Please note: Ernest O’Boyle and I completed a lengthy review of the original EPQ (Forsyth, 1980), and based on that review, we recommend researchers now use a shortened version of the original instrument: The EPQ-5. The instrument is described in detail in this open source paper at PLOS ONE:
O’Boyle, E. H., & Forsyth, D. R. (2021). Individual differences in ethics positions: The EPQ-5. PloS one, 16(6), e0251989.
I developed the Ethics Position Questionnaire to measure individual differences in moral thought, prompted in part by curiosity about the diverse reactions to one of my favorite social psychological studies: Milgram’s (1963) classic studies of obedience to authority. Milgram (1964), in defending his work, noted the wide range of opinions on the morality of his methods. Some critics were openly hostile, condemning him for deceiving people and putting them in such a stressful situation, but others argued that his findings more than justified the temporary discomfort that his subjects experienced. To examine this variation, Barry Schlenker and I (Schlenker & Forsyth, 1977) asked people who had never before heard of Milgram’s studies to indicate their degree of agreement with 50 items drawn from various philosophical analyses of ethics. The items ran the gamut from ones concerned with the feasibility of universal ethical codes to ones concerned with deception to those pertaining to harm to research participants. We identified two robust dimensions when we factor analyzed these data. Items with substantial loadings on the idealism-pragmatism factor pertained to concern for consequences (e.g., “The ethicality of a study depends on the amount of psychological harm that could potentially occur to participants”). Items that loaded on the rule-universality (relativism) factor pertained to moral principles (e.g., “It is possible to develop rigid codes of ethics that can be applied without exception to all psychological research”). (For full citations and papers, please visit my profile page.).
I reworded the items so that they applied to any behavioral domain, rather than just psychological experimentation, and added new items to more broadly sample the idealism and rule-universalism domains. After item analysis and a second revision the Ethics Position Questionnaire (EPQ) included 20 items, 10 for each of the idealism and relativism scales.
Forsyth, D. R. (1980). A taxonomy of ethical ideologies. Journal of Personality and Social Psychology, 39, 175-184.
Recently, working with Dr. O’Boyle of the University of Indiana, we have developed a shorten version of the EPQ: The EPQ5. If you are interested in completing the EPQ, and learning what your idealism and relativism scores are, please visit this Survey page.
Using the Items in Research
If you are interested in using the EPQ in research, in most cases respondents indicate degree of agreement with each item using a scale that ranges from disagreement (1) to agreement (9). Idealism scores are calculated by summing responses from items 1 to 10. Relativism scores are calculated by summing responses from items 11 to 20. The original response scale used was a 9-point scale, although people often trim it back to a true Likert 5-point scale.
Analyzing the Results
Because this is a research scale, there are no “norms” available to interpret the meaning of the scores. However, if you want to compare your findings to a baserate, then it would be best to use the mean and median based on the studies reviewed by Forsyth, O’Boyle, & McDaniel in their 2008 paper. That way, you can say things like “80% of the respondents in my study scored below the median on idealism”. But, if all you want to talk about is differences among your respondents, then I would use the median from the population you studied. That way, you can say things like “Among the respondents in this study, those who were low in idealism were most likely to act immorally.” The analyses tend to be easier if you use the median of your own data because it makes the cell sizes more equivalent if you do any type of median split analysis (there are equal numbers of people who are high Is and low Is and high rs and low rs).
If you decide to use the normed median, then that median should be based on the version of the EPQ you used. People use varying numbers of items from the original EPQ, and they also change the response scale. The original scale had 10 items for each subscale, and the scale ranged from 1 to 9 for each item. Therefore, people could score from 10 to 90 on these scales, originally. The mean and median, assuming a 9-point response scale (so that scores could range from 10 to 90), based on a review of 139 samples drawn from 29 different countries, for a total sample of 30,230 respondents, are shown in the table below.
| Scale | Mean | Median |
| Idealism | 65.52 | 66.06 |
| Relativism | 52.74 | 54.54 |
If you used fewer items (sometimes, based on scaling work, people drop out a few of the lowest contributing items, but analysis does indicate that the fewer items used to assess idealism, the lower the idealism scores) and a different response scale (such as 1-5), your medians and means would not be comparable to those in the table. To recalculate the rescaled means and medians for your metric use the following formulae:
(Highest score possible) * .734 = Median for Idealism
(Highest score possible) * .606 = Median for Relativism
Most individuals conduct simple bivariate correlational analyses or regression analysis to test their hypotheses. These types of analyses permit them to draw such conclusions as “Individuals who were more idealistic were less likely to do X” or “Increases in relativism were associated with more positive attitudes toward Y.” In some cases, too, individuals conduct one sample t-tests to determine if their sample is significantly different from the means reported in the table above.
It should be noted, however, that the original theory maintained that the two dimensions of the EPQ interact with each other to predict judgment and behavior (see “ethics positions” for more information). So, even though most researchers do not conduct such tests (In a recent meta-analysis of research involving the EPQ we wanted to see if the prediction that idealism interacted with relativism to predict moral judgments held up, but so few researchers actually tested for the interaction that we had to give up on that goal.) some type of test of their interaction should be considered.
First, you could can carry out a median split of the two dimensions, and then test their interaction. That does mean that some people who are right at the cut-off are arbitrarily placed in the one category rather than another.
Second, you could choose “extreme scorers,” by dropping people who are not only at the median but also those who are close to the median. As you do that, you gradually lose power because your n drops, but the hope is you gain power as well by creating more differentiation among your subjects. There no rules, though, about how far you should move away from the median…perhaps, a quarter standard deviation on either side of the median would do it, but that will cost you as much as 30% of your subjects (and not to mention that you might end up dropping a person who is close to the median on idealism, for example, but has a very extreme score on relativism). If you do continue with the median splits approach, the best analysis is a 2 X 2 (high-low relativism X high-low idealism) and not a 1 X 4 (absolutism, subjectivism, situationism, exceptionism).
Third, probably the most statistically powerful, if complicated, approach, is to conduct a moderated multiple regression analysis. Leave the idealism and relativism scores as continuous variables, and use them to compute an interaction score (just multiply them together). In the first regression step enter the two main effect variables (idealism and relativism). In the second step, enter the interaction. If the interaction is significant, then you need to conduct additional tests to determine the shape of the interaction (see this page for additional information).
The Items: The instructions and items for the EPQ
|
The Ethics Position Questionnaire Please indicate if you agree or disagree with the following items. Each represents a commonly held opinion and there are no right or wrong answers. We are interested in your reaction to such matters of opinion.Rate your reaction to each statement by writing a number to the left of each statement where: 1. People should make certain that their actions never 2. Risks to another should never be tolerated, irrespective 3. The existence of potential harm to others is always 4. One should never psychologically or physically harm 5. One should not perform an action which might in any way 6. If an action could harm an innocent other, then it should not 7. Deciding whether or not to perform an act by balancing 8. The dignity and welfare of the people should be the most 9. It is never necessary to sacrifice the welfare of others. 10. Moral behaviors are actions that closely match ideals of 11. There are no ethical principles that are so important 12. What is ethical varies from one situation and society to 13. Moral standards should be seen as being individualistic; 14. Different types of morality cannot be compared as to 15. Questions of what is ethical for everyone can never be 16. Moral standards are simply personal rules that indicate 17. Ethical considerations in interpersonal relations are so 18. Rigidly codifying an ethical position that prevents 19. No rule concerning lying can be formulated; whether a lie 20. Whether a lie is judged to be moral or immoral depends |
Additional Background Psychometric Information
The internal consistency coefficients of the idealism and relativism scales, as assessed by Cronbach’s alpha, range from .73 to .84, and test-retest reliabilities were .67 and .66, respectively (Forsyth, 1980; Forsyth, Nye, & Kelley, 1988). Forsyth (1980) also found that the two scales were orthogonal to one another, and were only slightly correlated with social desirability. Age trends were significant, however, as older individuals tended to be slightly less idealistic in their outlook, but also less relativistic. The rs between and age and idealism and relativism were -.20 and -.25, respectively. Forsyth (1980) examined the relationship between the EPQ and moral maturity as measured by Rest’s Defining Issues Test (DIT; Rest et al., 1999) and Hogan’s Survey of Ethical Attitudes (SEA; Hogan, 1973). The P-score of the DIT did not correlate with the EPQ scores, and when individuals were classified into the 4 categories of the 2 (high vs. low on idealism) X 2 (high vs. low on realivism) model they did not occupy any developmental stage more frequently than another. The EPQ’s relativism scale was, however, negatively correlated at -0.31 with Hogan’s (1973) SEA. This finding confirms the meaning of the relativism scale since low scores on the SEA are indicative of a rejection of societal regulatory standards in favor of an “ethics of personal conscience.” Forsyth et al. (1988) found that the idealism scale correlated .53 with scores on an “ethic of caring” (Gilligan, 1982). Subsequent analyses of these responses revealed differences among the 4 IMPs, in that absolutists scored significantly higher than situationists, who had higher scores than both the subjectivists and exceptionists.
Background: The original scaling work for the development of the Ethics Position Questionnaire was part of my dissertation conducted at the University of Florida, under the direction of Barry Schlenker. The committee included Dr. Marvin Shaw, Dr. Joel Cohen, Dr. Thomas Simon, Dr. Larry Severy, Dr. William Yost, and Dr. R. I Watson. Items from the EPQ were originally published in Forsyth, D. R. (1980). A taxonomy of ethical ideologies. Journal of Personality and Social Psychology, 39, 175-184.
Permission: Permission is granted for use of the scale for research purposes, including theses and dissertations. Permission to publish the items or to use them for any commercial purpose must be secured from the American Psychological Association.
Donelson R. Forsyth 