Data Collection

Along with the study of the concept of bank loyalty in relation to customer satisfaction and confidence, in this paper we intend to explain the reasons why customers of financial institutions may drop from one institution to another in order to meet their needs of credit and savings.

For that, the dependent variable in the statis­tical model we propose is dichotomous, with a value of 0 for customers who have finished their savings and credit activities with the institution under study and 1 for customers who are still customers of the institution. For both situations a maximum period of 18 months is considered since the collection of information in the satisfaction survey, not exceeding the limit of two years used by previous studies (Stewart, 1998; Malhotra et al., 2008; Santiago, 2008).

The data provided is derived from the general activity satisfaction survey that Caja Rural of Granada³ makes to customers every other year, and it was carried out between September and October 2009.

The Web form consisted of a total of 10 Iikert questions with five graduations and was uploaded on the servers of Caja Rural of Granada Computer Service. The survey was limited to the assessment of a number of items, which we gather in Table 2 of the present chapter.

As indicated in Table 3, the final sample consisting of 946 individuals was randomly divided into two subsamples with the intention of performing a post-validation on the estimated logistic regression model by assigning 75% of the total cases to estimate the statistical model and the remaining 25% to the validation of the same (Bensic et al., 2006).

Table 2. Website survey developed for e-Banking customer of Caja Rural of Granada

Table 3. Random sample data

selection was 3.19% in the proportion estimate (with a confidence level of 95% and P = Q = 0.5).

On the other hand, Table 4 shows the variables separate from the issue we want to explain. The database includes information on economic and socio-demographic characteristics of the custom­ers of the entity, as well as qualitative variables drawn from the above survey. The latter are known as categorical variables with 5 possibilities or states expressed by customers, with 1 indicating strongly disagree and 5 strongly agree.