CAPÍTULO II: DISEÑO, ELECCIÓN E INSTALACIÓN DEL EQUIPO
Foto 8: Reactor Teqma
2.3 Comparativa de los distintos equipos Ventajas e inconvenientes Elección del equipo.
This section explains the results of the analysis for testing hypotheses 1 to 8. As explained in Chapter Three (Section 3.3 and 3.4), hypotheses 1 to 4 (or Stage A) are related to the mediational effects of business-level capabilities (exploratory R&D, exploratory marketing, exploitative R&D, and exploitative marketing) and their integrations in the relationships between exploratory strategy, exploitative strategy, new product performance, and established product performance. Hypotheses 5a, 5b, 6a, and 6c are related to the mediational effects of marketing ambidexterity and R&D ambidexterity in the relationships between strategic ambidexterity, new product performance, and established product performance. Further, hypotheses 7a, 7b, 8a, and 8c are related to the mediational effects of product performance and established product performance in the relationships between marketing ambidexterity, R&D ambidexterity, and firm performance.
The mediational effect represents the extent that the relationship between two variables is more complex than a simple bivariate relationship, and it might be formed through one or more additional variables (Fairchild & MacKinnon, 2009). Instead of demonstrating the existence of an effect between exogenous and endgenous variables, the mediational effect seeks to understand the mechanism(s) by which an effect between two variables operates via one or more mediator variables (Preacher & Hayes, 2004, 2008). In other words, the mediational effect seeks to answer the question of the extent an exogenous variable influences an endgenous variable (Fairchild & MacKinnon, 2009).
This study examined the mediational effect following the approach suggested by James and Brett (1984; Kenny, Kashy, & Bolger, 1998; Siren et al., 2012). Following this approach, the mediation model was tested with a path from the exogenous variable to the mediator(s) and from the mediator(s) to the endgenous variable. The mediational effect occurs when the relationship between exogenous variable-mediator and mediator-endgenous variable is
183
significant. Further, within this approach a direct relationship between exogenous and endgenous variables does not need to be included in the model, but it can be controlled (James & Brett, 1984; Kenny et al., 1998).
To examine the significance of the mediational effect, Sobel’s (1982) test and the bootstrapping test as outlined by Preacher and Hayes (2004) were adopted. Sobel’s (1982) test determines the significance of the mediational effect of the mediator (t-value) by testing the hypothesis of no difference between the total effect and the direct effect (Fairchild & MacKinnon, 2009). However, Sobel’s test suffers from low statistical power in small samples (Meyer & Baker, 2010). Recently, Preacher and Hayes (2004) have proposed a non- parametric bootstrapping method to overcome the limitation of Sobel’s test that derives confidence intervals for total and specific indirect effects. If the confidence intervals of mediational effect exclude zero, the mediational effect will be significant and different from zero (Preacher & Hayes, 2004).
Sections 5.6.1 and 5.6.2 present the Hypotheses results related to Stage A (Hypotheses 1 to 4) and Stage B (Hypotheses 5 to 8), respectively.
5.6.1. Stage A: Hypotheses 1 to 4
Hypotheses 1a, 1b, and 2 are related to the mediational effect of exploratory R&D, exploratory marketing, and their integration in the relationship between exploratory strategy and new product performance. Table 5.18 reveals that exploratory strategy significantly influenced exploratory R&D (β= 0.41, t-value= 5.05), exploratory marketing (β= 0.51, t- value= 6.33), and their integration (β= 0.60, t-value= 8.89). In addition, new product performance was predicted by exploratory R&D (β= 0.26, t-value= 2.31), exploratory marketing (β= 0.50, t-value= 5.11), and their integration (β= 0.60, t-value= 1.96). However, exploratory strategy had an insignificant direct effect on new product performance (β= -0.24,
184
t-value= 1.73, ρ > 0.05). Following the arguments of James and Brett (1984), the results indicate that exploratory R&D, exploratory marketing, and their integration fully mediate the effect of exploratory strategy on new product performance, supporting hypotheses 1a, 1b, and 2.
Hypotheses 3a, 3b, and 4 are related to mediational effect of exploitative R&D, exploitative marketing, and their integration in the relationship between exploitative strategy and established product performance. Table 5.19 reveals that exploitative strategy significantly influenced exploitative R&D (β= 0.28, t-value= 3.57), exploitative R&D (β= 0.45, t-value= 7.92), and their integration (β= 0.57, t-value= 3.93). In addition, established product performance was predicted by exploitative R&D (β= 0.45, t-value= 5.69) and exploitative marketing (β= 0.35, t-value= 2.72). However, the integration of exploitative R&D and exploitative marketing had an insignificant effect on established product performance (β= 0.14, t-value= 0.74, ρ > 0.05). Exploitative strategy also had an insignificant direct influence on established product performance (β= -0.10, t-value= 0.95, ρ > 0.05). Following James and Brett (1984), the results indicate that exploitative R&D and exploitative marketing fully mediate the effect of exploitative strategy on established product performance, supporting hypotheses 3a and 3b. However, the integration of these two capabilities does not mediate the effect of exploitative strategy on established product performance, rejecting hypothesis 4.
As shown in Table 5.22, Sobel’s (1982) test and the bootstrapping test provide evidence for significance of hypotheses 1 to 3, as the t-value of all mediational effects were greater than the cut-off value (± 1.96) and the confidence interval of all mediational effects did not include zero. This shows hypotheses 1a, 1b, 2, 3a, and 3b are significant.
185
Table 5.22- Sobel’s test and bootstrapping test for Hypotheses 1a, 1b, 2, 3a, and 3b
Hypothesis Mediational effect Effect
Normal theory test Bootstrapping SE t-value ρ SE LL UL
H1a Exr S→Exr R&D→ NPP .29** .05 5.25 .00 .07 .15 .43
H1b Exr S→Exr M→ NPP .11** .03 2.81 .00 .04 .04 .18
H2 Exr S→Exr R&D × Exr M→ NPP .36** .06 5.64 .00 .07 .21 .52
H3a Exi S→Exi R&D→ EPP .15** .04 3.39 .00 .04 .06 .26
H3b Exi S→Exi M→ EPP .13* .05 2.37 .02 .05 .03 .22
Notes: Notes: ExrS= Exploratory strategy, ExiS= Exploitative strategy, ExrR&D= Exploratory R&D, ExrM= Exploratory marketing, ExiR&D= Exploitative R&D, ExiM= Exploitative marketing, NPP= New product performance, EPP- Established product performance, UL= Upper-level, LL= Lower-level, * p <.05, ** p < .01.
5.6.2. Stage B: Hypotheses 5 to 8
Hypotheses 5a, 5b, 6a, and 6c are related to the mediational effects of marketing ambidexterity and R&D ambidexterity in the relationships between strategic ambidexterity, new product performance, and established product performance. Table 5.20 reveals that strategic ambidexterity significantly influenced marketing ambidexterity (β= 0.36, t-value= 3.83) and R&D ambidexterity (β= 0.40, t-value= 4.82). In addition, new product performance was significantly predicted by marketing ambidexterity (β= 0.23, t-value= 3.42) and R&D ambidexterity (β= 0.52, t-value= 6.25). However, strategic ambidexterity had an insignificant direct effect on new product performance (β= 0.18, t-value= 1.84, ρ > 0.05). Following James and Brett (1984), the results indicate that marketing ambidexterity and R&D ambidexterity fully mediate the effect of strategic ambidexterity on new product performance, supporting hypotheses 5a and 6a.
Hypotheses 7a, 7b, 8a, and 8c are related to the mediational effects of new product and established product performance in the relationships between marketing ambidexterity, R&D ambidexterity, and firm performance. Table 5.21 reveals that strategic ambidexterity significantly influenced marketing ambidexterity (β= 0.36, t-value= 4.84) and R&D ambidexterity (β= 0.49, t-value= 7.04). In addition, established product performance was significantly predicted by R&D ambidexterity (β= 0.25, t-value= 2.23). However, strategic
186
ambidexterity (β= .016, t-value= 1.82, ρ > 0.05) and marketing ambidexterity (β= -0.22, t- value= 1.13, ρ > 0.05) had insignificant effects on established product performance. Therefore, the results indicate that R&D ambidexterity fully mediates the effect of strategic ambidexterity on established product performance, supporting hypothesis 5b. Since marketing ambidexterity had an insignificant effect on established product performance, marketing ambidexterity does not mediate the effect of strategic ambidexterity on established product performance, rejecting hypothesis 6b.
Regarding the mediational effect of new product performance, Table 5.21 indicates that marketing ambidexterity and R&D ambidexterity significantly influenced new product performance, while new product performance also significantly affected firm performance (β= 0.45, t-value= 9.28). However, marketing ambidexterity (β= 0.05, t-value= .47, ρ > 0.05) and R&D ambidexterity (β= .19, t-value= 1.50, ρ > 0.05) had insignificant direct effects on firm performance. Therefore, the results indicate that new product performance fully mediates relationships between marketing ambidexterity, R&D ambidexterity, and firm performance, supporting hypotheses 7a and 7b.
Regarding the mediational effect of established product performance, Table 5.22 shows that R&D ambidexterity significantly influenced established product performance, while established product performance also significantly affected firm performance (β= 0.43, t- value= 7.70). However, R&D ambidexterity (β= 0.09, t-value= 1.18, ρ > 0.05) had an insignificant direct effect on firm performance. Therefore, the results indicate that established product performance fully mediates the relationship between R&D ambidexterity and firm performance, supporting hypothesis 8a. Since marketing ambidexterity has an insignificant effect on established product perofrmance, established product performance does not mediate the relationship between marketing ambidexterity and firm perofrmance, rejecting hypothesis 8b.