*5.4. Confirmatory Factor Analysis, CFA*

The CFA modeling results on the data set produced from the EFA are shown in Figure 1. Due to low standardized regression weights: SM21 and SM12 model fit measures present unacceptable indices (NFI = 0.828, CFI = 0.910, SRMR = 0.0737, RMSEA = 0.089, and PClose = 0.006). For that reason, SM8, SM21 and SM12 from the cultural and enforced choice construct were dropped from this CFA assessment.

The final model fit measures present acceptable indices (NFI = 0.851, CFI = 0.933, SRMR = 0.079, RMSEA = 0.04, PClose = 0.0694) with the Chi-Square test of 1.66. The results from this CFA assessment show that all indices are at acceptable levels. We also observed results looking for standardized residuals with an absolute value greater than 3.00, which are usually alarming as potential threats to one-dimensionality. Therefore, we took into account that large residuals related to specified items in CFA are indicators of a model's inability to adequately explain the relationships in the model (Hair et al. 2010). By checking the standardized residuals, all absolute values are less than 3.00. The highest absolute value is 2.64, while the majority values are between +1.50 and −1.50.

**Figure 1.** CFA model output from AMOS 22.

All parametric estimates were statistically significant at a level of 0.05 or better. Furthermore, both samples' latent variables achieved an acceptable level of Composite Reliability (>0.6). Convergent validity takes two measures that are supposed to be measuring the same construct and shows that they are related. The convergent validity has been tested with the AVE value, with four factors, I.P., SI, S.F. and S.C. AVE values are all above the 0.5 thresholds. Conversely, discriminant validity shows that two measures that are not supposed to be related are, in fact, unrelated. The square roots of the AVE values are 0.71, 0.71, 0.84 and 0.84, which is higher than the correlation estimate for each pair of constructs. Therefore, this study can conclude that all items met convergent and discriminant validity.

#### *5.5. Structural Equation Modeling, SEM*

At this point of our analysis, we have obtained a set of items that measure four factors of the strategic process. This final set of constructs and items was used to test the study's research model hypotheses, which were derived based on previous theoretical research:

**Hypothesis 1 (H1).** *Institutional pressures are negatively associated with strategic intent.*

**Hypothesis 2 (H2).** *Institutional pressures are negatively associated with the strategic formulation.*

**Hypothesis 3 (H3).** *Strategic intent is positively associated with strategic formulation.*

Moran and Brightman (2001) define change management as "the process of continually renewing an organization's direction, structure, and capabilities to serve the ever-changing needs of external and internal customers.", which, by association, extends to the evolving nature of the strategic management process. Strategic change comprises strategic evaluation and control, performance evaluation, exercising control and recreating strategies. The strategic planning process, once strategic formulation is set up and strategy is implemented, subsequently leads to strategic control and evaluation. In Bailey et al.'s (2000) model, the element of incremental dimension is highly associated with the concept of strategic change. It is expected that strategic intent and formulation will impact strategy-driven change. Since strategic intent is usually the ultimate goal of an organization, it should be more stable and will not change easily as a part of the strategic planning phase, but it will positively influence change. Therefore, the following hypotheses, shown in Figure 1, are proposed.

**Hypothesis 4 (H4).** *Strategic intent is positively associated with change.*

**Hypothesis 5 (H5).** *Strategic formulation is positively associated with change*.

**Hypothesis 6 (H6).** *Strategic formulation mediates the relationship between strategic intent and change.*

SEM is chosen because of its ability to examine the entire research model as a whole, and not only the relationships between separate variables (Schumacker and Lomax 2010). Regarding the normality of the variables, Skewness and Kurtosis values of composite observed variables were examined using SPSS 22, and overall, based on the results, no large discrepancies from normal data were observed.

We would like to note that structural equation models, especially maximum likelihood estimations, are relatively robust about modest departures from a normal distribution (Diamantopoulos and Siguaw 2000). Since Likert scales were used in the survey, it is reasonable to accept that a continuous variable underlies each measurement scale. Most of the hypothesized relationships in SEM are tested through linear relationships. Therefore, it makes sense that there is a linearity assumption regarding the variables used in this study, because there is no evidence to suggest otherwise.

Attributable to the complexity of the relationships represented by the hypotheses, SEM assessments of the relationships are conducted in two phases. First, we test our proposed research model hypotheses. Then, we examine the mediation effects. An assessment of the model is conducted and presented in Figure 2 and Table 7.

**Figure 2.** SEM path analysis results.


**Table 7.** Hypotheses testing results.

Note: \*\*\* *p* < 0.01.

With the exception of H1, the other four hypotheses, H2, H3, H4 and H5, are strongly supported. H1 has a *p*-value of 0.172, which is not acceptable. H2, H3, H4 and H5 are strongly supported, with a *p*-value < 0.01. These results indicate that institutional pressures may not have a negative influence on strategic intent and may entail other secondary effects that moderate this relationship. Such effects may be the individual motivation that may persist despite institutional pressures. However, results show that S.F. is affected by I.P. (negatively) and S.I. (positively). In conjunction with H1, we may interpret these results, as S.I. does not mediate the influence of I.P. on S.F., despite direct influences. What is interesting to note is that the relationship coefficient of H2 and H3 is very close in strength but opposite in direction, which begs the question of whether the effects of I.P. on S.I. are equally and inversely proportional, and in which way? It seems that there are underlying complex (maybe behavioral, psychological, etc.) mechanisms describing the relationship between I.P. and S.I. which are not captured in our model.

The model fit indices are scrutinized. The criterion of the Chi-square test for acceptance varies across researchers, ranging from less than 2.00 (Ullman 2001) to less than 5.00 (Schumacker and Lomax 2004). In the Chi-square (χ2) test of S.M., the ratio is over three, but it is on the edge of the acceptable range (<5.00). All fit indices showed that the model fitting the data is acceptable. Since standardized RMR is within an acceptable threshold, that compensates RMSEA. Therefore, we determine that the research model fit indices are acceptable, suggesting that this model is suitable for further hypothesis testing and model development.

The paths were assessed through standardized estimates and associated *p*-values. Given that all of the hypothesized relationships in the model were one-directional, all critical *p*-values are shown significance, except that the relation between SM1F and SM3F showed an insignificant effect. The probability of getting a critical ratio as large as 2.237 in absolute value is 0.025. In other words, the regression weight for SM1F in the prediction of SM3F is significantly different from zero at the 0.05 level (two tailed), while all others in absolute value are less than 0.001, which means significantly different from zero at the 0.001 level (two tailed). We summarize the hypotheses' results below.

#### **Hypothesis H1 (I.P. negatively influences S.I.).** *Not Supported.*

The influence is negative, but not significant. This means that the institutional pressures that the Secretariat is facing have not negatively influenced its strategic intents. Considering strategic intentions are more stable and have not been easily changed, institutional pressures may not influence much strategic intent.

#### **Hypothesis H2 (I.P. negatively influences S.F.).** *Supported.*

Based on various IGO reports and observations from member states, political and power influence play a major role in the IGO strategic management process, which also influences the strategic formulation phase. The relationship is to test how significant the effect is. From the results of this independent model, it shows that the relationship does significantly exist in the IGO context.

#### **Hypothesis H3 (S.I. positively influences S.F.).** *Supported.*

In the literature, it is commonly agreed that strategic management comprises three management components, namely strategic intent, strategic formulation and strategic implementation. The results from this independent model confirm that the strategic intent factor significantly influences strategic formulation in the IGO context as well.

#### **Hypothesis H4 (S.I. positively influences S.C.).** *Supported.*

Strategic intent sets up the direction for strategic change and will positively impact the possibility of strategic change. The results from this independent model show that the relationship is significant.

#### **Hypothesis H5 (S.F. positively influences S.C.).** *Supported.*

#### **6. Discussion and Conclusions**

This study proposed a conceptual model that integrated DiMaggio's institutional pressures framework and Bailey et al.'s (2000) multidimensional strategy model to examine critical factors influencing the effectiveness and efficiency of strategic management in IGO organizations. First, we examined six dimensions of an organization management model, including external and internal forces as they relate to an IGO context, in which command, political, organizational, cultural and enforced choice elements in Barley's model are associated with institutional coercive, normative and mimetic pressures. Planning elements refer to strategic intent and strategic formulation, and incremental dimension refers to strategic change. The conceptual model posited six hypotheses, which we analyzed using EFA, CFA and SEM. Through a factor reduction process, the strategic management model instrument was optimized by around 50%. This new model promises an efficient and effective explanation of factors and influences on strategic organizational management.
