112 111.16
San Antonio 895 0.307 0.242 1.269 126.86 p
Memphis 1,266 0.333 0.242 1.376 137.60

Area 8
West 9,843 0.214 0.242 0.884 88.43

Los Angeles 1,813 0.204 0.242 0.843 84.29

Portland 2,090 0.231 0.242 0.955 95.46
San Francis. 2,718 C 202 0.242 0.835 83.47
Seattle 1,864 C 209 0.242 0.854 86.36
San Diego 1,358 0.235 0.242 0.971 97.11

Source: Developed from data from the Youth Attitude

Tracking Study, 1976-1984.

31

I
 A t-test indicates that there is significant evidence of

a difference in the percent national average and the area

percent intend to join for the New England, Southeast,

Midwest, Southwest and West recruiting areas. The Northeast

recruiting area indicated no significant difference between

area percent intend to join and the percent national

average.

As a comparison the same procedure was applied on Army

recruiting battalions (N=56) and brigades (N=5). Table 4.10

provides the results for Army battalions and brigades. This

table indicates that the propensity index is the highest in

the southeast (2nd Brigade) and the southwest (5th Brigade).

Citizen found the highest propensity for the Army to be in

the southeast and the northeast (1st brigade). Substantial

variation across recruiting battalions within brigades is

also observed for the Army.

B. SPECIFIC INTENTION TO JOIN THE NAVY

In the tables above the calculated propensity refers to

all services. Table 4.11 provides the propensity index of

intention to join the Navy by Navy recruiting areas and

districts. The total sample size was 79,242, with 10,952

indicating a positive propensity to join the Navy. The

national average Navy propensity was calculated to be 13.8

percent. Once again note the Navy propensity index is

highest in the southeast and the southwest recruiting areas.

32

N
 II

TABLE 4.10
PROPENSITY INDEX OF GENERAL INTENTION TO JOIN THE MILITARY
BY ARMY RECRUITING BATTALION AND BRIGADE

Army Recruiting Percent Percent

Brigades/ Intend National Index
Battalions N to Join Average Ratio (Ratio X 100)
1st Brigade 20,734 0.235 0.242 0.971 97.11

Albany 679 0.236 0.242 0.975 97.52

Baltimore 2,938 0.271 0.242 1.120 111.98
Boston 2,191 0.233 0.242 0.963 96.28
Brunswick 987 0.317 0.242 1.310 130.99
Harrisburg 1,609 0.239 0.242 0.988 98.76
New Haven 1,427 0.245 0.242 1.012 101.24
Long Island 1,812 0.199 0.242 0.822 82.23
Newburgh 1,508 0.180 0.242 0.744 74.38
Ft Monmouth 1,492 0.235 0.242 0.971 97 11
Philadalphia 2,568 0.212 0.242 0.876 87.60
Pittsburgh 2,268 0.227 0.242 0.938 93.80
Syracuse 2,003 0.251 0.242 1.037 103.72
2nd Brigade '3,446 0.295 0.242 1.219 121.90
Atlanta 1,664 0.288 0.242 1.190 119.01
Beckley 896 0.278 0.242 1.149 114.88
Charlotte 965 0.333 0.242 1.376 137.60
Columbia 1,505 0.333 0.242 1.376 137.60
Jacksonville 1,985 0.310 0.242 1.281 128.10
Louisville 1,658 0.233 0.242 0.963 96.28
Miami 1,159 0.305 0.242 1.260 126.03
Montgomery 1,081 0.321 0.242 1.326 132.64
Nashville 1,253 0.277 0.242 1.145 114.46
Raleigh 828 0.338 0.242 1.397 139.67
Richmond 881 0.256 0.242 1.058 105.79
4th Brigade 18,318 0.211 0.242 0.872 87.19

Chicago 2,062 0.203 0.242 0.839 83.88

Cincinnati 1,020 0.242 0.242 1.000 100.00
Cleveland 1,558 0.232 0.242 0.959 95.87
Columbus 1,079 0.235 0.242 0.971 97.11
Des Moines 2,164 0.212 0.242 0.876 87.60
Detroit 1,172 0.204 0.242 0.843 84.30
Indianapolis 1,028 0.227 0.242 0.938 93.80
Lansing 960 0.227 0.242 0.938 93.80
Milwaukee 2,596 0.191 0.242 0.789 78.93
Minneapolis 2,282 0.192 0.242 0.793 79.34
Omaha 1,606 0.208 0.242 0.860 85.95
Peoria 1,409 0.206 0.242 0.851 85.12

33

22
 TABLE 4.10 (CONTINUED)

Army Recruiting Percent Percent

Brigades/ Intend National Index
Battalions N to Join Average Ratio (Ratio X 100)
5th Brigade 16,139 0.259 0.242 1.070 107.02
Albuquerque 1,554 0.298 0.242 1.231 123.14
Dallas 1,612 0.224 0.242 0.926 92.56
Denver 1,547 0.221 0.242 0.913 91.32
Houston 1,117 0.280 0.242 1.157 115.70
Jackson 1,312 0.333 0.242 1.376 137.60
Kansas City 1,615 0.206 0.242 0.851 85.12
Little Rock 1,881 0.267 0.242 1.103 110.33
New Orleans 2,213 0.262 0.242 1.083 108.26
Oklahoma City 1.194 0.238 0.242 0.983 98.35
San Antonio 916 0.307 0.242 1.269 126.86
St. Louis 1,624 0.248 0.242 1.025 102.48
6th Brigade 9,846 0.214 0.242 0.884 88.43
San Francis. 1,477 0.185 0.242 0.764 76.45
Los Angeles 1,590 0.205 0.242 0.847 84.71
Phoenix 735 0.255 0.242 1.054 105.37
Portland 1,107 0.243 0.242 1.004 100.41
Sacramento 1,340 0.222 0.242 0.917 91.74
Salt Lake Cy 1,403 0.230 0.242 0.950 95.04
Santa Ana 1,001 0.207 0.242 0.855 85.54
Seattle 1,562 0.196 0.242 0.810 80.99

Source: Developed from data from the Youth Attitude

Tracking Study, 1976-1984.

34
 TABLE 4.11
PROPENSITY INDEX OF INTENTION TO JOIN THE NAVY
BY NAVY RECRUITING AREA AND DISTRICT

Percent Percent
Navy Recruiting Intend National
District/Area Index
N to Join Average Ratio (Ratio X 100)
Area 1
New England 14,444 0.125 0.138 0.904 90.42
Albany 1,804 0.137 0.138 0.993
Boston 99.28
3,247 0.152 0.138 1.101 110.14
Buffalo 2,131 0.128 0.138 0.928 92.75
New York 2,584 0.089 0.138 0.645
Philadelphia 64.49
2,644 0.125 0.138 0.906 90.58
New Jersey 2,034 0.116 0.138 0.841 84.06
Area 3
Southeast 10,969 0.172 0.138 1.246 124.64
Montgomery 1,052 0.167 0.138 1.210
Columbia 121.01
1,438 0.204 0.138 1.478 147.83
Jacksonville 1,952 0.160 0.138 1.159
Atlanta 115.94
1,621 0.185 0.138 1.341 134.06
Nashville 1,218 0.129 0.138 0.935 93.48
Raleigh 1,779 0.187 0.138 1.355 135.51
Richmond 866 0.142 0.138 1.029 102.89
Miami 1,043 0.184 0.138 1.333 133.33
Area 4
Northeast 11,941 0.134 0.138 0.971 97.101
Harrisburg 1,547 0.184 0.138 1.333 133.33
Wash. D.C. 2,744 0.145 0.138 1.051
* 105.07
Cleveland 1,503 0.121 0.138 0.877
* Columbus 87.68
2,035 0.140 0.138 1.014 101.45
Pittsburgh 2,011 0.120 0.138 0.870
Michigan 87.96
2,101 0.124 0.138 0.899 89.86

35

4i
 TABLE 4.11 (CONTINUED)

Percent Percent
Navy Recruiting Intend National Index
District/Area N to Join Average Ratio (Ratio X 100)

Area 5
Midwest 18,298 0.122 0.138 0.884 88.41

Glenview 3,085 0.120 0.138 0.870 86.96

St. Louis 1,572 0.141 0.138 1.022 102.17
Louisville 2,497 0.136 0.138 0.986 98.55
Kansas City 1,573 0.123 0.138 0.891 89.13
Minneapolis 2,599 0.123 0.138 0.891 89.13
Omaha 3,317 0.119 0.138 0.862 86.23
Indianapolis 1,141 0.122 0.138 0.884 88.41
Milwaukee 2,514 0.103 0.138 0.746 74.64

Area 7
Southwest 12,990 0.155 0.138 1.123 112.32

Denver 1,509 0.133 0.138 0.964 96.38

Albuquerque 1,742 0.164 0.138 1.188 118.84
Dallas 1,567 0.124 0.138 0.899 89.86
Houston 1,083 0.172 0.138 1.246 124.64
Little Rock 1,681 0.136 0.138 0.986 98.55
New Orleans 3,245 0.175 0.138 1.268 126.81
San Antonio 895 0.145 0.138 1.051 105.07
Memphis 1,268 0.174 0.138 1.261 126.09

Area 8
West 9,837 0.132 0.138 0.957 95.65

Los Angeles 1,825 0.136 0.138 0.986 98.55

Portland 2,097 0.142 0.138 1.029 102.89
San Francis. 2,694 0.120 0.138 0.870 86.96
Seattle 1,857 0.125 0.138 0.906 90.58
San Diego 1,364 0.143 0.138 1.036 103.62

Source: Developed from data from the Youth Attitude

TrackinQ Study, 1976-1984.

36

p
 A t-test indicates that there is significant evidence of

a difference in the percent national average and the area

percent intend to join for the New England, Southeast,

Midwest and Southwest recruiting areas. The Northeast and

West recruiting areas indicated no significant difference

between area percent intend to join and the percent national

average.

Table 4.12 provides a comparison of the propensity index

for general intention versus Navy intention by Navy

recruiting areas and districts. The simple correlation

coefficient between the two propensity indexes is +.835,

which is statistically significant. This positive

correlation indicates that general military propensity is

highly correlated with Navy propensity in most districts and

could serve as a proxy for it.

Table 4.13 provides a comparison between the coeffi-

cients of variation for general propensity index and for the
Navy propensity index. Compared to general propensity there

is a larger variance for Navy propensity in the New England,

Southeast, West and Northeast areas. The variance for the
Midwest and Southeast areas is lower, however, the

difference is only 1 percent and 1.3 percent lower,

respectively. Though there appears to be greater variation
between Navy recruiting districts and between Navy

recruiting districts within areas for Navy propensity, the

variation for general propensity between Navy recruting

37
 TABLE 4.12

COMPARISON OF MILITARY PROPENSITY INDEX

AND NAVY PROPENSITY INDEX
BY NAVY RECRUITING AREA AND DISTRICT

Propensity Index Propensity Index

General Intention Navy Intention

Navy Recruiting
Area/District

Area 1
New England 94.6 90.4

Albany 97.5 99.3

Boston 108.3 110.1
0Buffalo 102.1 92.8
New York 79.3 64.5
Philadelphia 88.0 90.6
New Jersey 90.1 84.1
Area 3
Southeast 126.4 124.6

Montcomery 132.6 121.0

Columbia 135.5 147.8
Jacksonville 127.7 115.9
Atlanta 119.0 134.1
Nashville 114.5 93.5
Raleigh 139.3 135.5
Richmond 105.4 102.9
Miami 126.4 133.3

Area 4
Northeast 99.2 97.1

Harrisburg 98.8 133.3

District of Columbia 111.9 105.1
Cleveland 95.9 87.7
Columbus 98.3 101.4
Pittsburgh 94.6 86.9
Michigan 90.1 89.9

38

%0ZZ
 TABLE 4.12 (CONTINUED)

Propensify Index Propensity Inaex

General Intention Navy Intention

Navy Recruiting
Area/District

Area 5
Midwest 88.0 88.4

Glenview 83.9 86.9

102.2
St. Louis 102.9
102.9 98.6
Louisville 89.1
Kansas City 85.1
83.1 89.1
Minneapolis 86.2
Omaha 84.7
91.3 88.4
Indianapolis 74.6
78.9
Milwaukee
Area 7
Southeast 110.3 112.3

Denver 91.3 96.4

118.8
Albuquerque 120.2
92.6 89.9
Dallas 124.6
Houston 115.7
99.2 98.6
Little Rock 126.8
New Orleans 111.2
126.9 105.1
San Antonio 126.1
Memphis 137.6

Area 8
West 88.4 95.7

Los Angeles 84.3 98.6

Portland 95.5 102.9
San Francisco 83.5 86.9
90.6
Seattle 86.4
San Diego 97.1 103.6

Correlation Coefficient +.835

Source: Developed from data provided by the Youth

Attitude Trackinct Study, 1976-1984.

39
 TABLE 4.13

COEFFICIENT OF VARIATION OF PROPENSITY INDEX

General Navy
Propensity Index Propensity Index

Coefficient of Variance
Between Navy Recruiting
Districts by Areas
New England 11.1% 17.1%
Southeast 9.1% 14.8%
Northeast 7.5% 17.5%
Midwest 10.3% 9.3%
Southwest 14.8% 13.5%
West 7.2% 7.7%

Coefficient of Variance
Between Navy Recruiting
Areas 14.7% 13.9%

Coefficient of Variance
Between Navy Recruiting
Districts 16.9% 18.0%

Source: Developed from data provided from the Youth

Attitude Tracking Study, 1976-1984.

areas is lower than that between Navy recruiting areas.

There is a greater variation between districts than there is

between areas for both general and Navy propensity.

C. SPECIFIC INTENTION TO JOIN THE ARMY

Table 4.14 presents the propensity index of intention to

join the Army by Army recruiting battalions and brigades.

The sample size was 79,240, with 10,325 indicating positive

intention to join the Army. The national average positive

Army propensity was 13.0 percent. As indicated in Table

40
 TABLE 4.14
PROPENSITY INDEX OF INTENTION TO
JOIN THE ARMY
BY ARMY PFCRUITTNG BATTALICN AND
DRIGADE
Army Recruiting Percent
Brigades/ Percent
Intend National
Battalions Index
N to Join Average Ratio (Ratio X 100)
Ist Brigade 20,734 0.116 0.13 0.892 89.23
Albany 646 0.149
Baltimore 2,754 0.136 0.13 1.146 114.62
0.13 1.046 104.62
Boston 2,124 0.107 0.13 0.823 82.31
Brunswick 964 0.168
Harr-isburg 0.13 1.292 129.23
1,536 0.133 0.13 1.023
New Haven 1,320 102.31
0.120 0.13 0.923
Long Island 1,773 92.31
0.074 0.13 0.569
Newburgh 1,448 56.92
0.088 0.13 0.677
Ft Monmouth 1,456 0.111 67.69
0.13 0.854 85 38
Philadalphia 2,586 0.091 0.13 0.700 70.00
Pittsburgh 2,197 0.123 0.13 0.946 94.62
Syracuse 1,930 0.131 0.13 1.008 100.77
2nd Brigade 13,446 0.179 0.13 1.377 137.69
Atlanta 1,622 0.186 0.13 1.431 143.08
Beckley 871 0.186 0.13 1.431 143.08
Charlotte 937 0.203
Columbia 0.13 1.562 156.15
1,470 0.235 0.13 1.808
Jacksonville 1,898 180.77
0.171 0.13 1.315
Louisville 1,619 131.54
0.145 0.13 1.115
Miami 1,096 111.54
0.163 0.13 1.254
Montgomery 1,050 125.38
0.165 0.13 1.269
Nashville 1,214 0.152 126.92
0.13 1.169 116.92
Raleigh 805 0.234 0.13 1.800 180.00
Richmond 861 0.154 0.13 1.185 118.46
4th Brigade 18,318 0.116 0.13 0.892 89.23
Chicago
Cincinnati 1,977 0.102 0.13
978 0.140 0.785 78.46
Cleveland 0.13 1.077 107.69
1,510 0.116 0.13 0.892
Columbus 1,053 89.23
0.134 0.13 1.031 103.08
Des Moines 2,133 0.120
Detroit 1,140 0.13 0.923 92.31
0.096 0.13 0.738 73.85 A
Indianapolis 987 0.127 0.13 0.977
Lansing 97.69
892 0.120 0.13 0.923
Milwaukee 2,517 92.31
0.099 0.13 0.762
Minneapolis 2,217 0.102 76.15
0.13 0.785 78.46
Omaha 1,581 0.135 0.13 1.038 103.85
Peoria 1,333 0.133 0.13 1.023 102.31

41
 TABLE 4.14 (CONTINUED)

Army Recruiting Percent Percent

Brigades/ intend National Index
Battalions N to Join Average Ratio (Ratio X 100)

5th Brigade 16,139 0.146 0.13 1.123 112.31

Albuquerque 1,525 0.153 0.13 1.177 117.69

Dallas 1,566 0.130 0.13 1.000 100.00
Denver 1,520 0.091 0.13 0.700 70.00
Houston 1,079 0.157 0.13 1.208 120.77
Jackson 1,269 0.206 0.13 1.585 158.46
Kansas City 1,575 0.108 0.13 0.83.L 83.08
Little Rock 1,836 0.181 0.13 1.392 139 23
New Orleans 2,124 0.159 0.13 1.223 122.31
Oklahoma City 1,172 0.135 0.13 1.038 103.85
San Antonio 896 0.158 0.13 1.215 121.54
St. Louis 1,577 0.139 0.13 1.069 106.92
6th Brigade 9,846 0.095 0.13 0.731 73.08

San Francis. 1,411 0.076 0.13 0.585 58.46

Los Angeles 1,513 0.106 0.13 0.815 81.54
Phoenix 711 0.110 0.13 0.846 84.62
Portland 1,087 0.099 0.13 0.762 76.15
Sacramento 1,285 0.107 0.13 0.823 82.31
Salt Lake Cy 1,363 0.101 0.13 0.777 77.69
Santa Ana 970 0.082 0.13 0.631 63.08
Seattle 1,502 0.081 0.13 0.623 62.31

Source: Developed from data from the Youth Attitude

Tracking Study, 1976-1984

4.14 the highest Army positive propensity index is in the

southeast and the southwest.

Table 4.15 compares the propensity index for general

military intention versus Army intention by Army recruiting
4 battalions and brigades. The simple correlation coefficient

between the two propensity indexes is +.885, which is

statistically significant. This correlation indicates that

42

II
I

TABLE 4.15

COMPARISON OF MILITARY PROPENSITY INDEX AND ARMY PROPENSITY

INDEX BY ARMY RECRUITING BATTALION AND BRIGADE

Propensity Index Propensity Index

Gi-neral Intention Army Intention
Army Recruiting
Battalions/Brigades

1st Brigade 97.1 89.2

Albany 97.5 114.6

Baltimore 111.9 104.6
Boston 96.3 82.3
Brunswirk 130.9 129.2
Harrisburg 98.8 102.3
New Haven 101.2 92.3
Long Island 82.2 56.9
Newburgh 74.4 67.7
Ft. Monmouth 97.1 "5.4
Philadelphia 87.6 . u.0
Pittsburgh 93.8 94.6
Syracuse 103.7 100.8
2nd Brigade 121.9 137.7

Atlanta 119.0 143.0

Beckley 114.9 143.0
Charlotte 137.6 156.1
Columbia 137.6 180.8
Jacksonville 128.1 131.5
Louisville 96.3 111.5
Miami 126.0 125.4
Montgomery 132.6 126.9
Nashville 114.5 116.9
Raleigh 139.7 180.0
Richmond 105.8 118.5

43
 TABLE 4.15 (CONTINUED)

Propensity Index Propensity Index

General Intention Army Intention
I.

Army Recruiting
Battalion/Brigade

4th Brigade 87.1 89.2

Chicago 83.9 78.5

Cincinnati 100.0 107.7
Cleveland 95.9 89.2
Columbus 97.1 103.1
Des Moines 87.6 92.3
Detroit 84.3 73.9
Indianapolis 93.8 97.7
Lansing 93.8 92.3
Milwaukee 78.9 76.2
Minneapolis 79.3 78.5
Omaha 85.9 103.9
Peoria 85.1 102.3

5th Brigade 107.0 112.3

Albuquerque 123.1 117.7

Dallas 92.6 100.0
Denver 91.3 70.0
Houston 115.7 120.8
Jackson 137.6 158.5
Kansas City 85.1 83.1
Little Rock 110.3 139.2
New Orleans 108.3 122.3
Oklahoma City 98.4 103.9
San Antonio 126.9 121.5
St. Louis 102.5 106.9

6th Brigade 88.4 73.1

San Francisco 76.5 58.5

Los Angeles 84.7 81.5
Phoenix 10i.4 76.2
Portland 100.4 76.2
Sacramento 91.7 82.3
Salt Lake City 95.0 77.7
Santa Ana 85.5 63.1
Seattle 81.0 62.3

Correlation Coefficient +.885

Source: Developed from data from the Youth Attitude

Tracking Study, 1976-1984.

44
 general military propensity is highly correlated with Army

propensity in most battalions and could serve as a proxy for

it.

D. INTENTION TO JOIN THE MILITARY BY STATE

Table 4.16 provides the propensity index of intention to

join the military by states. This does not provide any

information that could be used with the current Army or Navy

recruiting command structure, however, it may be of interest

in the future if boundaries are changed.

E. LOGIT/PROBIT ANALYSIS
A probit regression analysis was conducted on the

independent variables, excluding the Navy recruiting

districts dummy variables. Due to the large sample size the

probit analysis could not be completed for all years

combined. Thus, separate probit regressions were estimated

on three separate year groups. The groupings were not

determined by the number of years, but by the sample size

within each group. However, it might be expected that

differences in propensity would emerge if trends are present

in the data. Table 4.17 provides the sample size of each

group.

Table 4.17 provides the regression coefficients for the

0 three year groups. The signs of the coefficient models are

the same for the three year groups. As age increases, the

lower the probability of the intention to join the military.

45

*1"
 TABLE 4.16

PROPENSITY INDEX OF GENERAL INTENTIONS

TO JOIN THE MILITARY BY STATE

Percent Percent
Intend National Index
State N to Join Average Ratio (Ratio X 100)

1,109 0.296 0.242 1.223 122.31

Alabama
623 0.252 0.242 1.041 104.12
Arizona
863 0.277 0.242 1.145 114.46
Arkansas
5,269 0.207 0.242 0.855 85.54
California
1,340 0.213 0.242 0.880 88.02
Colorado
1,083 0.234 0.242 0.961 96.69
Connecticut
347 0.291 0.242 1.202 120.25
Delaware
Wash. D.C. 343 0.329 0.242 1.360 135.95
Florida 3,025 0.332 0.242 1.372 137.19
2,319 0.297 0.242 1.227 122.73
Georgia
418 0.244 0.242 1.008 100.83
Idaho
3,955 0.208 0.242 0.860 85.95
Illinois
1,660 0.228 0.242 0.942 94.22
Indiana
Iowa 2,180 0.207 0.242 0.855 85.54
828 0.199 0.242 0.822 82.23
Kansas
1,862 0.251 0.242 1.037 103.72
Kentucky
3,320 0.268 0.242 1.107 110.74
Louisiana
622 0.333 0.242 1.376 137.60
Maine
Maryland 2,047 0.257 0.242 1.062 106.19
Massachusett 2,106 0.238 0.242 0.983 98.34
Michigan 2,340 0.215 0.242 0.888 88.84
Minnesota 2,001 0.194 0.242 0.802 80.16
Mississippi 773 0.344 0.242 1.421 142.15
Missouri 1,707 0.240 0.242 0.992 99.17
Montana 363 0.262 0.242 1.083 108.26
660 0.180 0.242 0.744 74.38
Nebraska
131 0.237 0.242 0.979 97.93
Nevada
New Hampshire 338 0.287 0.242 1.186 118.60
2,637 0.218 0.242 0.901 90.08
New Jersey
759 0.329 0.242 1.360 135.95
New Mexico
5,159 0.216 0.242 0.893 89.26
New York
1,680 0.335 0.242 1.384 138.43
No. Carolina
291 0.230 0.242 0.950 95.04
North Dakota

46
 TABLE 4.16 (CONTINUED)

Percent Percent
Intend National Index
State N to Join Average Ratio (Ratio X 100)

Ohio 3,316 0.231 0.242 0.955 95.45

Oklahoma 956 0.240 0.242 0.992 99.17
Oregon 1,014 0.243 0.242 1.004 100.41
Pennsylvania 5,161 0.225 0.242 0.930 92.98
Rhode Island 313 0.256 0.242 1.058 105.79
So. Carolina 1,089 0.345 0.242 1.426 142.56
South Dakota 259 0.251 0.242 1.037 103.72
Tennessee 1,017 0.283 0.242 1.169 116.94
Texas 4,348 0.261 0.242 1.079 107.85
Utah 605 0.215 0.242 0.888 88.84
Vermont 173 0.306 0.242 1.264 126.44
Virginia 1,569 0.252 0.242 1.041 104.13
Washington 1,552 0.193 0.242 0.798 79.75
W. Virginia 909 0.272 0.242 1.124 112.40
Wisconsin 2,697 0.194 0.242 0.802 80.17
Wyoming 116 0.302 0.242 1.248 124.79
Missing or
Unidentified 96

Total 79,354

Source: Developed from data from the Youth Attitude

Tracking Study, 1976-1984.

47 '
P AJP.P.
 TABLE 4.17

PROBIT COEFFICIENTS BY YEAR

(t statistic)

YEARS

1976-1978 1979-1980 1981-1984

Variable

Intercept 8.474 7.899 7.509

(39.2)*** (43.5)*** (35.l)***
Age -.197 -.158 -.134
(-18.2)*** (-17.4)*** (-12.7)***
Race .404 .399 .477
(10.8)*** (12.5)*** (12.9)***
Cursch .154 .151 .208
2.9)*** ( 3.1)*** (3.6)***
Gradhs .212 .167 .252
(3.8)*** (3.4)*** (3.6)***
Avggrd -.129 -.217 -.272
(-6.7)*** (-12.9)*** (-12.8)***
Math -.112 -.070 -.064
(-9.)***(-8.)***(-5.8)***
Father -.154 -.125 -.099
(-9.)***(-8.)***(-5.8)**
N 27,619 25,150 26,585

** Statistically significant at the .01 level.

Source: Developed from data extracted from the Youth

Attitude Tracking~ Study, 1976-1984.

48
This is consistent with the simple cross-tabulation

displayed in Table 4.1. If a respondent is black he will be

more likely to have a positive propensity to join the

military. Table 4.3 indicated that blacks were almost twice

as likely to have positive intentions to join the military.

If the respondent is not currently in school the results

indicate an increase in his military propensity. This shows

that for individuals of a given age, currently in school has

a positive coefficient. This clearly indicates that age and

currently in school are highly correlated. A non-high

school graduate is likely to have a higher propensity to

join the services, which is consistent with Table 4.2. As

the average grade in high school increases, the positive

enlistment intentions also decreases. This can be confirmed

with the results from Table 4.5. Similarly, as the number

of math courses taken in high school increases the

propensity for enlistment intentions decreases. This

confirms the bivariate results in Table 4.6. Finally, as

the respondent's father's education level increases, the

results indicate there will be a decrease in the positive

propensity.

The probit equations were highly significant for all

year groups. Appendix C contains the contingency tables,

comparing actual and predicted intention to join the

military by year groupings. The probit model correctly

49
classifies approximately two-thirds of all cases in each

year group.

The same model was then run with the additon of Navy
recruiting districts as independent variables. Again the

sample size would not allow the model to be run as a simple

probit regression. The model was run separately for the six

Navy recruiting areas. Table 4.18 contains the results of

the probit estimations for the six Navy recruiting areas.

The signs for the independent variables are consistent with

the results contained in Table 4.17. However, the magnitude

of the coefficients vary across areas. There is a 1.2

variation between the smallest and largest area intercept.

There is only a slight ariation between the smallest and

largest intercept in the number of math courses completed

and father's highest level of education, .07 and .04

respectively, but a large variation in race of .357.

Currrent high school status , average grades in high school

and high school graduate status varied between .I and .2

across areas. The base case Navy recruiting district for

each area is as follows:

1. Area 1 Philadelphia Recruiting District

2. Area 3 Jacksonville Recruiting District

3. Area 4 District of Columbia Recruiting District

4. Area 6 Omaha Recruiting District

5. Area 7 New Orleans Recruiting District

6. Area 8 San Francisco Recruiting District.

50

N 5N
 TABLE 4.18

PROBIT COEFFICIENTS BY NAVY RECRUITING AREA

(t statistic)

Navy Recruiting Areas

1 3 4 5 8

Variable

Intercept 7.437 7.758 7.816 8.654 7.577 7.977

(26.7) (25.8) (26.2) (34.6) (26.2) (22.5)

Age -. 149 -. 151 -. 144 -. 196 -. 147 -. 164

(-10.9) (10.1) (-9.8) (-15.7) (-10.2) (-9.5)

Race .406 .506 .324 .366 .464 .149

(7.6) (11.2) ( 6.1) ( 6.5) (10.6) ( 1.7)

Cursch .223 .241 .052 .105 .154 .055

(2.7) (3.5) ( 0.7) ( 1.7) ( 2.2) (0.6)

Gradhs .265 .275 .183 .114 .185 .155

( 3.1) (3.7) ( 2.2) ( 1.7) ( 2.5) ( 1.5)
** ** ** * *** *

Avggrd -. 162 -. 170 -. 190 -. 266 -. 186 -. 259

(-6.3) (-5.8) (-6.8) (-11.7) (-6.9) (-7.7)

Math -. 125 -. 055 -.110 -. 059 -. 101 -. 095

(-8.1) (2.9) (-6.4) (-4.1) (-5.6) (-1.4)

Father -. 108 -. 141 -. 137 -. 122 -. 101 -. 117

(-4.9) (-5.9) (-5.7) (-6.3) (-4.5) (-4.3)

51

I
 TABLE 4.18 (CONTINUED)

Navy Recruiting Areas

13 4 5 7 8

Variable

Albany .163
(2.8)

Boston .316
(6.2)

Buffalo .229
(4.1)

New York -.089

* (-1.6)

New Jersey .031

(0.5)
Montgomery -.143
(-1.9)

Columbia -.119
(-1.8)

Atlanta -.149
(-2.5)

Nashville -. 150
(-2.2)

Raleigh -.101

Richmond -.237
(-3.1)

52
 WI

TABLE 4.18 (CONTINUED)

Navy Recruiting Areas

1 3 4 5 7 8
Variable

Miami .068
(0.9)
Harrisburg -. 038
(0.6)
Cleveland -. 121
(-2.0)

Columbus -. 106
(-1.9)

Pittsburgh -.084
(-1.6)

Michigan -. 129
(-2.3)

Glenview -. 077
(-1.6)

St Louis .102
( 1.7)

Louisville .134
2.6)

Kansas City .011

(0.2)
Minneapolis .027
(0.5)

53

0
 TABLE 4.18 (CONTINUED)

Navy Recruiting Areas

1 3 4 5 7 8

Variable

Indianapolis .117
(1.7)

Milwaukee -. 022
(-0.4)

Denver .070
(1.1)
Albuquerque .150
2.5)

Dallas -. 017
(-0.3)

Houston .172
(2.7)

Little Rock .022

(0.4)

San Antonio .146

( 1.8)

Memphis .157
(2.6)

Los Angeles .047

(0.7)

Portland .129
(2.1)

54
 TABLE 4.18 (CONTINUED)
Navy Recruiting Areas

1 3 4 5 7 8
Variable

Seattle .117
(1.1)
San Diego .081
(2.0)

N 14,472 10,910 12,068 18,348 12,948 9,843

* Significant at the .10 level.

** Significant at the .05 level.
* Significant at the .01 level.

Source: Developed from data extracted from the Youth

Attitude Tracking Study, 1976-1984.

Appendix D contains the contingency tables, comparing

actual and predicted intention to join the military for

individual recruiting areas. Again, approximately two-

thirds of the cases are correctly classified.

The entire sample was finally run using SAS logit

procedures. This provided sufficient time and disk space to

include all the independent variables and the full sample of

to be run.

Table 4.19 provides the logit regression coefficients.

The independent variables (other than recruiting district)

all have the same signs as in Tables 4.17 and 4.18. The

-55
 signs for the Navy recruiting districts did not necessarily

remain the same as Table 4.18.

56

0i .< ,<<.'. - :.U...,, ''-.:.,: <..'F . ' ;'b..;q.:

';<'<'< -,<
 TABLE 4.19

LOGIT REGRESSION COEFFICIENTS

(Pooled Sample)

INTERCEPT 5.168 ***

AGE -. 282 ***
RACE .681 ***
CURSCH .246 ***
GRADHS .342 ***
AVGGRD -. 352 ***
MATH -. 150 ***
FATHER -. 210 ***
ALBANY .063 GLENVIEW -. 307 ***
BOSTON .319 *** ST LOUIS -. 001
BUFFALO .180 * LOUISVILLE .033
NEW YORK -. 399 *** KANSAS CITY -. 175
PHILADELPHIA -. 241 ** MINNEAPOLIS -. 121
NEW JERSEY -. 178 * OMAHA -. 161 *
GONMGOMERY .190 INDIANAPOLIS .016
COLUMBIA .258 ** MILWAUKEE -. 207 **
JACKSONVILLE .422 *** DENVER .063
ATLANTA .233 ** ALBUQUERQUE .191 *
NASHVILLE .175 DALLAS -. 064
RALEIGH .269 ** HOUSTON .243 **
RICHMOND .028 LITTLE ROCK -. 024
MIAMI .562 *** NEW ORLEANS -. 044
HARRISBURG .133 SAN ANTONIO .190
WASHINGTON DC .170 MEMPHIS .217 *
CLEVELAND -. 005 LOS ANGELES -. 168
COLUMBUS .024 PORTLAND .019
PITTSBURGH .045 SAN FRANCISCO -. 248
MICHIGAN -. 037 SEATTLE .012

N = 45,682

Base Case: Navy Recruiting District - San Diego

• Significant at the .10 level.

•* Significant at the .05 level.
•** Significant at the .01 level.

Source: Developed from data extracted from the Youth

Attitude Tracking Study, 1976-1984.

57
 V. CONCLUSIONS AND RECOMMENDATIONS

This study has only opened the door for research on YATS

intentions data as an indicator for predicting new contracts

in Navy recruiting districts. The goal of this study was

to investigate possibile alternative methods which might be

used to forecast positive intention propensity to enlist in

the Navy. Emphasis was placed on determing the positive

intention propensity for recruiting areas and districts.

The measure of intention is not intended to replace other

statistical indicators currently used to estimate new

contracts or recruiter goal allowcation.

A. CONCLUSIONS

1. YATS survey responses can be used to forecast the

underlying local market propensity.

2. General military intention or intention to join the

Navy could be used in the predicting model because
military propensity and Navy propensity are so highly
correlated.
3. There have been changes in propensity during the last
three to four years of YATS; so it might be wise to
restrict data analysis to relatively current data.

B. RECOMMENDATIONS

1. Analysis should be done using the YATS to predict

Armed Services Vocational Aptitude Battery (ASVAB)
testing rates and/or actual assessions by Navy
recruiting districts.

58
 2. The logit and probit models should be used to predit
intention for 1985-1987 and then compared to actual
intentions from the YATS to validate the intention
prediction capability.

3. The YATS index should be added to the Navy recruiting

goal allocation model selected to predict new
contracts for past years to determine if the model
with intention included is a better predictor of new
contract, than the model currently used.

.559

!V
59 -
 APPENDIX A

NAVY RECRUITING AREAS AND DISTRICTS

Area District

Area 1 Albany
New England Boston
Buffalo
New York
Philadelphia
New Jersey

Area 3 Montgomery
Southeast Columbia
Jacksonville
Atlanta
Nashville
Raleigh
Richmond
Miami

Area 4 Harrisburg
Northeast Washington DC
Cleveland
Columbus
Pittsburgh
Michigan

Area 5 Glenview
Midwest St. Louis
Louisville
Kansas City
Minneapolis
Omaha
Indianapolis
Milwaukee
Area 7 Denver
Southwest Albuquerque
Dallas
Houston
Little Rock
New Orleans
San Antonio
Memphis

60
60

q |
 ql~

Area
District
Area West
Los Angeles
Portland
San Francisco
Seattle
San Diego

61
 APPENDIX B

NAVY RECRUITING DISTRICT SAMPLE SIZE

Area Navy Recruiting District Sample Size

New England Albany 1937

Boston 3347
Buffalo 2208
New York 2666
Philadelphia 2753
New Jersey 2080

4 Southeast Montgomery 1081

Columbia 1470
Jacksonville 2041
Atlanta 1664
Nashville 1253
Raleigh 1828
Richmond 880
Miami 1106
Northeast Harrisburg 1609
Washington DC 2939
Cleveland 1558
Columbus 2099
Pittsburgh 2063
Michigan 2211

Midwest Glenview 3235

St. Louis 1622
4 Louisville 2554
Kansas City 1617
Minneapolis 2678
Omaha 3374
Indianapolis 1185
Milwaukee 2596

62

V
Area Navy Recruiting District Sample Size
Southwest Denver 1547
Albuquerque 1775
Dallas 1612
Houston 1117
Little Rock 1725
New Orleans 3342
San Antonio 916
Memphis 1312
West Los Angeles 1912
Portland 2134
San Francisco 2819
Seattle 1940
San Diego 1414

Missing or
Unidentified 794

Total 82,013

'..It

63
-
 APPENDIX C

CONTINGENCY TABLES, COMPARISON OF ACTUAL AND PREDICTED

INTENTION TO JOIN THE MILITARY BY YEAR GROUP

2976-1978

Predicted Intention

Will Not Will

Join Join
Will Not
Join 59.8% 14.4%
Observed
Intention
Will 17.9% 8.0%
Join

N = 27,619
Proportion of the total correctly classified is 67.8%.

1979-1980
Predicted Intention

Will Not Will

Join Join
Will Not
Join 53.4% 23.2%
Observed
Intention
Will 12.9% 10.4%
Join

N = 25,150
Proportion of the total correctly classified is 63.8%.

64

64r
 1981-1984

Predicted INtention

Will Not Will

Join Join
Will Not
Join 63.2% 13.4%
Observed
Intention
Will 17.0%
6.4%
Join

N = 26,585
Proportion of the total correctly classified is 69.6%.

Source: Derived from data from the Youth Attitude

Tracking Study, 1976-1984.

"65

4.,

'p

5'
 APPENIX D

CONTINGENCY TABLES, COMPARISON OF ACTUAL AND PREDICTED

INTENTION TO JOIN THE MILITARY BY AREA

AREA 1 (NEW ENGLAND)

Predicted Intention

Will Not Will

Join Join

Will Not
Join 59.0% 18.1%
Observed
Intention
Will 15.0% 7.9%
Join

N = 14,472
Proportion of the total correctly classified is 66.9%

AREA 3 (SOUTHEAST)

Predicted Intention
Will Not Will
Join Join
Will Not
Join 55.7% 13.7%
Observed
Intention
Will 20.6% 10.0%
Join

N = 10,910
Proportion of the total correctly classified is 65.7%

66

Re P .,w
 AREA 4 (NORTHEAST)

Predicted Intention

Will Not Will

Join Join

Will Not
Join 58.9% 17.0%
Observed
Intention
Will 15.5% 8.6%
Join

N = 12,068
Proportion of the total correctly classified is 67.5%.

AREA 5 (MIDWEST)

Predicted Intention

Will Not Will

Join Join

Will Not
Join 55.7% 13.7%
Observed
Intention
Will 20.6% 10.0%
Join

N = 10,910
Proportion of the total correctly classified is 65.7%.

67
 AREA 7 (SOUTHWEST)

Predicted Intention

Will !ot Will

Join Join

Will Not
Join 58.4% 14.9%
Observed
Intention
Will 18.4% 8.4%
Join

N = 12,948
Proportion of the total correctly classified is 66.8%.

AREA 8 (WEST)

Predicted Intention

Will Not Will

Join Join

Will Not
Join 61.3% 17.3%
Observed
Intention
Will 15.0% 6.5%
Join

N = 9,843
Proportion of the total correctly classified is 67.8%.
Source: Derived from data from the Youth Attitude
Tracking Study, 1976-1984.

68

L N A--LAXI
 APPENDIX E

COMPARISON OF ACTUAL AND PREDICTED

INTENTION TO JOIN THE MILITARY

Predicted Intention

Will Not Will

Join Join

Will Not
Join 76.7% 01.4%
Observed
Intention
Will 20.5% 01.4%
Join

N = 45,682
Proportion of the total correctly classified is 78.1%

Source: Derived from data from the Youth Attitude

Tracking Study, 1976-1984.

-,. 69
'4

0 J-4""". .L''". -'. '-% ,' : : - £ ? -' I : i i; . --

 LIST OF REFERENCES

1. Borack, J.I., "A Technique for Profiling the Composition

of the 'High Quality' Military Market," Defense
Analysis, V. 2, No. 3, 1986.

2. Hosek, J.R., Fernandez, R.L., and Grissmer, D.W.,

"Enlisted Strength in the '80s--A Mid-Term Reassess-
ment," Defense Management Journal, p. 4, Second Quarter
91985.

3. Ivancevich, J.M. and Matteson, M.T., Organizational

Behavior and Management, Business Publication, Inc.,
1987.

4. Quester, G.H., "Women in Combat," International

Security, V. 1, pp. 80-91, Spring 1977.

5. Navy Recruiting Command, FY88 Allocation Models 1988.

6. Telephone conversation between Carl E. Kannepel, Navy

Recruiting Command, Arlington, Virginia, March 14, 1988
and the author.

7. Marine Corps Recruiting Command, FY88 Fairshare Quota

Distribution.

8. Army Recruiting Command, EPM Enlistment Proiection

Model.
9. Rand Corportation, N-2292-MIL, Relationship of
Enlistment Intention and Market Survey Information to
Enlistment in Active Duty Military Service, by B.R.
Orvis, and M.T. Gahart, June 1985.

10. Rand Corporation, N-1954-MRAL, Forecasting Enlistment

Actions from Intention Information: Validity and
Improvement, B.R. Orvis, p. v, December 1982.

11. Citizen, G.D., A New Device for Estimating Local Area

Enlistment Market Potential, Master's Thesis, Naval
Postgraduate School, Monterey, California, June 1985.

70

Q7
 INITIAL DISTRIBUTION LIST

No. Copies
1. Defense Technical Information Center 2
Cameron Station
Alexandria, Virginia 22304-6145

2. Library, Code 0142 2

Naval Postgraduate School
Monterey, California 93943-5002

3. Dr. Jules I. Borack 1

Code 63 Bldg 320C
Navy Personnel Research and Development Center
San Diego, California 92152-6800

4. Mr. Carl E. Kannapel 3

Navy Recruiting Command
0 4015 Wilson Boulevard
Arlington, Virginia 22203-1991

5. Dr. Stephen L. Mehay, Code 54Mp 10

Naval Postgraduate School
Monterey, California 93943-5000

6. Dr. Linda Gorman, Code 54Gr

Naval Postgraduate School
Monterey, California 93943-5000

7. Officer in Charge 2
Personnel Support Detachment
Naval Air Facility--Detroit
Mount Clemens, Michigan 48045-5057

71

1z