*4.1. fsQCA*

The present study used fsQCA [52,55,72] to assess the conditions of need and sufficiency. The need for a condition was based on seeing the impact on the achievement of PRAC (a consistency score greater than 0.90) [73]. The sufficiency of a condition was based on seeing the extent of its relationship with PRAC (configurations of various conditions that lead to PRAC). Acceptable solutions must respect the coverage level within the range of limits from 0.25 to 0.90 [72] and the coherence threshold of 0.75 [52,72].

In Model 1 (PRAC = f(HRC, OLC, EL, EXP, SIZ)), seven alternative causal configurations exist that lead to PRAC, and three alternative causal configurations exist for the absence of PRAC. Therefore, in Model 1, both H1 and H2 are supported.

In Model 2 (PRAC = f(HRC, ITS, EL, EXP, SIZ)), six alternative causal configurations exist that lead to PRAC, and four alternative causal configurations exist for the absence of PRAC. Therefore, in Model 2, both H3 and H4 are supported.

Finally, in Model 3 (PRAC = f(HRC, ITS, OLC, EL, EXP, SIZ)), eight alternative causal configurations exist that lead to PRAC, and seven alternative causal configurations exist for the absence of PRAC. Therefore, in Model 3, both H5 and H6 are supported.

Therefore, these results offer managers seven different options to reach PRAC for Model 1, six different options to reach PRAC adoption for Model 2 and eight different options to reach PRAC for Model 3. In short, all the hypotheses proposed in our research models are supported.

The results show that firms that invest heavily in establishing learning systems for their employees and achieve a high OLC are more likely to be able to perform PRAC because they will have more alternative paths that lead to it. Specifically, seven causal configurations lead to PRAC compared to only three that lead to its absence according to Model 1. This is a clear sign that PRAC in firms can be achieved through the implementation of a culture of learning where firms implement mechanisms that contribute to the development of knowledge and ethical values to obtain good adoption and implementation of PRAC. On the other hand, from the results obtained from Model 2, it is clear that technology is also a very necessary tool in firms for PRAC. In this case, Model 2 shows that there are six paths that lead to PRAC with technology, versus only four that lead to its absence.

If we interpret the results of both models separately, we can indicate that technology is important for PRAC in firms, but the factors derived from the learning capacity are essential, since with OLC, we can find more ways that lead to achieving PRAC (seven in Model 1) and less in its absence (three in Model 1) than with a model in which there is only a strong investment in technological and non-human factors, where we see that there is one less path that leads to PRAC (six in Model 2) and one more path than in Model 1 that can lead to its absence (four in Model 2). This means that only with the implementation of technology can firms achieve PRAC through different, alternative routes, but there are also a high number of ways that can lead to its absence—that is, there is a probability that the expected result can be achieved as much as other results that are not expected. However, with investment in OLC, it can occur because there are also causal configurations that lead us to the absence of PRAC. The probability of this being the case is lower because in this scenario, there are more paths that lead us to the expected result—that is, to the adoption of PRAC.

Finally, it should be noted that the results show that joint investment in OLC and ITS is optimal for the PRAC, since by integrating factors of both types, we add one more path that leads to the desired result with respect to Model 1. However, it is important to point out that Model 3, where OLC and ITS are included together to reach PRAC, is also the research model that presents more paths towards PRAC absence. From this, it can be interpreted that technology is an adequate tool to achieve the result, but always with adequate human control so as to not incur precisely the opposite effect.
