بررسی برخی عوامل اجتماعی مؤثر بر فساد اداری (مورد مطالعه: سازمان‌های اداری شهر یزد)

Introduction
Administrative corruption is a complicated and multi-faceted phenomenon that has many causes and ramifications, which manifest themselves in different forms in various countries and appear to be endemic of all social systems where authority is delegated. Although there might be different definitions of and approaches towards administrative corruption, the definition that has been put forward by the World Bank and Transparency International appears to be more general and comprehensive. It defines corruption as the abuse of public power, office, or resources by public or elected governmental officials for personal gain, e.g. by extortion, soliciting or offering bribes. Recent research findings indicate that organizational commitment and work satisfaction on the part of employees could be an effective antidote to administrative corruption. On the other hand, employee sense of equity and fairness also has been shown to be effective in enhancing organizational commitment. Last but not least, employee financial needs could play a significant part in their vulnerability to corruption.  Therefore, the main aim of this paper is as follows: by utilizing the ideas of Selznick on organizational commitment, Merton's views on work satisfaction, Taylor's and Schwartz's ideas on financial needs, and Barnard's views on managerial capabilities, the relationships of these variables on administrative corruption were studied. The other purpose of the paper is to examine the effects of some other related variables like age, sex, level of education, job seniority, and marital status on administrative corruption.
 



Material and Methods
The research method was quantitative, data collection technique was survey, and the tool of data collection was questionnaire. The research population of the study consisted of all 44116 employees of public organizations of the city of Yazd. Using a stratified sampling technique, 16 administrative organizations were selected. To optimize the generalizability of the findings, p (the relevant characteristic in the population, which in this study relates to the possibility of corruption) was considered to be 0.50. Then, with the aid of systematic random sampling, 400 employees were selected as research sample. Based on relevant theoretical orientations and the results of empirical research, the following variables were regarded as independent variable: (1) Employee organizational commitment, which refers to the extent that employees believe in organizational goals and values, and the extent to which they are willing to stay with the organization and help it in realizing those goals. Seven Likert-type questions were used to measure this variable. Cronbach alpha coefficient for reliability was 0.89.
(2) Employee work satisfaction, which refers to employees' sense of job security and contentedness with the organization, with his/her coworkers, and how much he/she felt that there are opportunities for progress in his/her job and the organization. Five Likert-type questions were used to measure this variable. Cronbach alpha coefficient for reliability was 0.84.
(3) Managerial capabilities, which refers to the extent of management's power and abilities to run the organization in an optimum level as regards to coordination, planning, leading, and controlling employee activity towards the attainment of organizational goals.  Six Likert-type questions were used to measure this variable. Cronbach alpha coefficient for reliability was 0.85.
(4) Employee financial status that referred to employee salary and whether they felt that their salary and fringe benefits were fair and suitable to have a relatively comfortable life.  Nine Likert-type questions were used to measure this variable. Cronbach alpha coefficient for reliability was 0.88.
The dependent variable was administrative corruption and it referred to situations where organizational officials and employees felt that the use of their official status for personal gain was possible and justifiable. For example, respondents were asked to rate their following question: "since officials should be able to use their time and expertise in much needed places, they should not have to wait in queues like everyone else to get their usual chores done". Nine Likert-type questions were used to measure this variable. Cronbach alpha coefficient for reliability was 0.90.
Analysis of variance, the T test, and Pearson correlation coefficients were used to test the relationships among variables, and to test the theoretical model, regression analysis and the Lisrel software were utilized.
 
Discussion of Results and Conclusions
Research findings showed that the average amount of corruption  33.78 was higher than the expected average of 27. The amount of corruption differed by educational level, that is, as educational level increased the average amount of corruption decreased. It could be said that since jobs requiring higher levels of education also pay higher levels of wages and salaries, therefore their holders may not be as vulnerable to corruption as their counterparts with lower educational levels. The average amount of corruption for singles was 36.89 and that of married individuals was 33.19, therefore it could be said that marital status and corruption are related. Although it could be argued that since married people have more life obligations and therefore have higher expenses, thus being more vulnerable to financial corruptions, but the results showed that their average corruption levels are actually lower than singles. The higher levels of corruption among singles could be due to the fact that singles are less experienced and/or may have lower salaries, since in all probabilities; they were recently hired and are at lower levels of job rankings. It could also be argued that since singles have many more needs such as houses, automobiles, etc. they may be more vulnerable to corruption. No association was found between corruption and years of employment. This is actually interesting, since one may suspect that experience and tenure on the job could positively or negatively affect vulnerability to corruption and it is something that needs further investigation. The average amount of corruption for men was 33.47 and that of women was 34.22, and the results of the t-test showed that this difference was not statistically significant. This may seem odd because women's participation in out-of-home jobs in Iran is relatively recent and is not very significant- as far as numbers are concerned- and other studies have not found significant amounts of corruption among them. Further research in this area could be illuminating.  The results of regression analysis indicated that organizational commitment (r= -0.61) and financial needs (r= 0.45) were the most important factors affecting corruption. The direction of these relationships is of importance. As one's commitment to the organization increases, his/her chances of committing corruption decreases. This means that organizational authorities should pay enormous attention to the ways and mechanisms that enhance employee commitment. Also, the direct association between financial needs and corruption also needs to be taken seriously, because it shows that as financial needs go up, so does the chances of committing corruption. There was a reverse association between job satisfaction and corruption (r= -0.35). This, like organizational commitment, seems reasonable and needs to be taken very seriously. Organizational authorities should be very keen to identify the factors that may lead to job dissatisfaction among employees and find ways to increase their job satisfaction. Managerial capabilities had the least amount of impact (r= 0.10). This is somewhat strange, since it appears that managerial capabilities should play a significant role in this area and other studies have found that managerial styles and capabilities do have an effect. Further research in this area can of course be very important. The coefficient of determination was 0.57 which, although is an acceptable level of goodness of fit, nevertheless, shows that all four independent variables could explain only about 57 percent of the variability of the dependent variable