*4.2. Causal Configurations*

Table 3 shows the reported intermediate solution related to Model 1. This solution shows the seven causal configurations that lead to PRAC. It can be seen that there are four causal configurations with four conditions and three configurations of three conditions. Table 3 shows that HRC, OLC and SIZ are the three relevant core conditions for PRAC. HRC is present in one causal configuration that leads to PRAC, whereas OLC is present in three and SIZ in two. Model 1 also shows that there are three causal configurations for the absence of PRAC (Table 4). Here, HRC, EL, EXP and SIZ are the core conditions for the absence of PRAC. However, OLC is not present in these solutions.

**Table 3.** Intermediate solutions for PRAC with OLC. Model 1(a): PRAC = f(HRC, OLC, EL, EXP, SIZ). Overall solution coverage: 0.641398. Overall solution consistency: 0.770815.


Note: Black circles (•) indicate the presence of a condition, and open circles (◦) indicate its absence. Large circles indicate core conditions (present in both the parsimonious and intermediate solutions), and small ones identify peripheral conditions (present only in the intermediate solution). Blank spaces indicate that the condition does not contribute to the configuration. PRAC = adoption of environmental practices; HRC = human resource cost; OLC = organizational learning capability; ITS = information technology support; EL = education level; EXP = experience; SIZ = firm size. In the three research models PRAC correspond to a function with combinations of the following variables. Research model 1: HRC, OLC, EL, EXP and SIZ. Research model 2: HRC ITS, EL, EXP and SIZ. Research model 3: HRC, ITS, OLC, EL, EXP and SIZ. However, in research model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)) and research model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), OLC has five dimensions (OLC-E, OLC-R, OLC-I, OLC-D and OLC-P), it is measure using the fsQCA "fuzzyand" function. It is a mathematical logical operation in Boolean algebra called "intercept." Therefore, when OLC appears in a causal configuration, it means that this condition is a five-dimensional cumulative condition.

**Table 4.** Intermediate solutions for~PRAC with OLC. Model 1(b): ~PRAC = f(HRC, OLC, EL, EXP, SIZ). Overall solution coverage: 0.393352. Overall solution consistency: 0.806792.


Note: Black circles (•) indicate the presence of a condition, and open circles (◦) indicate its absence. Large circles indicate core conditions (present in both the parsimonious and intermediate solutions), and small ones identify peripheral conditions (present only in the intermediate solution). Blank spaces indicate that the condition does not contribute to the configuration. PRAC = adoption of environmental practices; HRC = human resource cost; OLC = organizational learning capability; ITS = information technology support; EL = education level; EXP = experience; SIZ = firm size. In the three research models PRAC correspond to a function with combinations of the following variables. Research model 1: HRC, OLC, EL, EXP and SIZ. Research model 2: HRC ITS, EL, EXP and SIZ. Research model 3: HRC, ITS, OLC, EL, EXP and SIZ. However, in research model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)) and research model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), OLC has five dimensions (OLC-E, OLC-R, OLC-I, OLC-D and OLC-P), it is measure using the fsQCA "fuzzyand" function. It is a mathematical logical operation in Boolean algebra called "intercept." Therefore, when OLC appears in a causal configuration, it means that this condition is a five-dimensional cumulative condition.

> The intermediate solution in Model 2 (Table 5) shows six configurations that lead to PRAC. ITS and SIZ are the core conditions for PRAC. ITS is present in five causal configurations and SIZ in one. Model 2 also shows four causal configurations for the absence of PRAC (Table 6). Here, HCR, EL and SIZ are the core conditions for the absence of PRAC. However, ITS is not present in these solutions. The literature indicates that firms

that have a high investment in ITS tend to have a good PRAC [27,28], since technology provides the necessary mechanisms to convert inputs into sustainable outputs [25,29].

**Table 5.** Intermediate solutions for PRAC with ITS. Model 2(a): PRAC = f(HRC, ITS, EL, EXP, SIZ). Overall solution coverage: 0.786787. Overall solution consistency: 0.737200.


Note: Black circles (•) indicate the presence of a condition, and open circles (◦) indicate its absence. Large circles indicate core conditions (present in both the parsimonious and intermediate solutions), and small ones identify peripheral conditions (present only in the intermediate solution). Blank spaces indicate that the condition does not contribute to the configuration. PRAC = adoption of environmental practices; HRC = human resource cost; OLC = organizational learning capability; ITS = information technology support; EL = education level; EXP = experience; SIZ = firm size. In the three research models PRAC correspond to a function with combinations of the following variables. Research model 1: HRC, OLC, EL, EXP and SIZ. Research model 2: HRC ITS, EL, EXP and SIZ. Research model 3: HRC, ITS, OLC, EL, EXP and SIZ. However, in research model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)) and research model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), OLC has five dimensions (OLC-E, OLC-R, OLC-I, OLC-D and OLC-P), it is measure using the fsQCA "fuzzyand" function. It is a mathematical logical operation in Boolean algebra called "intercept." Therefore, when OLC appears in a causal configuration, it means that this condition is a five-dimensional cumulative condition.

**Table 6.** Intermediate solutions for~PRAC with ITS. Model 2(b): ~PRAC = f(HRC, ITS, EL, EXP, SIZ). Overall solution coverage: 0.634377. Overall solution consistency: 0.797445.


Note: Black circles (•) indicate the presence of a condition, and open circles (◦) indicate its absence. Large circles indicate core conditions (present in both the parsimonious and intermediate solutions), and small ones identify peripheral conditions (present only in the intermediate solution). Blank spaces indicate that the condition does not contribute to the configuration. PRAC = adoption of environmental practices; HRC = human resource cost; OLC = organizational learning capability; ITS = information technology support; EL = education level; EXP = experience; SIZ = firm size. In the three research models PRAC correspond to a function with combinations of the following variables. Research model 1: HRC, OLC, EL, EXP and SIZ. Research model 2: HRC ITS, EL, EXP and SIZ. Research model 3: HRC, ITS, OLC, EL, EXP and SIZ. However, in research model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)) and research model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), OLC has five dimensions (OLC-E, OLC-R, OLC-I, OLC-D and OLC-P), it is measure using the fsQCA "fuzzyand" function. It is a mathematical logical operation in Boolean algebra called "intercept." Therefore, when OLC appears in a causal configuration, it means that this condition is a five-dimensional cumulative condition.

> Finally, the intermediate solution in Model 3 (Table 7) shows eight configurations that lead to PRAC. Here, ITS, OLC and SIZ are the core conditions. ITS is present in five causal configurations, OLC in three and SIZ in one. In addition, there are seven causal configurations for the absence of PRAC (Table 8). Here, HRC, OLC, EL and SIZ are the core conditions. However, ITS is not present in these solutions.

**Table 7.** Intermediate solutions for PRAC with OLC and ITS. Model 3(a): PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ). Overall solution coverage: 0.721535. Overall solution consistency: 0.771006.


Note: Black circles (•) indicate the presence of a condition, and open circles (◦) indicate its absence. Large circles indicate core conditions (present in both the parsimonious and intermediate solutions), and small ones identify peripheral conditions (present only in the intermediate solution). Blank spaces indicate that the condition does not contribute to the configuration. PRAC = adoption of environmental practices; HRC = human resource cost; OLC = organizational learning capability; ITS = information technology support; EL = education level; EXP = experience; SIZ = firm size. In the three research models PRAC correspond to a function with combinations of the following variables. Research model 1: HRC, OLC, EL, EXP and SIZ. Research model 2: HRC ITS, EL, EXP and SIZ. Research model 3: HRC, ITS, OLC, EL, EXP and SIZ. However, in research model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)) and research model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), OLC has five dimensions (OLC-E, OLC-R, OLC-I, OLC-D and OLC-P), it is measure using the fsQCA "fuzzyand" function. It is a mathematical logical operation in Boolean algebra called "intercept."Therefore,whenOLCappearsinacausalconfiguration,itmeansthatthisconditionisafive-dimensionalcumulativecondition.

**Table 8.** Intermediate solutions for~PRAC with OLC and ITS. Model 3(b): ~PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ). Overall solution coverage: 0.631863. Overall solution consistency: 0.796658.


Note: Black circles (•) indicate the presence of a condition, and open circles (◦) indicate its absence. Large circles indicate core conditions (present in both the parsimonious and intermediate solutions), and small ones identify peripheral conditions (present only in the intermediate solution). Blank spaces indicate that the condition does not contribute to the configuration. PRAC = adoption of environmental practices; HRC = human resource cost; OLC = organizational learning capability; ITS = information technology support; EL = education level; EXP = experience; SIZ = firm size. In the three research models PRAC correspond to a function with combinations of the following variables. Research model 1: HRC, OLC, EL, EXP and SIZ. Research model 2: HRC ITS, EL, EXP and SIZ. Research model 3: HRC, ITS, OLC, EL, EXP and SIZ. However, in research model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)) and research model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), OLC has five dimensions (OLC-E, OLC-R, OLC-I, OLC-D and OLC-P), it is measure using the fsQCA "fuzzyand" function. It is a mathematical logical operation in Boolean algebra called "intercept." Therefore, when OLC appears in a causal configuration, it means that this condition is a five-dimensional cumulative condition.

> Therefore, the results show that the assumptions of fsQCA [53] are fulfilled. It can be seen how more than one configuration of causal conditions leads to PRAC, and thus, the existence of alternative combinations of causal conditions leading to PRAC is confirmed. Furthermore, since these alternative causal configurations can produce the same result, the principle of equifinality is confirmed. The solutions present coverage and coherence values that respect the limits set by the literature [72].

> Furthermore, our research models show that there are no pathways leading to PRAC when ITS or OLC are not included in the model as a causal condition. In this way, our work reveals the key role that technology and human resources managemen<sup>t</sup> have in achieving PRAC. The increasing introduction, improvement and application of new technologies in firms help, in a clear and simple way, to improve and apply more environmental practices within firms. New technologies, greater investment in human resources, a culture of organizational learning and other firm and manager characteristics such as education, experience or firm size can actually encourage the adoption of environmental practices [2].

## **5. Discussion and Conclusions**

Few empirical studies have shown the different pathways for PRAC [8,37,53]. This paper contributes to the theoretical and empirical research with a relevant and original qualitative method, fsQCA. The fsQCA in our research models shows new relations, pathways and alternative configurations for PRAC that other quantitative tools do not discover. Therefore, fsQCA offers alternative complex configurations of conditions for PRAC. Our research models show that both sources are relevant: technological- and humanresources-based. Contrary to quantitative approaches that only show single linear models of PRAC in SMEs [73], our results show several pathways for PRAC and its absence. They are relevant findings: the former points SMEs towards the ways to reach PRAC and the latter warns SMEs about the pathways that do not lead to PRAC.

Given the results from our research models, all hypotheses are supported. In Model 1(a), OLC is a core condition for PRAC in three pathways. It is the main core condition in this model for PRAC. However, in Model 1(b), the absence of OLC is a core condition in two of three pathways for the absence of PRAC. Similar considerations can be made about ITS in Model 2(a and b). In Model 2(a), ITS is a core condition for PRAC in five pathways of six. However, in Model 2(b), the absence of ITS is a core condition in all pathways for the absence of PRAC, totaling four pathways.

It is notable that OLC and ITS are complete substitutes (configuration 1 in Table 3 and configuration 5 in Table 5). Such results reveal similar substitutable contributions of the technological- and human-resources-based sources to PRAC. Yet, SMEs have other options that lead them to PRAC. Regarding Tables 4 and 6, the evidence shows that the absence of either OLC or ITS is associated with not adopting environmental practices.

Finally, in the full and integrative Model 3(a), ITS is a core condition for PRAC. ITS is present in five pathways for PRAC, and OLC is present in three pathways for PRAC, for eight total pathways. It is a relevant finding because technology is the most important factor leading to PRAC. Model 3(b) also shows that the absence of ITS is a core condition in six pathways for the absence of PRAC.

Therefore, our separate research models show that there are more ways to reach PRAC if we take into account human learning factors within the organization, such as OLC, than if we only adopt technological factors, such as ITS. In this sense, with Model 1 (OLC), seven paths lead us to PRAC, and with Model 2 (ITS), six paths lead us to PRAC. It is also relevant to note that the model that incorporates OLC without ITS (Model 1) offers firms fewer avenues in which they face the absence of PRAC, as there are only three. On the other hand, the model that incorporates ITS without OLC (Model 2) offers firms one more way to avoid PRAC; that is to say, in this case, four routes can condition the absence of PRAC, one more than in the previous case. Therefore, if we analyze these results separately, we could advise firms that, mainly, the optimum component that can shape and be present in different combinations to arrive at PRAC is the OLC. Additionally, we may interpret it such that depending on technology involves a higher risk of not achieving PRAC.

On the other hand, the main conclusions drawn from the complete and integrative research model shown in this paper to obtain PRAC show that it is desirable for firms to dedicate their efforts to joint investment in personnel managemen<sup>t</sup> practices and the adoption of new technologies for obtaining PRAC. As can be seen in Model 3, when the two sources (OLC and ITS) are integrated, more paths are obtained to reach PRAC than when a firm decides to focus more on one source than on another to obtain PRAC. Our third model makes it clear that the adoption of information technologies such as key training, storage and communication can help firms to be more efficient when implementing PRAC. Regarding the lack of PRAC (Table 8), it is worth noting that all the configurations present the lack of OLC or ITS or both. Such findings seem to validate a contrario the relevancy of such conditions to achieve PRAC.

Finally, there seems to be a special contribution of ITS to PRAC, since there is no configuration leading to PRAC in the absence of ITS (Tables 5 and 7). On the contrary, it is possible to reach PRAC with the absence of OLC (configurations 3, 4 and 5 in Table 3, and configurations 2, 3 and 4 in Table 7). Such findings give strong support to arguments in the literature on the technological basis of PRAC [74].

In addition, in the three research models, other conditions such as EL, EXP and SIZ are relevant for obtaining PRAC. For example, in Model 1, SIZ is a central condition to reach PRAC in two causal alternatives. EL and EXP are important as well, but they are not central conditions. In addition, it must also be taken into account that the conditions related to the manager and the size of the firm, if not managed properly, can lead to other combinations or paths heading towards the absence of PRAC. In Model 2, SIZ is also a central condition for obtaining PRAC in at least one path. The same happens in Model 3.

In summary, our findings indicate that by focusing on variables such as OLC, a firm can find more paths that lead to PRAC. Additionally, with a joint focus on OLC and ITS, it must be taken into account that only investing in ITS and not in OLC is riskier when trying to implement PRAC.

The main goal of this paper was to find complementary pathways and conclusions about PRAC by applying fsQCA [73–75]. Our study shows that fsQCA is a strong tool that allows discernment of how factors are combined to generate (or not) PRAC. Moreover, fsQCA allows the important roles of OLC and ITS in inducing PRAC to be checked. Likewise, fsQCA discovers asymmetrical relations between conditions.

For managers, the findings indicate that to adopt PRAC in firms, it is necessary to consider a TBL approach and factors derived from it. States should focus on new formulas and regulations that help firms to be able to adopt PRAC through a clear commitment to their human resources and technology. In addition, several practical recommendations can be derived from our study:

	- ∗ Training in sustainability for the application of PRAC.
	- ∗ Management of motivation and values of employees oriented towards PRAC.
	- ∗ Foster a corporate culture of sustainability.
	- ∗Creation of internal rules for correct application of PRAC.

	- ∗ Implementation of environmental information collection systems to determine the needs internal and external to the firm.
	- ∗ Code of ethics for all members of the firm that facilitates the adoption of PRAC.
	- ∗ Implementation of energy-saving practices.
	- ∗ Implementation of reverse logistics practices.
	- ∗ Diversification of products towards economic products.
	- ∗ Training on social and environmental needs.

We are aware that this study has some limitations. This paper focused on a limited population segment: 349 Portuguese SMEs certified as innovative in various industry categories (manufacturing, power and gas supply; water supply and pollution; building; vehicle trade and repair; transport and storage; catering; information and communication; housing; scientific activities; administrative activities; health activities; other services). Future studies should focus on other segments but also on other economic agents (large businesses, transnational corporations, etc.). Although our study to analyze the representativeness of the sample in the population was based on an analysis of variance (ANOVA) between firms that responded early and those that responded late, future studies could perform further analyses of representativeness, for example, between how the firms of one

industrial sector and another behave among the twelve that were represented in this study to ensure that there are no significant differences between the responses of the different industrial sectors. In addition, our study focused only on three factors that can affect the adoption of environmental practices in firms: company size and the level of education and experience of the manager. Future studies may include more business factors such as the type of firm according to its legal and commercial form and more characteristics of the manager, such as political ideology, among others, that may affect the adoption of environmental practices. Another constraint is that this research is confined to Portugal. The conclusions may have been slightly different if the survey had a wider geographical extension or if it were answered in another country, so the use of an international database may allow improvement of the conclusions that we have extracted in this work. Although the response rate was representative, it can be considered slightly low (5.1%). Another issue that could be addressed in future research is the sustainability of ITS. It is certain that ITS leads to PRAC, but the development of technology can also affect PRAC negatively. Technology requires intensive computation resources with large energy consumption. Therefore, the findings of this research should be interpreted with the above considerations and new studies. Finally, let us point out that the use of alternative analytical tools to fsQCA based on fuzzy sets, such as fuzzy correlation indexes and fuzzy multiple criteria decision-making, and other quantitative methods, such as SEM, etc., should be considered in future research.

**Author Contributions:** Conceptualization, L.M.-P., J.G. and C.C.; methodology, L.M.-P., J.G. and C.C.; software, L.M.-P.; validation, L.M.-P., J.G. and C.C.; formal analysis, L.M.-P.; investigation, L.M.-P., J.G. and C.C.; data curation, L.M.-P.; writing—original draft preparation, L.M.-P.; writing— review and editing, L.M.-P., J.G. and C.C.; visualization, L.M.-P., J.G. and C.C.; supervision, J.G. and C.C.; project administration, J.G. and C.C.; funding acquisition, L.M.-P., J.G. and C.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors are grateful for the economic support provided by FCT (Fundação para a Ciência e Tecnologia Portugal) gran<sup>t</sup> number UIDB/04521/2020, by "Agencia Estatal de Investigación" (AEI) of the Ministerio de Ciencia e Innovación del Gobierno de España gran<sup>t</sup> number PID2019-107546GA-I00, by Consejería de Educación of the Junta de Castilla y León and the "Fondo Europeo de Desarrollo Regional" (FEDER) gran<sup>t</sup> number SA106P20, and by Junta de Castilla y León and the European Regional Development Fund for the financial support to the Research Unit of Excellence "Economic Management for Sustainability" (GECOS) gran<sup>t</sup> number CLU-2019-03.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Authors acknowledge helpful comments of anonymous reviewers.

**Conflicts of Interest:** The authors declare no conflict of interest.
