The impact of intrinsic and extrinsic religiosity on ethical decision-making in management in a non-Western and highly religious country

Abstract

The primary purpose of this study was to explore the indirect effect of intrinsic religiosity and extrinsic religiosity on ethical intention through ethical judgment. A review of the literature shows the need for more research at the intersection of religiosity and ethics, especially in non-Western, highly religious contexts. This research, therefore, addresses the research question: Do intrinsic religiosity and extrinsic religiosity indirectly impact ethical intention through influencing the ethical judgment of management professionals? Data were gathered from members of the Management Association of Pakistan through a questionnaire. Pearson correlation results show the overall trend between the constructs of interest. Multiple regression results show that both intrinsic religiosity and extrinsic religiosity are significant positive predictors of ethical judgment. Ethical judgment was also found to be a significant, positive predictor of ethical intention. The main contribution of the study is evidence that ethical judgment acts as a mediator between religiosity (whether intrinsic or extrinsic) and ethical intention in a non-Western highly religious context. This research also found that intrinsic religiosity impacts ethical intention directly as well as indirectly through ethical judgment, but extrinsic religiosity influences ethical intention only through its effect on ethical judgment. We discuss our results along with practical and research implications, and limitations of this research are highlighted to guide future research.

Appendices

Appendix 1

Scenarios

Scenario 1

The pricing committee of a large video game marketer suggests that prices be raised 20–30% during the holiday season and that a highly popular game be kept in short supply.

Action: The vice president of marketing decides that this is a good idea since consumer demand indicates that consumers will be likely to pay the higher prices.

Scenario 2

XYZ Company has been barred from entering the market in a large Asian country by collusive efforts of the local bicycle manufacturers. XYZ could expect to earn 500 million rupees per year from sales if it could penetrate the market. Last week, a businessman from the country contacted the president of XYZ and stated that he could smooth the way for the company to sell in his country for a price of Rs. 50,000,000 (50 million).

Action: The president of XYZ pays the Rs. 50,000,000 (50 million).

Scenario 3

Mr. A has recently accepted a job with a new, vigorous microcomputer manufacturer. Microcomputer manufacturers are engaged in intense competition to become the market leader with a software package which utilizes the English language and thus is easily used by the average consumer. Mr. A’s former employer is generally known to be the leader in this software development. When Mr. A was hired, he was led to believe that his selection was based upon his management potential. The morning beginning the third week on the new job, Mr. A received the following memo from the president:

Please meet with me tomorrow at 8:15 AM for the purpose of discussing the developments your former employer has made in microcomputer software.

Action: Mr. A reveals the new product developments made by his former employer.

Scenario 4

Ted Jones, senior editor of J & P Publishing Company, has just received a manuscript from one of his most successful authors. It provides the most authoritative account yet published of the history of the development of the atomic bomb. However, the final chapter contains a detailed description of how the bomb is made. Jones has tried to convince the author to omit the last chapter stating that such information should not be made readily available to the mass market in paperback form. The author believes the chapter is critical to the success of the book and thus will not agree to its deletion.

Action: Ted Jones publishes the book.

Note: All the responses to the following items were sought on a Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree).

Ethical judgment and ethical intention items

For each of the four scenarios above, managers respondents were asked right after each scenario:

• I consider the action taken to be ethical. (Ethical judgment item)

• I would be likely to take the same action in this situation. (Ethical intention item)

Intrinsic religiosity

I enjoy reading about my religion.

It doesn’t matter what I believe in as long as I am fair in my dealings. (Reverse coded)

It is important for me to spend time in private worship, for example Salaat or prayer.

I always have a strong sense of God’s presence.

I try to live all aspects of my life according to my religious beliefs.

My whole approach to life is based on my religion.

I believe in my religion but many other things are more important in life. (Reverse coded)

Extrinsic religiosity

I go to religious service (for example, Jumma/Sunday service) mainly because it helps me to make friends.*

Prayer is for peace.

Prayer is for happiness.

I go to religious services (for example, Jumma/Sunday service) mostly to spend time with my friends.*

I go to religious services (for example, Jumma/Sunday service) mainly because I enjoy seeing people whom I know.*

What religion offers me the most is comfort in times of trouble/sorrow.

I pray to gain relief and protection.

*Deleted items

Appendix 2

Section A: data processing and assumption testing of parametric data

The responses from the usable questionnaires were initially input into an Excel sheet. In SPSS 21, appropriate variable names along with their titles were defined, specifying their data types and sets of permissible values. The data was then entered into the data view of SPSS. Since the scenarios each contain an ethically questionable action by a manager, the items essentially measure (un)ethical judgment and (un)ethical intention. Each item’s score, as provided by the respondent, was reverse coded by subtracting each item score from the highest permissible value + 1. For instance, if an item response is 2, recoding it produced 6 on a scale from 1 to 7 achieved through 7 + 1 − 2 = 6. Missing value analysis revealed the following pattern.

Table 8 indicates that missing values constitute only 0.09% of all the values; thus dictating no need to compensate for them. The missing values analysis proceeded by calculating the average score for each construct. Correlation, multiple regression, and mediation analysis were then run on these average scores only.

Table 8 Missing values analysis

In order to perform parametric tests on the data collected, a multi-pronged approach was applied whereby visual tools (P-P plots) along with normality tests (using 1-sample Kolmogorov-Smirnov tests [K-S]) were undertaken to check for the normality of the data. This procedure revealed that the outcome, ethical intention, D(109) = 1.20, p > 0.05, is significantly normal. Ethical judgment, D(109) = 1.25, p > 0.05, and intrinsic religiosity, D(109) = 1.09, p > 0.05, all appeared significantly normal. Extrinsic religiosity, however, D(109) = 2.04, p < 0.001, showed nonnormality, probably owing to the sensitivity of 1-sample K-S, but the normal P-P plot of extrinsic religiosity clearly showed normality. For brevity, we have not shown the normal P-P plots for the variables of the study. The individual values, however, converge to the normal line that pictorially depicts that the data is normal in all cases. This observation is also in agreement with the 1-sample K-S statistic described above except for extrinsic religiosity.

Assumption testing for generalizing multiple regression model

In order to generalize the findings of the sample to the population, the regression models should meet certain assumptions. The subsequent discussion will highlight how the data of this study show these assumptions are met. After performing case-wise diagnostics, the assumptions of nonzero variance, normally distributed data, no perfect multicollinearity, homoscedasticity, independent errors, normally distributed errors, independence, and linearity were checked.

Outlier treatment. The case-wise diagnostics were performed by computing the standardized residual value, Cook’s distance, average leverage value, and Mahalanobis distance for identifying outliers. Based on these statistics, no case unduly influences the regression model, and thus, no case was identified as an outlier.

Nonzero variance. All the predictors have nonzero variance, as can be seen through their mean and standard deviation values. All the predictors and the outcome variable follow a normal distribution, as discussed at the start of this appendix.

Multicollinearity. In the case of multiple regression, all the tolerance statistics—0.63, 0.67, and 0.93 in case of ethical judgment, intrinsic religiosity, and extrinsic religiosity respectively—were well above 0.2, indicating that there was no severe problem of multicollinearity as Menard (1995) suggests that only values below 0.2 are of concern.

Homoscedasticity. This assumption requires that at each level of the predictor variable, the variance of the residual terms should be constant. The scatterplot for each model shows that all the values are evenly distributed around 0, so the homoscedasticity assumption holds.

Independent errors. This assumption requires that residuals be uncorrelated. This assumption can be tested using the Durbin-Watson (D-W) test statistic that should produce a result between 1 and 3. In the case of multiple regression model, the D-W value was 1.5, showing that the independence of errors can be assumed.

Normally distributed errors. This assumption requires that the residuals in the model be random and normally distributed with a mean of 0. To check this assumption, standardized residuals were generated when running the regression model, and the residuals were found to form a normal distribution, as shown in the P-P plots. Also, the 1-sample K-S test statistic was nonsignificant, p > 0.05, thus showing residuals to be normally distributed in all models.

Independence. This assumption requires that the values of the outcome variable are independent. Each value comes from a separate entity, the respondent, so independence can be assumed owing to the correlation design of the study.

Linearity. The scatter plots depict linear relationships between all the predictors and ethical intention, the outcome variable.

Section B: factor analysis assumption testing

Table 9 Factor analysis assumption testing