• No results found

4. Methodology

4.4 Control variables

Multiple control variables were added to the model to account for variables that could potentially explain the relationships being tested. Academics like Porter (1980) have pointed out the fact that business characteristics significantly determine the performance of a firm.

Additionally, features of the entrepreneurial team in a social enterprise can substantially affect the characteristics of the social enterprise (Saebi, Foss, & Linder, 2019). As this study is

looking into the results of and behaviour in enterprises, it is imperative to include several characteristics to avoid the effects of confounding variables.

4.4.1 Firm size

Literature has indicated that firm size might affect the innovativeness and performance of firms due to older firms having a more extensive knowledge base (An, Zhang, You, &

Guo, 2018).Schumpeter (1943) has also argued firm size to positively influence the ability of a firm to deal with new business processes. For these reasons, it can be expected that firm size might influence the various variables being measured. The firm size was measured by the number of employees which is in line with existing strategic management literature (Zhang &

Luo, 2013). Firm size was added as a continuous variable to the model.

4.4.2 Firm age

The age of the firm can substantially influence bricolage behaviour and the overall performance of a firm. Older firms often have larger resource stock, and over time, firms get better at developing, acquiring, and managing resources (Peng & Luo, 2000). The

consequence is older firms having more options when handling resources. It has been found that entrepreneurs tend to engage less in bricolage behaviour when they have more options (An et al., 2018). Moreover, older firms tend to have less flexible routines and steady patterns of resource allocation. Resource stock and organisational flexibility can influence bricolage (Steffens, Davidsson, and Fitzssimmons, 2009). Therefore, firm age was included as a control variable.

Firm age was measured by the number of years the enterprise has been operating.

Participants answered a multiple-choice question with the following answer options - 0 to 3 years, 3 to 8 years, 9 to 15 years, 15 years or more. The variable was made into a

dichotomous variable to fit in regression analysis. Following existing studies, 8 years was used as a threshold between old and new social enterprises (An et al., 2018). Firms younger than 8 years were coded as ‘0’ and firm older than 8 years were coded a ‘1’.

4.4.3 Continent (EU vs USA)

The country that the social enterprise is registered influences the work culture, which can substantially influence the work behaviour. The way social entrepreneurs work heavily depends on the country of residence’s social, economic, political, and cultural context (Cozma

& Bocancia, 2020). Most social entrepreneurship literature focuses on the USA and there is a

great need to study it in a European context (Adro & Fernandes, 2021). Due to resources and time constraints, this study cannot make conclusions on the differences between different countries but can attempt to control the differences in work culture between Europe and America. Participants were asked to indicate in which country the social enterprise is a registered business. A dummy variable was created to indicate which social enterprise belonged to which continent. Europe was coded with “0” and the United States of America with “1”.

4.4.4 Enterprise type (product vs service)

The difference between service and product-based business has been discussed extensively in the innovation literature. Service-related businesses rely more heavily on mutual communication, joint engagement, and divided responsibilities between different individuals (Easingwood, 1986). On the other hand, product-related services tend to rely more on structured relationships by which resources and information are shared through existing stable routines. These characteristics could potentially result in significant differences concerning bricolage behaviour as bricolage involves a lot of flexibility (Schleimer &

Schulman, 2011). A dummy variable for enterprise type was created to indicate whether the social enterprise involves a product or service. Service was coded as “0” and product as “1”.

4.4.5 Goal (social vs environmental)

Social entrepreneurs can have a purpose more related to a social or environmental goal. The databases used to find respondents include social entrepreneurs with either of these goals. Literature suggests that there can be a difference between social enterprises with a social or an environmental goal. Social entrepreneurship has been known to transform social causes for a long time (Schumpeter, 1942); social entrepreneurs dealing with environmental issues are more recent. One of the main differences between these goals, as implied in literature, is the dissimilar process of opportunity discovery (Hall et al., 2010). For these reasons, an enterprise having a social or environmental goal was also included as a control variable to control for the differences between the two. Respondents were asked if their main goal was primarily focused on a social issue or environmental issue. A dummy variable was created with “0” for an environmental goal and “1” for a social goal.

4.4.6 Social desirability

A measurement for social desirability was also included as the type of questions in the survey might result in participants answering in a socially desirable way (Crowne & Marlowe, 1960). Answers that are given to create a more favourable impression could significantly comprise the survey’s validity (Crowne & Marlowe, 1960). This is particularly relevant in survey studies as it includes reported data (Tourangeau, 2000). Additionally, the self-reported data in this study includes information on organisational behaviours and outcomes. It is known that employees tend to present themselves in a favourable manner in research (Zerbe, & Paulhus, 1987).

The BIDR-40 scale by Paulhus (1991) is a widely used scale to measure social desirability. Due to its impractical length, this survey included the shorter version of the BIDR, namely the BIDR-16. This version has been found to be an excellent substitute (Hart, Ritchie, Hepper, & Gebauer, 2015). The BIDR exists out of two components - self-deception and deliberate self-presentation. Respondents answered on a 7-point scale ranging from

“totally disagree” to “totally agree”.

Two methods have been used to score the participants, namely continuous scoring and dichotomous scoring. Paulhus (1991) himself suggests the use of a dichotomous method by which only extreme scores count. This guarantees scores to include only exaggerated responses on already highly desirable items. However, many researchers have used

continuous scoring as dichotomous scoring can potentially lead to extremity bias and errors of discreteness (Cohen, 1983; Hart et al., 2015). On top of that, a study by Stöber and Dette (2002) have found, on average, higher reliability scores for continuous scoring.

For each respondent, two sets of scores were computed - a continuous and

dichotomous score. For the continuous score, negatively keyed items were reverse coded and all answers were summed up to give a sum score for each participant. For the dichotomous coding, negatively keyed items were also reverse coded. The dichotomous score was computed by giving one point for each 6th or 7th score. These scores were summed up to compute a dichotomous score for each participant.

A reliability test was conducted for both scores. The continuous score had a

Cronbach’s Alpha of .365. Such a score indicates low and unacceptable reliability. However, the Cronbach’s Alpha of the Dichotomous score indicated more acceptable reliability, Cronbach’s Alpha = .643. As Paulhus (1991) recommended the dichotomous scoring and resulted in higher reliability, this study will use it as a control variable instead of the continuous scoring.

GERELATEERDE DOCUMENTEN