*2.6. Statistical Methods*

After collecting the required data, Cronbach's alpha (a measure of internal consistency) was used to check how closely related a set of items (questions) are. Cronbach's alpha, which ensures that either the response given from each respondent was consistent or not, was used to validate and ensure that the data is reliable. Based on the collected data, qualitative and quantitative data analysis techniques were employed. In undertaking the quantitative analysis, a regression model was used to investigate the extent to which independent variables affect the dependent variable. The correlations among the independent variables were also investigated to see which variable has more effects. Data collected from interviews, open-ended questionnaires, observations, and focus group discussions were described qualitatively.

In order to make sure that the collected data are correct, consistent, and useful for its accuracy and reliability against respondents, the data were verified through Cronbach's alpha in SPSS. The alpha coefficient for this study is 0.883, suggesting that the items have relatively high internal consistency. A reliability coefficient of 0.70 or higher is considered "acceptable" in most research studies. After ensuring that the data reliability was correct and certain, Equation (1) was used to calculate the overall organizational achievement of the cadastral system. The equation represents the influence of independent variables *X* on the dependent variable *Y*, which is a fundamental concept in ordinal regression.

$$Y = \beta\_0 + \beta\_1 X\_1 W\_1 + \beta\_2 X\_2 W\_2 + \dots + \beta\_n X\_n W\_n \tag{1}$$

where *Y* is the dependent variable (overall organizational performance), β<sup>0</sup> is the intercept, β*<sup>n</sup>* is coefficients (estimates), *Xn* is the mean value of independent variables, and *n* refers to the number of independent variables, which in this case is 8. The intercept for ordinal regression model is zero since it starts from the origin.

Table 1 shows basic descriptive statistical information about the response statistics. The minimum and maximum numbers are bounded between 1 and 5 to represent level of satisfaction. The Mean describes the average of the responses; median explains the value separating the higher half from the lower half of a response, while the mode explains the number that appears most frequently.


**Table 1.** Response statistics of the questionnaire (*n* = 150).

Assumptions of Ordinal logistic regression model: ordinal logistic regression (often just called "ordinal regression") is used to predict an ordinal dependent variable given one or more independent variables. There are four assumptions to validate this model. (1) Dependent variables should be measured at the ordinal level; (2) independent variables must be treated as either continuous or categorical, they cannot be treated as ordinal variables; (3) two or more independent variables that should not be highly correlated with each other (no multicollinearity); (4) each independent variable has an identical effect at each cumulative split of the ordinal dependent variable.

Regression model fitting information: Since the collected data from questionnaire are in the form of order (rank) through Likert scale, ordinal regression model was performed to extract meaningful information. This type of regression model has five conditions to be fulfilled.

Model fitting: this is the measure of how well the model fits the data. The significance level of alpha is 0.05, which limits the level of significance value. The result from this model is 0.00 (See Table 2), which is less than the common alpha level of 0.05, which indicates that it is statistically significant, telling that the model gives better predictions. The statistical significance indicates that changes in the independent variables (enablers) correlate with shifts in the dependent variable (organizational performance).

Goodness of fit: The goodness of fit of a statistical model describes how well it fits a set of observations. Measures of goodness of fit typically summarize the discrepancy between observed values and the values expected under the model in question. The significance value for the goodness of fit (Pearson) is 1 (see Table 2), which is greater than the common alpha level of 0.05, which indicates that it is statistically significant and suggests that the model fits the data very well.



The null hypothesis states that the location parameters (slope coefficients) are the same across response categories. <sup>a</sup> Link function: Logit. <sup>b</sup> The log-likelihood value cannot be further increased after maximum number of step-halving. <sup>c</sup> The Chi-Square statistic is computed based on the log-likelihood value of the last iteration of the general model. Validity of the test is uncertain.

Pseudo R2: This statistic indicates the percentage of the variance in the dependent variable (organizational performance) that the independent variables (enablers) explain collectively. If R2 (Nagelkerke) is greater than 0.7, it indicates that 70% of the independent variables explain the dependent variable, which in this case is 0.719.

Test of parallel line: this is the test according to the assumption of proportional odds. This is a key assumption in ordinal regression. The assumption is that the effects of any explanatory variables are consistent (proportional) across the different thresholds (by thresholds we mean the splits between each pair of categories of your ordinal outcome variable). In other words, that the explanatory variables have the same effect on the odds regardless of the threshold.
