Anchoring Sweden’s Downsized Military – People’s Attitude to, Knowledge About, and Trust in Our Military Defense

The hypothesis was that post-Cold War downsizing of the military defense is associated with low levels of public anchoring, and that better anchoring is associated with a perception of a better performance by the Swedish Armed Forces (SAF) and with a SAF connection and military experience. An additional research question concerned the level of concordance between the military and political elites and the population on these issues. A questionnaire was sent to 3,000 randomly selected residents. The results showed that post-Cold War downsizing was only partly associated with low levels of anchoring. Better anchoring was associated with a perception of a better performance by the SAF and with SAF connection and experience. There was significant concordance between the elites and the public. The study shows that anchoring is best supported by increased communication of the results and benefits of the military defense to all subgroups of the population.


Introduction
Since the end of the Cold War, many countries have dramatically downsized their armed forces. This means that the distance between the military and large groups of the population has increased. Knowledge, interest, and trust in the military may be negatively affected when a much smaller proportion of the cohort does military service and when people see less military activity. Accordingly, an important question is how to maintain healthy civil-military relations with a downsized force.
For a long time, the relationship between the military and civil society has attracted scholarly attention. Recent reviews address the multidimensionality of the field, which is described as a sprawling, complex mix of various themes and analytical approaches (Pion-Berlin & Dudley 2020), a field divided across multiple dimensions (Brooks 2019), or a field with many models, but none appropriate for all nations (Rukavishnikov & Pugh 2018).
One challenge that research has addressed is how to keep the power of the military under democratic control, and there has been debate about the appropriate level of independence or integration of the military in relation to civilian society (Burk 2002;Feaver 1999;Avant 1998;Rahbek-Clemmensen et al. 2012). A central question is whether, or to what degree, the military elite will follow orders from the political elite (Burk 2002). Another central question concerns the civil-military gap that exists, according to most scholars. This gap deals with the level of military influence on policy, differences between military and civilian populations, the degree to which the military is representative of society, and the amount of friction between civilians and the military in day-to-day interactions (Szayna et al. 2007;Feaver & Kohn 2001).
The present civil-military issue in Sweden is not the risk of a coup or that the military becomes too strong and independent. The problem is how to maintain a healthy societal anchoring in a situation of downsizing and distancing from the defense. This study will address this specific civil-military problem.
Most recent research shows that the armed forces in many countries enjoy a high level of public confidence and support (Hill, Wong, & Gerras 2013;Hines et al. 2015;Tiargan-Orr & Eran-Jona 2016). Western Europeans SCANDINAVIAN JOURNAL OF MILITARY STUDIES confidence, trust, respect, and pride. A third is public interest, knowledge, and understanding. In our previous study, the term societal anchoring was defined as follows: The general public (1) is familiar with the SAF and their mission, (2) trusts the SAF to act in accordance with political wishes, and (3) is confident that the SAF reflect the values of the civilian society (Wallenius et al. 2019). Accordingly, we do not equate societal anchoring with a generally positive view of the defense. Neither do we define any political position in itself as more or less anchored. It is, however, desirable that people's opinions are based on knowledge and involvement, and that people trust government agencies.
Hence, it is in the interest of a nation to maintain good civil-military relations or, to put it differently,to have a satisfying societal anchoring of the national defense. While we have some knowledge about the perceptions of the elites, we lack corresponding knowledge about the population in general. Hypotheses and questions are: Hypothesis 1: In the general population, the post-Cold War downsizing and distancing of the military defense is associated with low levels of anchoring (public interest, knowledge, and trust). This hypothesis is based on the research reported above, indicating that especially interest and knowledge are limited, although few question the need for a military defense. Hypothesis 2: On the individual level, better anchoring is associated with a perception of better performance by the SAF. This hypothesis is based on the research reported above, showing that trust in public institutions and in the military defense is associated with perceived performance. Hypothesis 3: On the individual level, better anchoring is associated with an SAF connection and with military experience. This is a logical conclusion, though it indicates a problem, since a historically small proportion of the population presently has such connections and experience. Accordingly, the character of this relationship is important. Question 1: What level of concordance is found between the military and political elite, on the one hand, and the general public, on the other, when it comes to issues related to anchoring? Schiff's (2009) concordance theory emphasizes the importance of conformity between the population, the political elite, and the military. Since this has never been studied in the Swedish context, it is formulated as an open question rather than a hypothesis.

Method Participants
The participants in the study consisted of a random selection of persons aged 18 to 84 years registered in Sweden. The sample size was 3,000. The service was provided by Skop-Research AB, which in turn purchased a sample from the Swedish population register (the National Register of Personal Addresses). The survey was conducted as a postal survey with two reminders. Each participant in the survey received an introductory letter, which explained the purpose of the survey and provided a personal password for an online questionnaire containing the same questions as the postal questionnaire. The participant could then choose between the paper and online forms.
Of the 3,000 people surveyed, 56 were unable to answer the questionnaire due to living abroad, illness, and the like. Of the remaining 2,944 people, 1,170 chose to answer the questionnaire, giving a response rate of 40%. In eight cases, only the background questions and/or single questions were answered. These eight respondents were excluded from the analyses. This means that 1,162 (39%) questionnaires have been analyzed. Of these, 1,038 (89%) responded by post and 124 (11%) online.
The gender distribution was largely even, and the age distribution is shown in Table 1 below. Of the respondents, 18% reported having completed primary and lower secondary school, 34% had completed upper-secondary school, and 47% had a post-secondary degree as the highest level of education. On the issue of main employment, 50% stated that they were employed, 40% that they had retired, and others that they were entrepreneurs, students, or job seekers. The proportion of married couples or cohabitants was 71%, while 29% were single. The proportion of respondents with children under the age of 18 was 23%. In addition, 81% of the respondents had been born in Sweden and had Swedish-born parents, 7% had been born in Sweden and had at least one foreign-born parent, and 12% had been born abroad.

Representativeness
Respondents with a post-secondary degree were overrepresented in the sample (48% of the sample versus 36% of the population). Respondents with a secondary education were underrepresented (34% of the sample versus 45% of the population). Respondents with a primary education were somewhat under-represented (17% of the sample versus 19% of the population) (Statistics Sweden 2015b). Foreign-born respondents were underrepresented (12% of the sample versus 19% of the population) (Statistics Sweden 2015d). The gender distribution was representative of the general population with 49.9% female and 50.1% male respondents (Statistics Sweden 2015c). Finally, respondents aged 18-45 years were slightly underrepresented, while respondents aged 55-75 years were overrepresented (see Table 1). Compared to the general population, the respondents were thus somewhat older, more educated, and included slightly more Swedish-born individuals.
To correct for any selective non-responding, post-stratification was conducted for region, gender, and age. A weighting variable was constructed. However, the weight did not affect the mean values of the indices. Since no real differences were detected, respondents were assumed to be representative of the population, and unweighted data could thus be used.

Measures
The survey consisted of 95 questions/statements and one open question. The majority were answered on a five-point scale (agree entirely -do not agree at all), and most were derived from the results of a previous interview study within the project (Wallenius et al. 2019). The factors which the interviewees in these studies (the executive and political elites) considered important for anchoring, and which were relevant to the public, were included. This means that problems that were considered defense-internal or internal political issues were left out of this survey.
The survey contained the following blocks: • Assessment and concern for societal risks and threats (disasters, military attacks, terrorism) • Interest in the SAF and the defense policy • Knowledge about the SAF and the defense policy • View of the SAF's focus, tasks, and cost • Perception of the SAF's focus, tasks, and cost, and the defense policy's focus and long-term perspective • Opinion of the SAF's international missions • Trust in the SAF and the defense policy • Statements of ways to strengthen the trust In addition, the questionnaire included background questions (gender, age, education, employment, marital status, place of birth, size of the place of residence, and personal connection to the SAF) and one open question, where the respondents could state if they saw any other or more significant threats to the country than a military attack.
The data collection was conducted between 4 June and 4 September 2015.

Converting Questions/Statements into Indices
Indices were formed through factor analyses of the various blocks of questions. Questions with a low response rate (more than 30% missing) were removed from this part of the analysis. The indices were largely identical to the survey blocks. In some cases, however, the blocks were divided into several indices (see Table 2).

Analysis
The study is cross-sectional, and the purpose is to show statistical correlation between different factors as well as to highlight differences between subgroups. The indices were created through block-to-factor analysis (direct axis factoring with direct oblimin rotation). The indices' reliability was measured using Cronbach's alpha. The relationship between the indices was studied through correlation analysis (Pearson bivariate), and the differences between subgroups were mapped via t-tests (independent samples) or variance analysis (one-way ANOVA).
Effect size was calculated to get a better indication of the relative size of the mean value differences. The measure of one square (η 2 ) can vary between 0 and 1. A value of .01 is considered a small effect, a value of .06 a moderate effect, and a value of .14 indicates a great effect (Pallant 2013). How do you perceive the risk that a major crisis/disaster will occur in Sweden in the next five years?
Worries about different threats to society How big is your concern for a major crisis/disaster to occur in Sweden? Note: Most items were answered on a five-point scale (agree entirely -do not agree at all). Items formulated as questions were answered on different five-point scales (non-existent -very large; not at all -very much).

Results
The results section starts with a section comparing different subgroups, before identifying the statistical relationships between the different indices.

Index Means and Standard Deviations
Even if there is reason to be careful when comparing the different scales, we can conclude that the majority of respondents consider both the SAF's value-based work and openness as important for increased trust (see Table 3).
Generally, respondents have a relatively positive attitude to the defense, military service, the defense industry, and international missions. However, with regard to international missions, there is more variation in the responses. The mean for experienced knowledge about the defense policy and SAF is remarkably low. It should also be noted that the public clearly trusts the SAF more than the defense policy, even though this trust is generally limited. Simply put, people have "partial" trust in the SAF and between "to a low degree" and "partial" trust in the defense policy.

Open Question About Perceived Threats to Society
The survey also gave the respondents a chance to provide examples of threats to the Swedish society which they considered bigger than military threats. 568 respondents answered this open question. Terrorism was the most common answer (45%). Other commonly occurring responses were environmental and climate changes (16%), socioeconomic differences and threats to the welfare society (13%), infrastructure threats, including cyber threats (10%), crime (10%), immigration (8%), xenophobia (6.5%), integration and conflicts between (ethnic) groups (5.8%), and economic threats (5.5%).

The Relationship Between the Indices
The relationship between the indices is shown in Table 4. It is also illustrated in Figure 1, where medium or strong correlations are shown. The limit of .350 was chosen for this illustration. Some of the relationships shown in the correlation analysis were expected. Examples include the positive correlations between perceived risk and perceived concern, between interest in and knowledge about the SAF/the defense policy, and between trust in the SAF and trust in the defense policy. Also expected was the correlation between a positive attitude to the SAF and a positive attitude to defense alliances and international missions, as well as to interest in the SAF and the defense policy.
The correlation between a positive attitude to the SAF and worrying about/risk of different threats to society is notable. We do not know the causal direction, but a reasonable hypothesis is that some degree of Note: Large Index 7, 11: Scale 1-5, where 1 = nonexistent, 3 = moderate, and 5 = very. Index 9: Scale 1-5, where 1 = not at all, 3 = quite interested, and 5 = very interested. Index 1, 3: Scale 1-5, where 1 = not important, 3 = quite important, and 5 = very important. Others: Scale 1-5, where 1 = do not agree at all, 3 = agree in part, and 5 = agree entirely. worrying/risk makes a military defense more relevant. Also worth noting is the fact that trust in the SAF is strongly connected to a positive attitude to international missions, whereas its connection to other indices is weak or insignificant.

Age
Age has little or no impact on the different indices. The index with the largest correlation with age is trust in the defense policy (r = .183; p < .001). Younger respondents showed slightly less trust in the defense policy than older respondents, but the correlation is weak.

Gender
For several indices, there is a significant difference between women and men, albeit with a small to moderate effect size (see Table 5). Men claim to be more interested in the SAF and the defense policy, as well as having greater knowledge of these areas. Men also have a somewhat more positive attitude to the defense, military service, and the defense industry, international missions and defense alliances. Women consider the SAF's value-based work as more important for their trust in the SAF. Part of the gender gap remains, even if you check for military service (see further below on military service).

Education
The questionnaire defines four levels of education: primary and lower-secondary school (n = 175), uppersecondary school n = 381), post-secondary education of less than three years (n = 209), and post-secondary education of three years or more (n = 319). The higher the education, the greater the respondent's interest in and knowledge about the SAF and the defense policy (see Table 6). Higher education also implies that the SAF's transparency is an important factor with regard to trust. For other indices, the post hoc test (Scheffé) shows the following significant differences between the groups (p ≤ .05). Concern about international missions is significantly greater among respondents with an upper-secondary education than among respondents with a post-secondary education of three years or more. Trust in the defense policy is significantly greater among respondents with a primary and lowersecondary education than among respondents with an upper-secondary education. The SAF's core values are considered significantly more important by those with a post-secondary education of three years or more than by those with a primary or secondary education.

Place of residence
The questionnaire included items about the size of the place of residence. The division was: • City with more than 200,000 inhabitants (n = 279) • City with 81-200,000 inhabitants (n = 199) • City with 21-80,000 inhabitants (n = 248) • Rural area, less than 20,000 inhabitants (n = 408) The larger the hometown, the greater the respondent's interest in the SAF and the defense policy and knowledge about the latter (see Table 7). For other indices, the post hoc test (Scheffé) shows the SAF's core  values to be more important in cities with more than 200,000 inhabitants than in cities with 21,000-80,000 inhabitants.

The SAF's presence and visibility at place of residence
Respondents living in a town that is home to a military unit or military school (n = 235) show a greater interest in the SAF and the defense policy and have more knowledge about both than respondents living in a town that is not home to a military unit or military school (n = 704) (see Table 8). The effect size is small for all indices.
Respondents who reported that the SAF are visible in the neighborhood (n = 156) are more interested in the SAF and the defense policy than respondents who do not experience the SAF as visible (n = 917). The former also report greater knowledge about the SAF (see Table 9). However, the power size is small.

Foreign descent
A total of 11% (n = 133) of the respondents are foreign-born, and 18% (n = 210) have at least one foreign-born parent. Of those who were born abroad, almost all (n = 124) also have a parent who was born abroad, and therefore, the group of foreign-born respondents is almost entirely included within the group of respondents with one foreign-born parent.
Foreign-born respondents report lower perceived risk of threats to society and less worrying than those born in Sweden (see Table 10). They perceive themselves as having less knowledge about the defense policy. They are more positive toward defense alliances and have greater trust in the defense policy. The effect size for foreign descent or not is small for trust in the defense policy and very small for the remaining indices.
Respondents with at least one foreign-born parent have a slightly lower mean score for worrying than others (see Table 11). They also have a more positive view of defense alliances and largely report that they trust the defense policy. However, the variable foreign-born parent has little effect on these indices.

Personal connection with the SAF
The questionnaire included questions about whether the respondent or a close relative of the respondent was connected to the SAF, and whether the respondent had been an employee of the SAF or active in the Home Guard. 15% (n = 154) had such a link, and 85% (n = 887) did not.   Respondents with a personal connection with the SAF experience greater interest and knowledge about the SAF and the defense policy (see Table 12). They have a more positive attitude to the defense, military service, the defense industry, and defense alliances, but less trust in the defense policy. Respondents who are closely linked to the SAF also experience threats to society as greater risks and consider core values less important factors when it comes to increasing trust in the SAF. However, the effect size is small on these indices.

Completed military service
Respondents who have completed military service (n = 350) differ from those who have not (n = 691). They experience greater interest in and more knowledge about the defense policy and the SAF (see Table 13). The power size is moderate. Those who have completed military service also have a more positive attitude to the defense, military service, and the defense industry, as well as to international missions, though the size of the impact is small.
As 338 of the 351 respondents who have completed military service are men (98.6%), comparing men who have completed military service (n = 338) with men who have not (n = 187) can isolate the effect of military service. The men who have completed military service experience greater interest in the SAF and the defense policy compared to those who have not completed military service (see Table 14). They also believe to have more knowledge about the defense policy and the SAF, and they have a more positive attitude to the defense, military service, the defense industry, and international missions. However, the effect size of conscription experience is moderate.
The results also show that completed military service explains a part of, but not the whole difference between men and women. Even the men who have not completed military service have a higher mean than women with regard to interest in and knowledge about the SAF and the defense policy.

Level of anchoring
The respondents have a relatively positive attitude to the defense, military service, the defense industry, and international missions. People generally feel the need for a military defense, but their attitudes to military  alliances are more varied. Many people have little knowledge about both the defense policy and, above all, the SAF. Their trust in the SAF is somewhat greater than in the defense policy, though the latter can be regarded as remarkably low.

Relations between the indices
Some of the correlations were expected, such as the correlations between worrying and risk, between a positive attitude to the SAF, to military alliances, and to international missions, between trust in the SAF and trust in the defense policy, and finally, between interest and knowledge. Worth noting here is that trust in the SAF is strongly connected to a positive attitude to international missions. Table 15 provides a summary of the significant differences between subgroups (male, foreign-born respondents, respondents with a longer education, respondents living in a large city, respondents with a personal connection to the SAF, and respondents who have completed military service). It is clear that interest and knowledge are greatest among male respondents, respondents with a longer education, respondents with a personal connection to the SAF, and respondents who have completed military service. Male respondents, respondents with a personal connection to the SAF, and respondents with military experience are also somewhat more positive to the defense, military service, the defense industry, military alliances, and international missions. These groups are, of course, overlapping, as mainly men have military experience. Foreign-born respondents show greater trust in the SAF and the defense policy than Swedish-born respondents. Core values are considered more important among female respondents and respondents with no connection to the SAF.

Differences between demographic categories
At the same time, it should be noted that although the differences between groups are statistically significant, they are generally small to moderate. They should consequently be understood more as trends than as all-or-nothing relationships.

Discussion
According to the study, hypothesis 1 -that post-Cold War downsizing is associated with low levels of anchoring -is only partly true. The defense's current level of societal anchoring is acceptable, judging from the present results. Confirmation of this hypothesis is restricted by a lack of established measures and of reference values from previous eras, though. What could threaten this anchoring is lack of knowledge and the distance between the SAF and large groups of the population.  Hypothesis 2 -that better anchoring is associated with a perception of better performance by the SAF -was confirmed, as trust in the SAF is correlated with a positive attitude to international missions. The Afghanistan mission (ISAF) was for a long time the most visible of the SAF's activities and affected the public view of the SAF's performance, even though Sweden has now entered a phase with more focus on national defense.
Hypothesis 3 -that better anchoring is associated with a personal SAF connection and experience -was also confirmed, as those who experience relatively more interest and knowledge are men who have done military service, have a personal connection with the SAF and a higher level of education.
Turning to question 1 -concordance between the elite and the general public -we can conclude that there are significant similarities between the results of the present study and the results of a previous study focusing on how the military and civilian elites perceived societal anchoring (Wallenius et al. 2019). The public is relatively positive to the idea of a military defense and considers it necessary. The elites feel that the public wants a military defense, but that the opinion varies somewhat, depending on the current threats to the nation. Furthermore, the public reports a relatively positive attitude to international missions. Among the elites, international missions enjoy great support, but the fact that their purposes are perceived as unclear has impaired the anchoring.
The public perceives their interest to be low to moderate. The elite groups hold a similar view: that public interest in the SAF and the defense policy is generally low. The public reports having limited knowledge about the SAF and the defense policy, and the elites find that the public knowledge is inadequate and outdated, albeit varied. The public reports partial trust in the SAF, and the elites find that the general public has inadequate, albeit varied views of the capacities of the SAF, in particular regarding national defense.
Lack of trust in the defense policy is assumed to reflect a general lack of trust vis-à-vis politicians and political parties. However, we have not measured public trust in other policy areas and therefore do not know whether public trust in defense politicians differs from public trust in politicians in general.
One conclusion of the study is that politicians, as well as the SAF, must be better at communicating the results achieved in their activities, including international missions. What seems to be critical is not whether we should have a military defense or not, but that this defense delivers reasonable and clear results. Note: + means that the subgroup has significantly higher value, -implies significantly lower + respectively -means that p ≤ .05; ++ respectively --means that p < .001.
As for target groups in the population, we find that there are differences between the subgroups. However, it should be noted that these differences are consistently moderate and do not give reason to exclude any group from future communication efforts. The results suggest that it is the content of the information provided that is important, not to whom it is addressed. If you were to formulate some general advice on the basis of this study, it would be to communicate the results and benefits of the military defense and to do so widely.
This study has confirmed the conclusions of the studies referred to in the introduction: that downsizing in itself does not necessarily have a significant negative effect on the societal anchoring of the armed forces. The important gap seems to be not between civilians and the military, but between those with and those without knowledge of the latter. Most people simply do not have up-to-date knowledge of the present status of the armed forces and how it relates to the defense policy. Only a minority within the political and military elites do.
When it comes to conceptual issues, this study has deliberately avoided outcome variables like support for higher defense spending. This would be to regard certain political views as more correct than others, which is not a scientific task. We believe, however, that knowledge and trust are generally preferable in any democracy, despite ideological views. Finally, this study has highlighted the importance of the armed forces informing the public of what they do, why they do it, and how successful they are when doing it.
A cross-sectional study based on self-reporting has its strengths and weaknesses. It provides a good picture of the views of the public, including different subgroups, of their own knowledge and attitudes at a specific point in time. You can also see connections between the different variables that have been identified as relevant to understanding people's anchoring. However, there are also limitations. For example, people's views of their own knowledge may differ from their actual knowledge. In addition, a cross-sectional study never provides knowledge about causes and effects. On the other hand, it can provide a basis for formulating hypotheses of causal relationships.
This study was conducted at a specific point in time: 2015. Since then important political changes have been made concerning defense policy and rearmament, but the main problems -lack of public knowledge and the limited number of people connected to the SAF -have not changed dramatically. The deteriorating security situation, especially after the Russian aggression against Ukraine, represents one change from 2015 till today. Still, armed conflict or a military attack is low on the list of perceived threats, and the level of worrying in the public has been quite stable since 2015 (Swedish Civil Contingencies Agency 2018). These facts may indicate that the results of the study are still relevant. How anchoring and attitudes to defence are influenced by current events should nevertheless be an issue for further research, preferably with a longitudinal approach.
There is also a need for measuring both subjective and objective variables. For instance, this study did not include any measurement of the respondents' actual knowledge, only their perception of their own knowledge. Further research could address the need to measure actual knowledge about the defense and the SAF, or actual and observable behaviors of people as a result of an experienced threat or interest in defense issues.