https://orcid.org/0000-0002-2871-3358
![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
What does it mean to be a veteran, and how does serving in the armed forces condition how veterans view their sense of identity? In a national survey sample with an oversample of veterans, I find that veterans have a stronger sense of identity as a veteran, measured in terms of self-identification as both a veteran and a feeling of closeness to the veteran group, compared to civilians. I also find that, among military veterans, combat experience and valuing time in the military leads to higher veteran identity. Moreover, I find that even some non-military members report a greater sense of identity with veterans than others. I compare the effect of this “veteran” identity to that of partisan identity and find that, for most veterans, there is a greater sense of attachment to the veteran identity than to their partisan identity. Finally, I find that veteran identity has an important, independent influence on veterans’ and civilians’ views on military spending. These findings suggest that there is a veteran identity that military members and civilians attach themselves to that is stronger than partisanship for some individuals and is associated with certain policy positions.
Acknowledgements
The author would like to thank Conor Dowling, Robert Brown, Roger Henley, Kevin Samuels, the editors, and the anonymous reviewers for their valuable comments.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Human subjects approval was granted by the Institutional Review Board at the University of Mississippi.
2 Qualtrics recruits participants through a number of ways including, but not limited to: Website intercept recruitment, Member referrals, Targeted email lists, Gaming Sites, Customer loyalty web portals, Permission-based networks, and Social Media. Participants are verified through a double-opt-in process and agree to take part in surveys for an incentive. Participants are also subject to other quality control measures such as LinkedIn matching, phone calls made to the participant’s place of business, and other third-party verification methods. Using profile information provided by the participants, Qualtrics sends them specific email invitations to applicable surveys. If they elect to participate, panelists are informed and agree at the beginning that they will only receive compensation upon completion of the survey.
4 The sample is fairly representative of the percentage of each branch’s makeup (Army [36.6%], Navy [24.8%], Air Force [24.3%], and Marines [14.2%]), according to data from the Council of Foreign Relations (https://www.cfr.org/article/demographics-us-military).
5 The 2012 ANES sample summary statistics can be found in Appendix Table A2. The 2012 ANES was the most recent National Elections Survey that included questions regarding the feeling thermometers for the Democratic Party, Republican Party, and the United States Military as well as questions concerning cutting military spending.
6 Respondents who identified as Democrats (Republicans) including those who subsequently identified themselves as “strong” or “weak” are coded as 1 for democrat (republican). Partisan “leaners” (those respondents who initially identified as Independent but subsequently stated they felt closer to the Democratic or Republican Party) are included with partisans (Petrocik Citation2009; Baum and Groeling Citation2009; Bullock Citation2011; Druckman, Fein, and Leeper Citation2012; Levendusky Citation2012).
7 Additional analysis conducted with ordered logit instead of OLS for the Closeness to Veterans dependent variable appears in Appendix Table A4. That analysis yields similar statistical and substantive results to the OLS models presented in first model of Appendix Table A3.
8 For example, Republican respondents from the 2012 ANES do not have as warm of feelings toward the military as compared to those in the Qualtrics study. This could be an artifact of the time at which the studies took place and who the commander in chief was at the time, President Obama in 2012 and President Trump in 2018.
9 Factor analysis of the five items that comprise the Combat Events variable results in an eigenvalue of 3.55 (and load on only one factor). Additionally, a principle components/correlation test found that on the singular factor the variables explain 71% of the variance. These analyses suggest that the items collectively measure a single latent trait (Combat Events). Moreover, since these are somewhat subjective measures (e.g., I cannot differentiate whether witnessing gunfire exchange with the enemy was “worse” for some respondents more than others), I do not wish to claim that any one of these events necessarily carries more explanatory power than any others.
10 See Table A4 for an analysis using logit analysis instead of OLS.
11 Although feeling thermometer ratings may not provide as direct a test for how closely a respondent identifies with a given group compared to other measures of closeness, the feeling thermometers are the only measures of identity I have that are similar for both identity with veterans and identity with the political parties. Further, as Greene (Citation1999, Citation2004) shows, feeling thermometers can provide an acceptable measure of a respondent’s social (partisan) identity. Moreover, my utilization of feeling thermometers regarding veteran and party identification measures is similar to what Mason and Wronski (Citation2018) use in order to test differences in attitudes toward in-party and out-party feelings. I do believe that this test, which is used in prior studies, does provide a new and informative approach to understanding respondent attitudes toward these different groups.
12 If respondents indicated they identified with either the Republican or Democratic Party and subsequently selected that they are “strong” Republicans/Democrats, they are in the “Strong” category. If they identified as “not very strong” Republicans/Democrats or as an Independent but felt closer to the Republican/Democratic Party, then they are placed in the “Weak” category (i.e., partisan “leaners” are in the “Weak” group). Appendix Tables A6 and A7 contain all partisan categories including Independents.
13 To test for statistical significance, I estimated a series of OLS regressions separately for each partisan group in which the difference score (i.e., feeling thermometer military minus feeling thermometer in-party) was regressed on an indicator for whether the respondent was a military member (1) or civilian (0).
14 In a supplemental analysis of the 2011 CCES, which contains variables asking about military service, familial relationship to service members, and attitudes about cutting defense spending, there are only two differences in the results where opposition to cutting defense spending is the dependent variable. In the 2018 Qualtrics data, education and white were not significantly different from zero, but in the 2011 CCES, they are significant and negative (more support to cutting defense spending). Being a military member (current or former) leads to more opposition to cutting defense spending than being a family member of a service member, but this is only different in the 2011 CCES. The largest predictor of opposition to cutting defense spending while controlling for military service or military relationship is among older respondents. This is found in both data sources.
15 Specifically, in Table A9, we observe similar decreases in mean support for decreasing military spending among veteran Republicans and Democrats as compared to their civilian counterparts. Respondents who identified as Democrat were more likely to support decreasing military spending, however, having a veteran association similarly decreased this effect.