Research Methodology | Trust in Knowledge Management and Systems in Organizations

< Day Day Up >

Sample Subjects

This research was conducted by using IS managers, including data processing managers, operation managers and check processing managers in American and Korean banks and their branches located in the United States. Those banks are using computer-based information systems in order to process and to perform their daily transaction processing function such as automated teller machines (ATMs), loan decision making systems and check processing systems and other information resources processing. Numerous researchers (Miller & Doyle, 1987; Scott, 1988; Singleton, McLean, & Altman, 1988) have used the banking industry as an information systems research institution in the literature. The banking industry can uniquely represent American IS managers and Korean IS expatriate managers in which this research attempts to study their goals and objectives. Sampled banks and IS managers were chosen from Thomson Bank Directory and The Rand McNally Bankers Directory. The sampled banks were selected from seven states: California, Washington, Florida, Pennsylvania, Illinois, Texas, and New York. These states are the ones that have Korean banks and American banks as well. Based on the number of banks available in the United States, the numbers of sampled banks were determined. In particular, every fifth American bank in the Thomson Bank Directory and The Rand McNally Bankers Directory were selected. On the other hand, for Korean banks, all the banks in the directory were selected. Number of IS managers selected from each bank was different since all the IS managers including operations, check processing and data processing managers were chosen. As a result, 567 American IS managers, and 126 Korean IS expatriate managers were selected respectively from those seven states.

Research Instrument

The research instrument includes a cover letter and a questionnaire. The research instrument consists of a questionnaire of 35 questions. As suggested by Kerlinger (1973), more than 10 times the number of items in the questionnaire, 945 questionnaires were sent to the selected subjects. As shown in Table 2, 19 predictors used in this study were as follows:

Organizationally mediated (eight questions) variables such as communications, job-related stress and participation in decision-making.
Interpersonally mediated (five questions) variables are job variety, and interpersonal relations.
Internally or personality mediated (six questions) attributes are independence and achievement.

Table 2: Measures Used in the Study.
Variables	Item Numbers in the Questionnaire	Sources
Predictor Variables:
Organizationally Mediated:		Sekaran & Martin (1982)
Participating in decision-making	10, 11	White & Ruh (1973)
Communication	12	Price (1972)
Sound company policies	25, 27, 28	Kanungo & Wright (1983); Hofstede (1983)
Self-esteem from job setting	1	Quinn & Shepard (1974)
Job-related stress	15	Lyon (1971)
Interpersonally Mediated:		Sekaran and Martin (1982)
Job variety	26	Sims, Szilagyi, & Keller (1976)
Interpersonal relations	9, 13	Kanungo & Wright (1983)
Respect and independence	4, 22	Kanungo & Wright (1983)
Internally Mediated:		Sekaran & Martin (1982)
Responsibility and independence	2, 3, 5	Steers & Braunstein (1976)
Achievement	3, 7, 8	Steers & Braunstein (1976)
Criterion Variables:		Sekaran (1983)
Values	6, 13, 14, 17, 28	Hofstede (1980, 1984)
Job satisfaction (Pay, promotion, coworkers, supervisor, work)	13, 16, 21, 23, 24	Smith, Kendall, & Hulin (1969)
Performances (Self-rated performance)	18, 19, 20	Porter & Lawler (1968)

On the other hand, the criterion variables of value, job satisfaction and performance were measured by five, five and three questions, respectively. The questionnaire has two parts. In Part I, each question was answered using a 5 point-scale ranging from 1 (Of Very Little Importance: VLI) to 5 (Of Utmost Importance: UI). This questionnaire also contains three questions to evaluate self-productivity, self-performance, and self-effort of IS managers. The purpose of three self-reported-ratings in the questionnaire was to examine the predictive validity. In Part II, there were seven demographic questions such as number of years working in the organization and in the current position, sex, education, age, nationality, and total assets of bank. The seven demographic variables were used to measure the devotion of IS managers and loyalty toward their organization. In addition, Table 2 shows the relationships between the variables and their item numbers in the questionnaire with their sources in the literature.

Methodology of Data Analysis

The data analysis has purposes as follows: to evaluate the objectives in conjunction with the values, job satisfaction and performances of the groups of two national IS managers' activities; to evaluate and analyze the psychometric properties of the instrument; and to examine the adequacy for cross-cultural studies from those IS managers. Thus, this research followed three phases. In Phase 1 of descriptive statistics, Korean IS expatriate and American IS managers' values, job satisfaction and performances are measured by the 28 questions with a 5-point Likert scale, ranging from 1 (Of Very Little Importance: VLI) to 5 (Of Utmost Importance: UI). Since Montazemi (1988) argued that unweighted scores were validated as reliable results and highly correlated to weighted scores, the unweighted scores were used in this study. The IS managers' values were measured by eight questions of "organizationally mediated" variables. In addition, IS managers' values were measured using five questions of "interpersonally mediated" variables and six questions of "internally mediated" variables. Job satisfaction also used the same three groups of mediated variables. Finally, performance was measured based on effort, productivity, and performance. Those three criterion variables were measured by their own self-evaluated variables. This research then compares the results with those of the independent variables which support them for the two groups of national IS managers. In Phase 2, the reliability and construct validity of the instrument were tested by utilizing Cronbach's α tests, Pearson's product-moment correlation coefficient analysis and factor analysis (Churchhill, 1979; Cohen, 1977; Cronbach, 1951). Content validity was tested through an exhaustive literature research to include representative questions for this research design. After developing the research instrument, it was pretested by 20 faculty and staff members to follow the suggestions made by Hunt, Sparkman, and Wilcox (1982). In addition to that, the instrument was corrected and edited according to the suggestions from the respondents of the pretest. The coefficient of reliability tells whether the instrument is consistent in what it does. The cut-off point of Cronbach's α ranges from 0.60 to 0.80 for this research analysis (Churchhill, 1979; Cohen, 1977; Cronbach, 1951; Ives & Olson, 1984; Nunnally, 1978). When the alpha is below the range of 0.60 to 0.80 out of 1.00, the corresponding item(s) was deleted from the instrument. After the Cronbach's α test, a factor analysis was tested with those items which are reliable variables. Coefficient of validity tells whether the instrument is doing what a researcher believed it was doing (Montazemi, 1988). In doing the construct validity, this research used the cut-off point at 0.45. For the content validity, internal consistency of the instrument would be tested by examining the inter-item correlations coefficient at 0.05 level (Doll & Torkzadeh, 1988; Doyle & Becker, 1983). In order to keep the variables in the questionnaire the instrument should have passed the basis based on the results of the validity and reliability tests. Montazemi (1988) also suggested that to measure the homogeneity of the questionnaire, calculate the correlation coefficients for each pair of scales for each factor. If the coefficients are significant at p-value > 0.0001 level, these coefficients indicate that the questionnaire is internally consistent. In addition, to measure the predictive validity or criterion-related validity of the questionnaire, three self-rated performance criterion variables were utilized.

In Phase 3 of regression analysis and ANOVA Test, this study also examined correlation of values, satisfaction and performance between the two national IS managers. The correlation coefficient of the variables that measure the values with the dependent variables was measured and the suggested cut-off point (Dunn, Norburn, & Birley, 1985) of each coefficient is equal to or greater than 0.3. For regression analysis, separate regression equations were used between each dependent variable (value, job satisfaction and performance) with the independent variables. All predictor variables were regressed without a predetermined order, thereby variables were included sequentially into the equation based on the respective contribution of each predictor to explained variance in the dependent variables. This regression procedure does not assume a definite causal ordering among the variables, and therefore would identify the most powerful independent variables of value, job satisfaction and performance. As a result, a stepwise regression was calculated and used to show the unique marginal contribution of criterion variables in terms of change in R². Besides, to measure criterion-related validity, cross-validated regression analysis procedures were employed. This technique was used to provide an average measure of the degree of association (R²) between a set of predictor variables and the criterion measures within each culture. In addition, ANOVA analysis was conducted. Through the multivariate analysis of variance, dependent variables, job satisfaction and performances of the two national IS managers were analyzed. The differences and similarities based on the convergence hypothesis and divergence hypothesis were examined.

< Day Day Up >