**1. Introduction**

Attention towards energy and energy security over the last decade increased as the efficient military energy usage is considered to be one of the key enablers of military operational capabilities. This trend is reflected in European Union (EU) and NATO strategic priorities and initiatives that underline the importance of ensuring energy security for military operations as well as reducing the environmental impact of military operations [1–3]. In addition, a portion of research, technology, and innovation (RTI) projects in military and defense are focused on energy and particularly on renewable energy solutions. In general, three key factors that affect energy usage can be listed: energy generation technologies, energy management, and energy data collection and analysis systems together with energy behavior at military units [4,5]. Resonating with an overall trend in energy transition towards pro-environment energy usage [6–8], military RTI initiatives stress on technological solutions. Meanwhile, the energy behavior remains an inadequately explored factor in reducing the energy usage and thus increasing military energy security [9] and energy e fficiency [10]. Even though energy behavior in the military is gaining more attention, overall attempts in changing energy behavior of military personnel remains a managerial challenge.

The Capability-Opportunity-Motivation-Behavior (COM-B) model proposed by Michie et al. [11] and tested in numerous research projects is applied in this paper to investigate the energy-saving behavior in the military. The application of COM-B model to energy behavior leads to a better understanding of pro-environmental behavior in the military and facilitates a detailed analysis of the factors a ffecting this behavior. The model works in a context where three factors of the behavior (capability, opportunity, and motivation) are surrounded by managerial interventions, organizational policies, and limitations.

In the military context energy behavior is rooted into the trilemma of: (i) how to assure energy security for the military operations, (ii) how to use energy e fficiently, and (iii) how to reduce the environmental impact of the operations [1]. In this context, energy behavior plays a critical role after the energy policies and standard operating procedures are introduced and related technologies are deployed. All three considerations are gradually translated into the requirements for the United Nations (UN) peacekeeping operations as well as EU and NATO military environment promoting not only technological development but also energy behavioral changes. Specifically, the UN peacekeeping forces had implemented environmental policies in all peacekeeping missions since 2009 [12]. This includes the requirements for environmental managemen<sup>t</sup> systems that include energy, water, and waste management. As a new step in promoting pro-environmental energy usage in the military, the UN introduced its Environment Strategy of the UN Department of Field Support (DFS) which came into e ffect in January 2017. Its energy related objective "to reduce overall demand for energy through efficiencies" is planned to be realized by 2023 [2] (p. 2). This includes not only the requirements for energy e fficient infrastructure but also the behavioral incentives where "awareness-raising and behavioral change" [2] (p. 2) play an important role. In general, the UN initiatives complement NATO's approach on energy security, energy resilience, and the protection of critical energy infrastructure. The improvement of energy e fficiency becomes one of the key priorities [10,13], therefore NATO's approach is also focused on the military by "reducing the energy consumption of military vehicles and camps, as well as minimizing the environmental footprint of military activities" [14]. At the Brussels Summit in 2018, those priorities were emphasized to the Member States underlining the importance of "more education and training opportunities" [3]. This highly resonates with energy priorities in security and other sectors [15]. Energy and energy security as the strategic priority was elaborated through the activities of NATO Energy Security Center of Excellence that was established in Lithuania in 2012 in order to assist Strategic Commands, other NATO bodies, nations, partners, and other civil and military entities by supporting NATO's capability development process, mission e ffectiveness, and interoperability by providing comprehensive and timely subject matter expertise on all aspects of energy security.

Given the situation where the energy policies and standard operating procedures are already introduced, and energy related technologies are deployed the COM-B model explaining the behavior change factors becomes an e ffective tool for the further research of military energy behavior. According to the model's designers Michie et al. [11], three factors heavily influence the behavior: capability (C), opportunity (O), and motivation (M) (Figure 1). The model explores individual's behavior in the organizational context and provides the basis for managerial interventions [16], as well as includes main steps for behavioral change [17]. According to this model, all three conditions must be met in order to make an influence on individual's energy behavior: the individual's physical and social capability, individual's social and physical ability to explore new opportunities, and self-motivation as the crucial part of the behavior change [18]. This model was theoretically grounded and applied in a wide variety of contexts: nutrition [19], smoking [20], physical activity [21], as well as for energy use by households [22] and other end-users [23]. The COM-B model was applied to improve energy behavior in the military too [14,24]. However, those behavioral interventions in military and defense were purely practical and lacked intrinsic validity. This suggests that the COM-B model should first of all be tested as a solution/construct in the military context.

**Figure 1.** The original construct of COM-B model for behavioral change [11,16,21].

Assuming that a military context differs from a civilian context, the COM-B model needs to be validated and factors influencing the pro-environmental energy behavior in the military need to be identified. Studies showed that military members are indoctrinated already at the beginning of their military career [25]. Consequently, military culture penetrates attitudes and behavior, whereas individuals report strong identification with the military [26–28]. The military's impact on an individual increases along the duration of the military service [29]. Based on this evidence it could be assumed that the awareness of energy criticality in the military is increasing over the years of service, as shown by research in military energy efficiency [10]. Capability, opportunity and motivation affecting pro-environmental energy behavior are positively strengthened with the understanding that energy in military is considered to be a critical combat's "tooth" [1]. These considerations are taken into account when planning and conducting military operations for professional soldiers as well as for military conscripts. It is also assumed that the conscripts' perception of the importance of energy in the military differs from the professional soldiers. The conscripts are serving for a short term and they can be not ye<sup>t</sup> fully indoctrinated [30]. Their capabilities, opportunities, and motivation to behave in a pro-environmental way could be linked to their civil life experiences of a green lifestyle [31] and not to the military service.

Efficient usage of military equipment and infrastructure is perceived as a potential for the freedom of action and an opportunity for enhanced capabilities of the military operations [1]. Consequently, important parts of the military's RTI are directed at addressing those issues. For example, new technologies that are used in expeditionary environment as well as in fixed military installations are focused on improving energy supply and reducing the usage [32,33], but the awareness of energy behavior remains limited. Despite the emphasis on technical solutions related to energy, the call for behavioral change at a unit level remains a priority [10,34]. Considering that at military bases soldiers are semi-isolated from their existing ties as well as their life outside the military, their bond with a military unit rises [35], and therefore, pro-environmental energy behavior must be analyzed not only from an individual's behavioral perspective, but also from the individual's social identification with the military unit. During the demanding military training, professional soldiers as well as conscripts are trained in order to increase their awareness of their own behavior in the context of the unit and their peers [35]. Therefore, COM-B model must focus not only on the individual's behavior, but also include the collective military behavior of the military unit.

The unique feature of our research is that we disaggregate two groups of energy behavior: individual energy behavior ("my" behavior) and unit energy behavior ("our" behavior). While the former follows the COM-B traditional paradigm used and validated in a series of studies (e.g., [11,16,21,36]), the latter relies exclusively on military research that puts a stress on soldier's bond with a military unit [28,37].

The purpose of this study is to explore how a politically and institutionally favorable environment that forces military transition towards the pro-environmental energy behavior is reflected at a military unit level. The rest of the paper proceeds as follows. First, we outline our research instrument and measurement model by extending the construct of COM-B model for behavioral change in the military. Next, we perform data analysis using a series of statistical tests. By applying the COM-B model we investigate energy behavioral factors and postulate that the three behavioral change factors—capability, opportunity, and motivation—are positively linked not only with individual soldier's behavior, but also with a collective energy behavior of the unit. Finally, discussion and conclusions are presented in the last section.

#### **2. Materials and Methods**

To measure the relationships between energy behavior in the military and a ffecting factors we collected data by making an annual survey for two consecutive years. The research was performed using traditional paper questionnaires at a selected Lithuanian military unit (fixed installation). The first survey was executed in October 2018 followed by the second survey in October 2019. During the first stage of data collection, 235 soldiers, non-commissioned o fficers, and o fficers were surveyed, while during the second stage the number of respondents was 219 soldiers, NCOs, and o fficers. Most respondents were conscripts (70 percent), followed by professional soldiers, NCOs, and o fficers (30 percent). Hence, the total sample was N = 454. Detailed demographic information of the respondents is presented in Table 1. In some categories, the sum of all answers is less than 100% due to some missing responses. They were eliminated from further analysis.


**Table 1.** Demographic information of the respondents.

The questionnaire consisted of Likert one-to-five scale questions where value 1 corresponded to "strongly agree" and value 5 to "strongly disagree". In total there were 18 randomly listed statements related to COM-B measuring three latent variables: Motivation to save energy, Capabilities of the respondents, and available opportunities to exhibit the energy-saving behavior. Five statements were related to the capability category, seven statements to the opportunity category, and four to motivation. Two questions were attributed to the behavior category; they are analyzed later in the text. The ordinal scale of Likert questions limited statistical comparisons, and therefore several questions were included to measure the same variable, e.g., if three questions are used then the sum of the answers to these questions goes from 3 to 15, therefore it can be treated as an interval variable. We used Cronbach's alpha (CA) coe fficient to measure the internal consistency of the composing questions. During the process, one motivation-related question and two capability-related questions were eliminated from

further analysis due to their detrimental effect on the CA coefficient of the corresponding variables. In other words, they were independently measuring something else than intended.

The three latent variables, their questions, and reliability (the CA coefficients) are presented in Table 2. Observed variables were also validated individually: none of them have more than 5% of the missing values. Average values of all questionnaire answers are presented in Figure 2.



1 The question had only two values—Yes and No. 2 The answers had to be inverted before performing the analysis.

**Figure 2.** Average values of all answers. "S" denotes overall satisfaction (max 10), "B1"—"my" behavior, "B2"—"our" behavior, other variables are defined in Table 2, and they range from 1 (best) to 5 (worst).

Additionally, we performed the principal component analysis on all answers of the questionnaire using SPSS software in order to assess its internal consistency. Sum of eigenvalues of the main five components covered almost 60% of the variance. The first component corresponded exactly with the motivation questions, explaining their high CA coefficient and the reliability of measurement. The questions have high loading factors above 0.7 (see Table 2) and their contribution to other latent variables is negligible. The second component grouped questions C1 and C3, with the C2 question also having a large load (0.71) in the third component and therefore significantly decreasing the reliability of capability measurements—it is the lowest among the latent variables. The remaining seven questions were grouped into three components separated in Table 2 by horizontal lines, where O5, O6 and O7 constituted the strongest component of the three, O1 went together with O4, and the last component consisted of O2 and O3. Together the seven questions were assigned to the opportunity variable as

initially intended. Overall, the analysis validated the independence of the three latent variables and the reliability of their measurement.

According to the COM-B model [38] the behavior is determined by the three constituting factors: the previously defined latent variables Capability (*C*), Opportunity (*O*), and Motivation (*M*). For the purposes of further analysis of the data, the three variables were combined into a single *COM* variable:

$$\text{COM} = 1 - \left[ (\text{C} - 3) / 12 + (\text{O} - 7) / 25 + (\text{M} - 3) / 12 \right] / 3 \tag{1}$$

where the scale of each variable is shifted to zero and normalized to 1 so that their contribution has the same weight, and then the normalized sum value is inverted to produce the combined *COM* variable with values from 0 (worst) to 1 (best).

Variable B (behavior) was tested using two separate questions that represent connotations related to "My" versus "Our" energy behavior. It differs from the original COM-B construct presented in Figure 1 where only one behavioral indicator is foreseen. We took into consideration the uniqueness of the military culture where collective behavior and soldiers' psychological bond with his/her unit is manifested and adopted the COM-B model to the military context. "My" behavior (B1) was measured using a statement "I turn off the electricity when I stop using it". For "Our" energy behavior (B2) we used a statement that represents a commitment to unit's performance: "We manage energy well in the unit". All together these two statements created a We-I behavioral construct for measuring energy behavior in the military (Figure 3). These statements were measured using the Likert scale. Taking into consideration that COM-B model works in a context, we additionally tested an overall workplace satisfaction (see variable "S" in Figure 2) of the soldiers using one 10-point Likert scale question.

**Figure 3.** An extended construct of COM-B model for behavioral change in the military.
