**3. Materials and Methods**

The study followed a sequential exploratory mixed-methods approach [62] and combined ethnographic case studies with a subsequent quantitative survey (Figure 2). To gain an understanding of how decisions on energy efficiency are constituted and how the enterprises deal with energy efficiency issues in everyday work life, firstly, ethnographic case studies [63] were carried out on 10 industrial SMEs using qualitative interviews and observations as the main methods. Secondly, and based on their key results and orienting on the discussed framework, the questionnaire was conceptualized, which was addressed to 500 SMEs.(An SME is here intended as an enterprise according to the 2003 recommendation of the European Council [64]).


**Figure 2.** Research design.

#### *3.1. Qualitative Analysis: Case Studies*

The ethnographic organizational analyses formed the starting point of the research work, which primarily aimed at the observation and reconstruction of situations in everyday work life in order to find out "how work is organized and how that organizing organizes people". [63] (p. 1). This qualitative approach is particularly suitable for studies of organizational cultures [65] (p. 20). Exploratively designed case studies focused on several basic questions: How do decisions for energy efficiency come about in the SMEs? Which driving (or constraining) processes and aspects can be identified? How are energy efficiency issues treated, organized, and communicated in everyday work life? The sample (Table 1) comprised 10 manufacturing enterprises from different industrial sectors (chemicals, minerals, engineering, and machinery). The cases were selected by theoretical sampling [66] according to the premise of "minimum/maximum contrast", especially with regard to energy intensity, sector, and number of employees of the enterprises. All participating enterprises are family businesses (ownership and/or control). (Ninety-one percent of all enterprises in Baden-Württemberg are family controlled and 88% of all enterprises in the manufacturing sector are family controlled [67].). The data generation was mainly based on qualitative interviews [68] with members from different divisions within the enterprises. Around seven to 10 interviews per SME (one-on-one and multi-person) were carried out in each enterprise. Interviews were conducted with managing directors, owners, and energy managers as well as production workers and controlling, marketing, and human resources staff (Table 2). In addition to the interviews, participating observations [69] and artefacts (e.g., company presentations, homepages, work instructions) were included in the analysis. Depending on what was appropriate from the perspectives of the enterprises and their members, the observations were either performed as fly-on-the-wall (e.g., at meetings of formal or informal energy teams or meetings with external energy efficiency consultants) or following the daily routines throughout the work day. The data (primarily interview text and observation protocols) were analyzed by system analysis [68,69]. This hermeneutic approach focuses on the interpretation of the data in two steps. In the first step, hypotheses are formed, from which subjective and organizational meanings and conditions could lead to a statement (e.g., "Only the boss is responsible for energy.") or observation. In the second step, hypotheses are formed, from which these meanings and conditions could have effects for the organization (e.g., centralization of competences and responsibility for measures). The field research was carried out by one person of the University of Reutlingen, taking about one year all together and spending around one work week in each SME.


**Table 1.** List of enterprises participating in the case studies.

> \* Self-evaluation of the enterprises.

**Table 2.** Shares on the roles of the interviewed persons of the SMEs.


## *3.2. Quantitative Analysis: Survey*

Based on the case study research, a quantitative survey was conceptualized through a questionnaire comprising questions on topics such as the importance of energy e fficiency, measures, support measures, the influence of the business environment, the relevance of employee behavior, financing, and others. The questionnaire consisted of 20 di fferent types of questions including multiple choice questions, Likert scale questions, and matrix questions as well as single-choice questions (see Appendix A). The survey took place from May to June 2018 and around 500 SMEs from the federal state of Baden-Württemberg, Germany, were surveyed. A market research institute was commissioned with the survey itself while the analysis was conducted by the Institute for Energy E fficiency in Production, Universität Stuttgart. On the basis of available data bases and selected by company size (micro-, small-, middle-sized) and sectors, the SMEs were reached by telephone. The energy demand of the surveyed SMEs is shown in Figure 3. The distribution of the responding SMEs with respect to the number of employees is shown in Table 3. Analogous to the sample procedure for the survey of the Energy E fficiency Index of German Industry [10], the distribution of the sample size did not correspond to the real distribution of SMEs by enterprise size in Germany [70]. In order to allow valid results for small- and medium-sized enterprises, the share of micro enterprises was reduced. Furthermore, a thorough coverage of the manufacturing sector including subsectors of particular importance for the state of Baden-Wuerttemberg (such as the manufacturing of metal products and processing, mechanical engineering, and the automotive sector) was targeted. The respondents were either owners, managing directors, technical managers, production managers, energy or environmental managers, controlling, or other persons of the SMEs. Naturally, not all enterprises were open to be interviewed. Therefore, a self-selection bias can be assumed. In addition to descriptive data analysis, a correlation analysis (using SPSS) investigating the factors driving the internal establishment of energy e fficiency was conducted. The correlation analysis was performed using ordinally scaled variables, with the Spearman rho rank correlation coe fficient as an indicator of correlation.

**Figure 3.** Energy demand of the surveyed SMEs.

**Table 3.** Sample composition, according to size.

