Does Moral Reasoning Represent Sociomoral Structure or Political Ideology? a Further Exploration of the Relations between Moral Reasoning, Political Attitudes, Consistency of Moral Thought, and the Evaluation of Human Rights in Dutch Young Adults
Raaijmakers, Quinten A. W., Van Hoof, Anne, Social Behavior and Personality: an international journal
The relations between moral reasoning and political attitudes were examined in a representative sample of 2520 Dutch adolescents and young adults. This study was repeated for a second, select sample, 210 Dutch university students. In the latter study both moral reasoning and political attitudes were related to the evaluation of the 30 articles of the Universal Declaration of Human Rights (UDHR). Moral reasoning was measured using the N2 index of the Defining Issues Test (Rest. 1979a). Political attitudes were measured using 10 Likert-type multiitem scales, representing both the cultural and the economic dimension of general political ideology. Participants' general moral thought was identified through their evaluation of the 30 articles of the UDHR. The intraindividual consistency of the evaluation of the 30 UDHR articles was used as a measure of the structure of moral thought. Factor analyses revealed that moral reasoning represented something different from political ideology. Regression analyses demonstrated that only a minor part of the variance in moral reasoning scores could be predicted by the combined effect of political attitudes. Moral reasoning scores were not related to the consistency of the evaluation of the 30 UDHR articles. Using SEM analyses, several models were tested describing the relations of moral reasoning, political ideology and consistency with the evaluation of human rights. It appeared that these variables all uniquely contributed to the explanation of moral thought.
Keywords: moral reasoning, political attitudes, consistency, human rights, construct validity.
In cognitive-developmental theory, moral reasoning is understood as the manifestation of an inner-psychological and cognitive-developmental structure that governs action in situations in which moral claims conflict (Kohlberg, 1969). As a structure, moral reasoning is sharply distinguished from the content of moral judgments. Moral structures refer to general organizing principles or patterns of thought rather than specific moral beliefs or opinions (Colby, Kohlberg, & Kauffman, 1987, p. 2). The same premises underlie the various measures that are more or less closely related to Kohlberg's cognitive-developmental theory, such as the Moral Judgment Interview (MJI: Colby & Kohlberg, 1987) or paper-and-pencil tests such as the Defining Issues Test (DIT: Rest, 1979a), or the Socio-Moral Reflection Objective Measure (SROM: Gibbs, 1991). Whatever the specific theoretical claims of these tests, they all have in common the assumption that a general cognitive structure is measured rather than certain moral beliefs or specific attitudes associated with the politico-moral domain. In the determination of moral behavior and moral choices, moral reasoning is claimed to have its unique contribution parallel to other processes (i.e., motivational and ideological processes: see Kohlberg, 1986; Rest, Narvaez, Bebeau, & Thoma, 1999).
Several authors have strongly disputed the claim that moral reasoning represents a general sociomoral cognitive structure rather than a specific politico-moral attitude. Early critiques focused on the "liberal" and "American" bias in Kohlberg's theorizing and research (e.g., Shweder, 1982; Simpson, 1974). More recently, criticism has concentrated on the measurement of moral reasoning, referring to the Defining Issues Test (Rest, 1979a) as the most widely used measure of moral reasoning. Two different, though related, questions are separately raised. Lind (1995, in press) argues that the prevailing paper-and-pencil tests of moral reasoning are essentially preference measures, and in that sense capable of measuring only attitudes. In his view, a more suitable approach of measuring cognitive (or structuring) capacity is to assess the consistency of the individual ratings instead of the expressed preferences. In addition, Emler and colleagues reasoned that the strong correlation of moral reasoning scores with scores on certain political attitude scales in fact demonstrates their mutual origin: an individualistic political-cultural ideology, predominant in Western societies. In their opinion moral reasoning and political attitudes are "by and large identical things" (Emler, 1983; Emler, Palmer-Canton, & St. James, 1998; Emler, Renwick, & Malone, 1983). Both evaluations agree that the DIT's moral reasoning scores represent attitudes instead of cognitive structures, but they differ in that they understand the content of these scores as attitudes toward specific moralities in the first case and toward general politics in the second.
In response to the critics, three kinds of defences have been employed (Rest, Narvaez, Bebeau, & Thoma, 1999; Rest, Narvaez, Thoma, & Bebeau, 1999). First, there is the examination of the correlations between moral reasoning scores and scores on measures of political attitudes to verify the degree to which these scores might overlap (e.g., Thoma, Narvaez, Rest, & Derryberry, 1999). Second, there is the determination of the unique contribution of moral reasoning scores - corrected for scores on measures of political attitudes - to the explanation of the variance in scores on different measures of general moral thought and specific sociomoral behaviors, respectively (e.g., Narvaez, Getz, Rest, & Thoma, 1999). And third, is the comparison of the explanatory power of moral reasoning scores with the power of genuine measures of cognitive structure (i.e., the intraindividual consistency of moral responses; see, for instance. Rest, Thoma, & Edwards, 1997).
In this study we briefly discussed and empirically tested the three kinds of defenses of the construct validity of moral reasoning scores, particularly the scores of the Defining Issues Test.
CORRELATIONS OF MORAL REASONING SCORES WITH MEASURES OF POLITICAL ATTITUDES
Numerous studies have been published reporting moderate to strong correlations (.40 < r < .65) between moral reasoning scores and measures of political attitudes (for an overview of more than 20 studies, see Rest, Narvaez, Bebeau, et al., 1999).
Two kinds of measures of political attitudes are used: firstly measures indicating the degree of agreement with specific aspects of political ideology (such as tolerance towards minorities, freedom of speech, law and order, abortion, women's roles and religious values; for instance. Rest. 1986; Rest, Copper, Coder, Masantz, & Anderson, 1974). and secondly, measures indicating the degree of agreement with general ideological orientations (such as liberalismconservatism, self-ratings as liberal-conservative and democratic orientation; for instance, Lind, 1995; Sparks & Durkin, 1987). Whatever measurement is used, the results are invariably interpreted as indicative of general political ideology, sometimes denoted as "liberalism-conservatism" (Emler et al., 1983), other times labelled as "civic libertarianism" (e.g.. Rest, Thoma, Narvaez. & Bebeau, 1997). Although the vast majority of these studies concern rather small samples with Ns typically varying between 60 to 200 participants (e.g., of the 23 studies reviewed in Rest. Narvaez. Bebeau et al., 1999, the mean N was 112 participants), the samples strongly differ from each other in composition (from graduate students to juridical experts). Despite these differences in sample characteristics, the general picture of the reported results remains the same.
Research connecting moral reasoning to political attitudes usually involves the measurement of a few (one to three) political attitudes. The employed measures of political attitudes represent a one-dimensional concept of political ideology in terms of cultural liberalism/conservatism. Middendorp (1991). however, referring to the work of Rokeach (1973), demonstrated in his extensive inquiry into the empirical structure of political ideology that political ideology was structured along two fairly independent dimensions, at least for the Dutch population: the cultural dimension dealing with questions of equality and freedom concerning cultural issues, and the economic dimension dealing with the same kind of questions regarding economic issues, including the facilitating and supervisory function of the state in these matters. As the research of Middendorp showed, measures of political attitudes concerning freedom of speech, tolerance towards minorities, law and order, women's roles, religious values, abortion, the respondent's self-rating as politically liberal or conservative, and so on, all belong to the same cultural dimension of political ideology. A complete picture of the relationship between moral reasoning and political ideology should also include measures specifically designed to assess the economic dimension of political ideology.
Up till now, a direct test of the probability of the hypothesis that moral reasoning belongs to a common factor of political ideology has been lacking, probably because of the employment of a small and selective number of measures of political attitudes, which complicates the statistical test of the derivation of such a common factor.
THE UNIQUE CONTRIBUTION OF MORAL REASONING SCORES
Moderate to strong correlations between measures of moral reasoning and political attitudes do not necessarily imply that moral reasoning is reducible to political attitudes or to some supposed general factor of liberalism/conservatism. One way to prove this contention is to demonstrate the unique relation of moral reasoning to some kind of measure of general moral thought, while at the same time controlling for political ideology. In defence of the construct validity of the Defining Issues Test as a measure of moral reasoning, this kind of research strategy has been both advocated and utilized by Rest and his colleagues (e.g., Narvaez et al., 1999; Thoma et al., 1999).
Thoma et al. (1999) mention four studies that actually used this research strategy to examine the unique contribution of moral reasoning to the explanation of moral thought. A closer look at these studies reveals that the measurement of moral thought has been limited to the respondents' opinions or attitudes towards selective facets of the domain of moral thought. Measures were based on scales rating the degree of agreement with certain statements (i.e., opinions). In three studies these statements referred to actual policy issues (i.e., the studies of Getz, 1985, G. Rest in Rest, 1979b, and Thoma, 1993); in the remaining study the dependent measure was a measure of libertarianism (i.e., Rest et al., 1974). Conceptually, these measures all belong to the domain of political ideology itself (Middendorp, 1991). Moreover, the measurement of the control variable, political ideology, was limited to either a one-item scale measuring self-defined liberalism/ conservatism (i.e., the studies of Getz, 1985, G. Rest in Rest, 1979b and Thoma, 1993) or a multi-item scale measuring a law-and-order attitude (i.e.. Rest et al., 1974). Indeed both measures belong to the domain of political ideology, but are specific aspects of political ideology without being representative for the whole domain (Middendorp). Thus, the results of these studies demonstrate, rather, that moral reasoning is still correlated with specific political attitudes while being controlled for other political attitudes.
Despite the questions regarding the measures used in these studies, their analytical assumptions are clear and justifiable: moral reasoning cannot be reduced to political attitudes whenever enough variance in moral reasoning scores remains after having controlled for scores on measures fully representing political ideology. A direct test of this assumption is still lacking, but can be achieved by running multiple regression analysis with moral reasoning score as the dependent variable and a representative sample of political attitudes as independent variables.
MORAL REASONING SCORES AND DIRECT MEASURES OF COGNITIVE CONSISTENCY
Theoretically, moral reasoning is assumed to be a cognitive structure governing general moral thought. Although numerous measures of cognitive structure have been developed over the years, each tapping different aspects of the structuring of perception and thought (see Fransella & Bannister, 1977; Scott, 1966), the discussion of the relationship between moral reasoning scores and the structuring of moral thought has been focused on the intraindividual consistency of moral test responses.
Lind (1979) was the first to pay attention to the systematic and developmental differences in the consistency of test responses regarding Kolbergian high-stage ratings relative to low-stage ratings. He demonstrated that the gap between these two kinds of ratings increased with age and education. To capture this structural phenomenon, Lind developed the "C" index, in which the intraindividual consistency of the discrimination between high- and low-stage ratings is measured using multivariate analysis of variance of individual test responses (Lind, 1995: Lind, Hartmann, & Wakenhut, 1985).
Rest, Thoma, Narvaez et al. ( 1997) partly adopted Lind's view and incorporated the consistency of the participant's discrimination between high- and low-stage ratings in a new index of moral reasoning, the N2 index. In the N2 index, the traditional P (principled) score, which measures the importance attached by the participant to high-stage items (i.e., a score based on the ranking of statements reflecting principled reasoning), is combined with a score reflecting the degree to which higher stage items are rated higher than lower stage items (a score based on the ratings of both high- and low-stage items).
Rest, Thoma, and Edwards ( 1997) compared the performances of the traditional P score with those of the C index on several outcome criteria (i.e., the capacity to differentiate between groups differing in moral judgment expertise, correlations with moral comprehension, developmental trends, sensitivity to educational interventions and correlations with behavior and attitudes). Generally, the P index outperformed the C index.
The performances of the traditional P index were also compared to those of the new N2 index, using the same set of criteria for comparison (Rest, Thoma, Narvaez et al., 1997). Although both indices correlated highly (i.e., rs above .90), the N2 index either appeared to be a slightly better index or revealed no differences from the original P score in predicting the criterion variables.
These results show that the moral consistency factor adds little to the explanation of, and association with, several outcome variables compared to the performance of the traditional P index, which ignores the consistency of the responses to the moral reasoning test. At the same time, these studies mainly demonstrate that moral reasoning scores capture structural components of the responses to the mural reasoning lest itself, in other words, these scores reflect cognitive consistency as a structural characteristic of the independent variable. It would be more accurate to directly test the assumption that moral reasoning scores incorporate structural components of moral thought by relating moral reasoning scores to cognitive-structural characteristics of the dependent responses; that is, the consistency of the responses indicative of general moral thought. Moral reasoning could then be considered to reflect structural components of moral thought when a) moral reasoning scores substantially correlate with measures of the consistency of the dependent responses and b) when moral reasoning scores mediate the possible effects of this consistency on the actual content of moral thought. To our knowledge, such a direct test of the theoretical proposition that moral reasoning represents a cognitive structure governing general moral thought, is still lacking.
PURPOSE OF THE STUDY
This study focused on three central issues in the debate on the construct validity of measures of moral reasoning in general, and the Defining Issues Test of moral reasoning in particular.
The first issue concerns the correlations between measures of moral reasoning and measures of political attitudes. Do these correlations justify the assumption of one common factor of political ideology underlying the responses to both kinds of measures?
The second question being researched concerns the unique contribution of moral reasoning scores to the explanation of moral thought, compared to and controlled for the contribution of a representative measure of political ideology. Does enough variance in moral reasoning remain after controlling for political ideology?
Thirdly, this study aimed to test the theoretical assumption that moral reasoning provides for the cognitive structure and internal consistency of general moral thought. Are moral reasoning scores related to measures that directly estimate the consistency in responses pertaining to general moral thought?
In summarizing the results regarding the previous questions, this study finally explored the viability of five different models explaining the relationships between moral reasoning, political ideology and consistency in moral responses on the one hand, and general moral thought on the other hand.
In this study, general moral thought is defined as the evaluation of the 30 articles of the Universal Declaration of Human Rights (UDHR). The purpose of our study then became to empirically investigate the relationships between moral reasoning, political ideology and the evaluation of the UDHR. In 1948 the UDHR was accepted by the Assembly of the United Nations as the principal document defining the fundamental rights, duties and responsibilities of individuals in their reciprocal relationships with other individuals, public organizations and the state. Therefore, the evaluation of the 30 articles of the UDHR. which taken together contain all relevant issues in the broad field of human rights, may be considered as fairly representative of general moral thought. In the next section, we briefly introduce the 30 articles of the UDHR.
In answering the research questions, data of two separate samples were used. The first sample was a selected sample of university students in law and humanities. The respondents of this sample were considered to be capable of fully comprehending the quite difficult language of the UDHR. The second sample was representative of the Dutch native population between the ages of 15 to 24 years. Except for the 30 articles of the UDHR, the same measures (moral reasoning and political attitudes) were administered to this representative sample. The results obtained from the representative sample are used to discuss the extent to which the results obtained from the selected sample may be generalized to the Dutch population of young adults.
CHARACTERIZATION OF THE 30 HUMAN RIGHTS ARTICLES
The Classification of Human Rights According to Their Subject Matter René Cassin. the principal designer of the Universal Declaration of Human Rights (UDHR), made an a priori distinction between six groups of human rights (see Agi, 1980). The first group of articles concerns the universal and moral principles of freedom, equality and dignity of mankind (Articles 1 and 2); the second group deals with the rights of the individual, both in a general basic sense regarding respect for life and abolition of slavery and torment (Articles 3, 4 and 5) and in a more applied juridical sense concerning the lawful protection of the individual (Articles 6 to 11); the third group focuses on the rights concerning the concrete relationships between human beings, such as the freedom of movement and the right to marry and start a family (Articles 12 to 17); the fourth group includes precisely defined public rights such as freedom of organization and the right to participate in public administration (Articles 18 to 21); the fifth group contains obvious economic and cultural rights, such as the right to labor, social security and education (Articles 22 to 27); and the sixth group again involves more general rights related to society as a whole and the international community, such as conformity to the principles and goals of the United Nations (Articles 28 to 30).
The Empirical Classification of Human Rights In an extensive survey of about 7000 respondents from 35 different countries all over the world, Doise, Spini and Clémence (1999) studied the empirical clustering of the 30 articles of the UDHR. The above-mentioned theoretical classification was not exactly replicated in the minds of the respondents. Five clusters of human rights appeared, which were defined as principled (Articles 1 and 2), basic (Articles 3 to 5), individual (Articles 6 to 11), social (Articles 12 to 27) and societal (Articles 28 to 30). In this classification the articles dedicated to the rights of the individual were clustered into two different groups, one containing general basic rights of the individual and one containing juridical rights of the individual. The articles 12 to 27, on the other hand, collapsed into one group of social rights instead of three different groups (i.e., social, public and political rights).
Our research involved two samples. The first, select sample consisted of 210 university students in law or humanities: 67 law students and 143 humanities students (164 females and 46 males; mean age 21.7 years, range 18 to 24 years). No claims are made concerning the representativeness of this sample. The purpose was to select a sample of young adults, who might be expected to fully understand the wording of the 30 articles of the UDHR. The results of this sample are used for answering all of the previously mentioned research questions.
The second sample was derived from the Utrecht Study of Adolescent Development (USAD) research project (Meeus & 't Hart, 1993). This sample consisted of 2,520 respondents and can be considered representative of the Dutch native population between the ages of 15 to 24 years (53% females and 47% males; mean age 19.5 years, range 15-25 years). With the exception of one measure (i.e., the evaluation of human rights), the same measures were administered to this sample. The results of this sample are used for answering the first and the second research question concerning the relationships between moral reasoning and political attitudes. In comparing the results of the representative sample with the results of the students' sample, some further indications might be obtained about the generalizability of the results obtained exclusively from the students' sample, which concern the relationships of moral reasoning and political attitudes with the evaluation of human rights. To enhance this comparison, a subsample from the representative sample was selected containing 421 respondents, comparable in age and education with the students' sample (52% females and 48% males; mean age 21.9 years, range 18-24 years).
Evaluation of human rights (students' sample only) Participants were asked to evaluate the 30 articles of the 1948 Universal Declaration of Human Rights by answering the same eight questions about each of these articles on 9-point bipolar scales. The eight questions were originated from a study of Doise, Spini, Jesuino, Ng, and Emler (1994) and aimed to assess the general evaluation of the articles. The English wording of the eight opposed poles were: (1) difficult-easy to understand, (2) difficult-easy for governments to apply, (3) clear-no clear implications for individuals' responsibilities toward others (reversed scale), (4) political parties can-cannot do much to enforce this article (reversed scale), (5) relevant-not relevant to my rights as a private individual (reversed scale), (6) I can personally do a great deal-very little for the respect of this article (reversed scale), (7) I don't agree-I agree with every aspect, and (8) I am willing-not willing to join other concerned people to defend this article (reversed scale).
Participants' responses were coded in such a way that higher scores were associated with a more positive evaluation of human rights.
Human rights category scores (students' sample only) According to the empirical classification of Doise et al. (1999), the 30 human rights articles include five categories of rights: principled (Articles 1 and 2), basic (Articles 3 to 5). individual (Articles 6 to 11), social (Articles 12 to 27) and societal (Articles 28 to 30). Therefore, the participants' general evaluation of these separate categories of human rights articles was summarized in a category score that was computed as the mean of the scores on the bipolar scales belonging to that category. Since the participant's responses to these scales were coded in a positive direction, a higher category score reflected a more positive evaluation of the human rights articles regarding that category. The reliabilities (as) of the resulting five scores were satisfactory (.76, .84, .91, .96 and .91, respectively).
Human rights component scores (students' sample only) To reduce the number of variables in analysis, the five category scores were examined to see whether they could be summarized in component scores. For the principal designer of the UDHR the different groups of human rights all together reflect one general and homogeneous concept of the universal rights of human beings (Agi, 1980). Others, however, maintain that the UDHR represents rather a multidimensional concept combining both moral-ethical and sociopolitical principles (e.g., Crowson, 2004; Diaz-Veizades, Widaman, Little, & Gibbs, 1995).
Using confirmatory factor analysis, two models were tested representing the possible latent structure underlying the five human category scores. In the first model, a one component structure was hypothesized; in the second model, a correlated two-components structure was assumed, with the principled and basic categories loading on one component (labeled as the moral component), and with the other three categories (i.e., individual, social, and societal) loading on the other component (labeled as the political component). The one-component model did not fit unambiguously well (χ^sup 2^(5) = 13.55, p = .019, AGFI = .924, NFI = .978 and CFI = .986). The two components model sufficiently fitted the data (χ^sup 2^(4) = 2.39, p = .664, AGFI = .983, NFI = .996 and CFI = 1.000). The latter model also proved to be a significant improvement compared to the former model (Δχ^sup 2^(1) = 11.16, p < .001). Therefore, the 5 human rights category scores were summarized into 2 human rights component scores, with each component score computed as the mean of the category scores belonging to that component.
Moral reasoning To assess the respondents' level of moral reasoning, a translated version of the short form of the Defining Issues Test (DIT; Rest, 1979a) was applied. The DIT contains moral dilemmas, adapted from Kohlberg's interview, which are administered as part of a paper-and-pencil test. The participant is asked to rate the importance of different arguments in solving each separate dilemma. Each presented argument is indicative of a specific stage of moral reasoning. It is assumed that individuals will attach more importance to those arguments that come closest to their own level of moral reasoning.
The DIT delivers several indices of the level of moral reasoning, of which the N2 index is the most recently advocated (Rest, Thoma, Narvaez et al., 1997). The N2 index is built upon two parts. The first part reflects the degree to which stage 5 or stage 6 (i.e., postconventional) arguments are prioritized in solving the presented moral dilemmas. This part of the N2 index resembles the traditional P (principled) score. The second part of the N2 index reflects the degree to which higher stage arguments are rated higher than the ratings of lower stage arguments: the average rating given to stage 2 and stage 3 arguments (lower stage reasoning) is substracted from the average rating given to stage 5 and stage 6 arguments (higher stage reasoning). After standardizing the scores of the second part in such a way that both parts show the same mean and standard deviation, the N2 score is computed by adding the resulting scores of the two parts (for a detailed description of the computing procedure of the N2 index, see Rest, Thoma, Narvaez et al.). A higher N2 score represents a higher level of moral reasoning.
Previous research demonstrated ample evidence for the validity and reliability of the DIT indices (for an overview, see Rest, Narvaez, Bebeau et al., 1999). Regarding the N2 index of the short form of the DIT, Cronbach's alphas varying between .70 and .74 are reported (Rest, Thoma, Narvaez et al., 1997). In the present study, the reliability coefficients were .76 for the students' sample and .68 for the representative sample (Cronbach's alphas). Correlations between the N2 index and the traditional P index were .95 and .92, respectively, which is quite similar to the values reported by Rest. Thoma. Narvaez et al.
Political ideology The scores on 10 multi-item, 5-point Likert-type attitude scales were used for the measurement of both the cultural and the economic dimension of political ideology. All scales were taken from previous studies of researchers in the field of Dutch political ideology. A brief description of the scales is given below including the reliability coefficients for both samples (first the students' sample and second the representative sample). A complete description of the scales can be found in the cited references.
Two scales represented the economic dimension of political ideology. The scale Equality of income and property (Middendorp, 1991; 6 items, Cronbach's αs: .87 and .88) measures the belief that differences in income, socioeconomic status and material possessions should be smaller than they are. The scale Socioeconomic radicalism (Felling. Peters. & Schreuder, 1983; 4 items, αs: .82 and .83) measures the striving for socioeconomic equality among people.
For the measurement of the cultural dimension of political ideology, 8 attitude scales were administered. The scale Sexism (Vollebergh, 1986; 14 items, αs: .89 and .89) measures political intolerance towards women and the feminine, in which the full political and social rights of women as a group are denied. The scale Division of domestic labor (Spruijt, 1993; 3 items, as: .88 and .78) measures the attitude towards an equal division of this labor between adult partners. The scale Tolerance of alternative lifestyles (Van der Avort, 1988; 8 items, αs: .93 and .88) assesses the acceptance of social relationships other than the traditional family. The scale Conventional male and female roles (Middendorp, 1991; 4 items, αs: .80 and .74) measures the feelings towards a rigid division of the social roles of men and women. The scale Division of power in intimate relations (Spruijt. 1993; 7 items, αs: .90 and .89) measures opinions on the equal distribution of material and immaterial power resources in intimate relationships. The scale Autlioritarianism was the shortened form of the Dutch version of the original F-scale (Vollebergh, 1991; 11 items, αs: .79 and .79). The scale Ethnocentrism (Hagendoorn & Janssen, 1983: 12 items, αs: .88 and .92) measures ethnocentric feelings or an intolerant attitude towards outgroups. Finally, the scale Freedom of political expression (Middendorp, 1991; 6 items, αs: .82 and .79) measures the belief that people should have the right to express their political opinions publicly.
Two summated scale scores were constructed, one consisting of the summation of the 2 economic scale scores measuring the political economic orientation (αs: .76 and .76), and one consisting of the summation of the 8 cultural scale scores measuring the political cultural orientation (αs: .90 and .82). Both scores were coded in such a way that high scores reflected conservative orientations.
Moral thought consistency (students' sample only) Over the years several measures of response consistency have been developed. Block (1961) developed a measure based on the percentage of explained variance by the first principal component. Others computed the mean correlation between comparable responses (Hauser, 1976; Van Hoof & Raaijmakers, 2002). We have already mentioned the study of Lind (1995), who performed a MANOVA on the individual responses.
Characteristic for all these measures is the statistical analysis of responses at an intraindividual level. In the present study, consistency was measured on the evaluative responses to the 30 human rights articles. For each individual participant a separate data matrix was constructed, with the 30 articles as rows (i.e., as cases) and the evaluative constructs as columns (i.e., as variables). Next, for each participant individually, consistency scores were obtained by conducting a principal component factor analysis of the evaluative constructs. The percentage of variance explained by the first component is used as the consistency score. This measure was chosen, because it still is the most widely used and accepted measure of consistency (e.g., Cross, Gore, & Morris, 2003; Dunkel, 2005: Epting, Prichard, Wiggins, Leonard, & Beagle, 1992). The consistency score represents a true structural measure, in the sense that it is based on the relationships among responses, rather than the direction, strength or number of responses. The measure is also a true intrapersonal measure, since it is based on the individual pattern of moral evaluations of the respondent rather than the pattern of relationships between variables across respondents.
FACTOR ANALYSES OF MORAL REASONING SCORES AND POLITICAL ATTITUDE SCALES SCORES
Several confirmatory factor analyses were performed, using the structural equation software package of AMOS, version 5 (Arbuckle, 2003). In the analyses, three different models of the relations between measures of moral reasoning and political attitudes were tested. In the first model (Model 1), the assumption of one common factor underlying both the moral reasoning scores and the political attitude measures was tested. The second model (Model 2) tested the assumption of two related factors, one economic factor and one cultural factor, with moral reasoning loading on the latter factor. In the third model (Model 3) moral reasoning was excluded from analysis to establish a two-dimensional model representing political ideology based solely on the political attitude scales.
Table 1 reports a summary of the statistics resulting from these subsequent analyses for the three samples. Because of both the large differences in sample size and the three models not being nested models, the usual fit statistics (based on the χ^sup 2^) are inaccurate. For evaluating and comparing the fit of the models for the three samples, the following fit statistics were considered: 1) AGFI (Adjusted Goodness of Fit Index) represents the proportion of the observed covariances explained by the model corrected for model complexity (i.e., models with more parameters). Values above .90 reflect a good fit. 2) CFI (Comparative Fit Index) indicates the relative improvement in fit of the model compared to the null model (i.e., the independence model), corrected for sample size. Values above .90 reflect a good fit. 3) AIC (Akaike Information Criterion) is especially designed for comparing the fit of different models, corrected for the number of parameters. The smaller the value of AIC, the better the fit.
The fit statistics reported in Table 1 are in favor of the models representing political ideology as a two-dimensional construct (i.e., Models 2 and 3). Moreover, the model excluding moral reasoning from the domain of political ideology (i.e., Model 3) fits the data best according to the value of the AIC index. Additional support for the power of Model 3 over Model 2 is obtained by the amount of explained variance by both models, the low factor loadings of the moral reasoning scores in Model 2 and the minimal amount of the variance in moral reasoning scores that is explained in Model 2 by the cultural component of political ideology (see the results in the last three columns of Table 1).
REGRESSION ANALYSIS OF MORAL REASONING SCORES WITH MEASURES OF POLITICAL ATTITUDES AS PREDICTORS
To verify whether enough variance in moral reasoning scores remained after controlling for measures of political attitudes, a regression analysis was performed for each sample with the N2 scores as dependent variables and the scores on the 10 political attitude scales as independent variables. For the students' sample this resulted in a multiple correlation coefficient (MR) of .30 and an R^sup 2^ of .09. For the USAD sample and the USAD subsample, MRs were .23 and .21 with R^sup 2^s of .05 and .05, respectively. In terms of effect size (Cohen, 1988), the joint effect of measures of political attitude scales on the variance of moral reasoning scores appeared to be fairly small for all three samples.
CORRELATION OF MORAL REASONING SCORES WITH CONSISTENCY OF THE EVALUATION OF HUMAN RIGHTS
The percentage of variance in human rights evaluative responses explained by the first component was used as the consistency score and correlated with the N2 score, measuring developmental level in moral reasoning. Data are available of the students' sample (N = 210). The correlation between both variables was neither statistically nor substantially significant: -.034, p = .312. This result also implies that N2 scores do not mediate the possible effect of consistency on the evaluation of human rights.
MODELLING THE RELATIONS BETWEEN MORAL REASONING, POLITICAL IDEOLOGY, CONSISTENCY, AND THE EVALUATION OF HUMAN RIGHTS
Five hierarchical models were tested and compared to obtain a more detailed picture of the relative effect of moral reasoning scores on the evaluation of human rights, as compared with similar effects of political ideology and consistency. Data are available of the students' sample (N = 210). Figure 1 depicts the Full Model, containing all plausible relations between the variables (i.e., the correlations between moral reasoning and both components of political ideology, the correlation between the two components of the evaluation of human rights, and the effects of moral reasoning, political ideology and consistency on the evaluation of the two human rights components). As reported in Table 2, the Full Model fitted the data well.
The second model, labelled Basic Model in Table 2. excluded the effects of both moral reasoning and political ideology by fixing the corresponding paths to zero. In this model, therefore, only consistency is allowed to affect the evaluation of human rights. Not surprisingly, this model performed worse than the Full Model (Δχ^sup 2^(9) = 84.82, p< .001). However, the results with this model, indicated a direct effect of consistency on the evaluation of both the moral and the political component of human rights (βs .14 and .16, respectively, ps < .05), explaining about 2% of the variance of each human rights evaluation component.
Next, a model was tested in which - compared to the basic model - moral reasoning additionally was allowed to affect the evaluation of human rights. This model, in Table 2 labelled as Moral Reasoning Model, improved the fit, compared to the basic model (Δχ^sup 2^(2) = 6.97, p < .05), but still performed worse than the Full Model (Δχ^sup 2^(7) = 77.85. p < .001). The results with this model demonstrated a direct effect of moral reasoning on the evaluation of the moral component of human rights (β = .13, p <.05), explaining about an additional 12% of its variance. However, moral reasoning did not relate to the evaluation of the political component of human rights (β = -.01, p =.89).
Next, a model was tested in which - compared to the basic model - the two components of political ideology additionally were allowed to affect the evaluation of human rights. This model, in Table 2 labelled as Political Ideology Model, improved the fit, compared to the basic model (Δχ^sup 2^(5) = 58.17, p < .001), but still performed worse than the Full Model (Δχ^sup 2^(4) = 26.65, p < .001). The results with this model showed a direct effect of political cultural ideology on the evaluation of both the moral and the political component of human rights (βs = -.17, and -.14, respectively, ps <.05); the economic dimension related to the political component only (βs = -.09, p=.20 and -.17, p<.05, respectively). Compared to the basic model, political ideology explained about an additional 5% of the variance in the evaluation of the moral component of human rights, and about an additional 6% of its political component.
Finally, the Final Model was constructed and tested. This model allowed both moral reasoning and political ideology to affect simultaneously the evaluation of human rights. This model not only showed a better fit than the Basic Model (Δχ^sup 2^(7) = 81.84, p <.001), but also did not differ significantly from the Full Model (Δχ^sup 2^(2) = 2.98, p = .23). The standardized regression coefficients and correlations resulting from this final model are shown in Figure 1. Compared to the Political Ideology Model, this model explained about an additional 1% of the variance in the evaluation of the moral component of human rights suggesting that a part of the effect of moral reasoning is mediated by political ideology.
This study demonstrated that moral reasoning, although related to political cultural attitudes, measures something different. The assumption of one common factor (i.e., political cultural ideology) underlying moral reasoning and political attitudes is not validated. Moreover, the common variance of moral reasoning and a broad range of political attitudes appeared to be rather small, which indicates the unique potential of moral reasoning in explaining certain outcome variables, in comparison with the explanatory power of political attitudes.
On the other hand, this study did not provide any evidence of moral reasoning actually structuring moral thought responses, in the sense that no correlation could be established between moral reasoning scores and the measurement of the consistency in the responses indicative of the evaluation of human rights. It appeared rather, that moral reasoning, political ideology and consistency of responses each contributed in a unique way to the explanation of the evaluation of human rights.
The first question we have to deal with is whether our results must be considered specific to our relatively small sample of students in the humanities. Compared to the research reviewed by Rest. Narvaez, Bebeau, and Thoma (1999). this study showed striking differences in the strength of the correlations between moral reasoning scores and measures of political ideology. Generally, the correlation coefficients in the students' sample were less substantive, ranging from .15 (for political economic ideology) to .30 (for political cultural ideology), compared to values between .40 and .65 repotted by Rest. Narvaez, Bebeau et al. (1999). Although this result is not unique for our study - for instance, Hayes (2004) also reports low correlation coefficients - we discuss several explanations.
One explanation for the relatively low correlations in the students' sample could be the restricted age range of this sample. This explanation, however, does not seem very plausible since several studies reported in the review of Rest, Narvaez, Bebeau, et al. (1999) concern the same or similar age ranges.
A second explanation could be the specific and rather high educational level of our sample of university students. This explanation is, however, very unlikely since the correlations obtained from the representative sample are even weaker (ranging from .01 to .18)
Two possible explanations remain: 1) the difference between the USA and The Netherlands, and 2) the difference in the measurement of political ideology. Associated with the more pronounced secularization and liberalization of Dutch society (Middendorp. 1991). political ideology might have become less morally inspired in the Netherlands. Besides, in this study political ideology was measured using a broad range of political attitudes, whereas in the studies reported by Rest, Narvaez, Bebeau et al. (1999) political ideology was measured using a restricted definition with measures more obviously related to the moral domain. It seems likely that the differences in the strength of the correlation between moral reasoning and political ideology can be explained by both differences in national political identity and the measurement of political ideology itself.
A second issue for discussion relates to the rather small amount of explained variance in the evaluation of human rights by moral reasoning (i.e., 2%) or by the combined measures of moral reasoning and political ideology (i.e., 8%). This result contrasts with previous studies of the relations between moral reasoning, political ideology and attitudes towards human rights (e.g., Derryberry & Thoma, 2005; Narvaez et al.. 1999). The most obvious reason for this difference lies in the measurement of thinking about human rights. Other studies have used the Attitudes towards human rights inventory (ATHRI) originally developed by Getz (1985) as a measure of general moral thought. Close inspection of the 40 items of this scale, which contains issues drawn from the American Constitution's Bill of Rights, reveals a specific definition of human rights. The 40 items of the ATHRI almost exclusively (i.e.. 38 of the 40 items) deal with issues regarding the political-cultural rights of the individual (freedom of thought and speech, discrimination on the basis of ethnicity, gender, age or religion, and the selfdetermination of the individual); virtually no item deals with economic, public and social rights, or the responsibility of the state and its institutions regarding the enforcement of these rights, both nationally and internationally. In that sense, it is difficult to consider the issues mentioned in the ATHRI as representative of the issues raised in the articles of the UDHR. In addition, the ATHRI is an attitudinal measure, measuring the degree of agreement with human rights statements. In our study, however, we measured the general evaluation of human rights by means of 8 bipolar scales of evaluation, incorporating other components of moral thought besides the attitudinal or judgment component. In fact, only one of these bipolar scales dealt with the degree of agreement with the human right articles. Considering these differences in measurement, our measure of the evaluation of human rights involves several components of general moral thought, as defined in the Four-Component Model of moral thought (Narvaez & Rest, 1995). In view of all this, it seems less surprising that only a small amount of the variance in the evaluative responses regarding human rights is explained by moral reasoning scores. After all, moral reasoning is hypothesized to be primarily related to the judgment component of moral thought (Rest, 1983; Thoma, 1994).
A third point of discussion concerns the lack of any relation between moral reasoning and the structure of moral thought as measured by the consistency of the evaluation of human rights. This result should not be misinterpreted as indicating that moral reasoning is not related to any structural component of general moral thought. Consistency of responses is just one of the many relevant structural components of moral thought. It might well be possible that moral reasoning still is related to other structural components, such as ambivalence, salience, centrality, and so on.
Finally, our results demonstrated that moral reasoning affects general moral thought, just as political ideology does. An individual's political ideology represents his/her position towards the core values of political ideology: equality and freedom (Middendorp, 1991 ). It seems, that moral reasoning provides for an additional concurrent value: justice. This value not only directly and uniquely affects moral thought, but also has an indirect effect mediated by political ideology. In that sense, moral reasoning seems to represent a more fundamental value in the determination of moral thought. Future research will have to demonstrate whether moral reasoning mainly concerns the attitudinal component of this value, or additionally represents a true structure or organizational principle of moral thought. The results of our study suggest, then, to focus on structural characteristics which are different from consistency.
Agi, M. (1980). De l'idée d'universalité comme fondatrice du concept des droits de l'homme: d'après la vie et l'oeuvre de René Cassin. Amibes (France): Editions Alp'Azur.
Arbuckle, J. L. (2003). Amos 5.0: Update to the Amos users' guide. Chicago: SmallWalcrs Corp.
Block, J. (1961). Ego-identity, role variability, and adjustment. Journal of Consulting and Clinical Psychology. 25, 392-397.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. (2nd ed.) Hillsdule NJ: Lawrence Erlbaum.
Colby, A., & Kohlberg, L. (1987). The measurement of moral judgment (Volume II). New York: Cambridge University Press.
Colby, A., Kohlberg, L., & Kauffman, K. (1987). Theoretical introduction to the measurement of moral judgment. In A. Colby & L. Kohlberg, The measurement of moral judgment (Volume I, pp. 1-61). New York: Cambridge University Press.
Cross, S. E., Gore, J. S.. & Morris, M. L. (2003). The relational-interdependent self-construal, self-concept consistency, and well-being. Journal of Personality and Social Psychology. 85, 933-944.
Crowson, H. M. (2004). Human rights attitudes: Dimensionality and psychological correlates. Ethics and Behavior. 14.235-253.
Diaz-Veizades, J., Widaman, K. E, Little, T. D., & Gibbs, K. W. (1995). The measurement and structure of human rights attitudes. The Journal of Social Psychology, 135, 313-328.
Derryberry, W. P., & Thoma, S. J. (2005). Functional differences: Comparing moral judgment developmental phases of consolidation and transition. Journal of Moral Education. 34, 89-106.
Doise, W., Spini, D., & Clémence, A. (1999). Human rights as social representations in a crossnational context. European Journal of Social Psychology. 29. 1-29.
Doise, W., Spini, D., Jcsuino, J. C., Ng, S. H., & Emlcr. N. (1994). Values and perceived conflicts in the social representations of human rights: Feasibility of a cross-national study. Swiss Journal of Psychology, 53, 240-251.
Dunkel, C. S. (2005). The relation between self-continuity and measures of identity. Identity : An International Journal of Theory and Research, 5, 21 -34.
Emler, N. (1983). Morality and politics: The ideological dimension in the theory of moral development. In H. Weinreich-Haste & D. Locke (Eds.), Morality in the making (pp. 47-71). New York: Wiley.
Emler, N., Palmer-Canton, E., & St. James. A. (1998). Politics, moral reasoning and the Defining Issues Test: A reply to Bamett et al. ( 1995). British Journal of Social Psychology. 37, 457-476.
Emler, N., Renwick, S., & Malone. B. (1983). The relationship between moral reasoning and political orientation. Journal of Personality and Social Psychology, 45, 1073-1080.
Epting, F. R., Prichard, S., Wiggins, S. C., Leonard, J. A., & Beagle Jr., J. W. (1992). Assessment of the first factor and related measures of construct differentiation. International Journal of Personal Construct Psychology, 5, 77-94.
Felling, A., Peters. J.. & Schreuder, O. (1983). Burgerlijk en onburgerlijk Nederland [Bourgeois and non-bourgeois in the Netherlands). Deventer (The Netherlands): Van Loghum Slaterus. Fransella, F.. & Bannister, P. (1977). A manual for repertory grid technique. London: Academic Press.
Getz, I. (1985). The relation of moral and religious ideology to human rights. Unpublished doctoral dissertation. University of Minnesota, MN, USA.
Gibbs, J. C. (1991). Towards an integration of Kohlberg's and Huffman's theories of morality. In W. M. Kurtines & J. L. Gerwitz (Eds.), Handbook of moral behavior and development. Vol. 1: Theory (pp. 183-222). Hillsdale, NJ: Lawrence Erlbaum.
Hagendoom, L., & Janssen, J. (1983). Rechtsomkeer [Turn right]. Baam (The Netherlands): Ambo.
Hauser, S. T. (1976). Self-image complexity and identity formation in adolescence: Longitudinal studies. Journal of Youth and Adolescence. S, 161-176.
Hayes, M. A. (2004). Students' moral judgment, cultural ideologies, and moral thinking at evangelical Christian liberal arts colleges. Dissertation Abstracts International, 65 (4A), 1273,
Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and research (pp. 347-480). Chicago: Rand McNally.
Kohlberg, L. (1986). A current statement on some theoretical issues. In S. Modgil & C. Modgil (Eds.), Lawrence Kohlberg: Consensus and controversy (pp. 485-546). Philadelphia: Palmer.
Lind, G. (1979). Moral development - A new issue in higher education research. In G. Framheim (Ed.), Report of a special seminar in university students - Their training and conception of life (pp. 52-61 ). Unpublished manuscript. University of Konstanz, Germany.
Lind, G. (1995). The meaning and measurement of moral competence revisited. Paper presented at the Annual Meeting of the American Educational Research Association, San Francisco.
Lind, G. (2000). The meaning and measurement of moral judgment competence revisited. A dual aspect model. In D. Fasko & W. Willis (Eds.), Contemporary philosophical and psychological perspectives on moral development and education. Cresskill, NJ: Hampton Press.
Lind, G., Hartmann, H. ?., & Wakenhut. R. (Eds.), (1985). Moral development and the social environment. Chicago: Precedents Publishing.
Meeus, W., & Hart, H. 't (1993). Jongeren in Nederland [Youth in the Netherlands). Amersfoort, The Netherlands: Acco.
Middendorp, C. P. (1991). Ideology in Dutch politics. Assen (The Netherlands): Van Gorcum.
Narvaez, D., Getz, I., Rest, J., & Thoma, S. J. (1999). Individual moral judgment and cultural ideology. Developmental Psychology, 35, 478-488.
Narvaez, D., & Rest, J. (1995). The four components of acting morally. In W. Kurtines & J. Gerwitz (Eds.). Moral behavior and moral development: An introduction (pp. 385-400). New York: McGraw-Hill.
Rest, J. R. (1979a). Revised manual for the Defining Issues Test. Unpublished manuscript, Minneapolis: University of Minnesota.
Rest, J. R. (1979b). Development in judging moral issues. Minneapolis: University of Minnesota Press.
Rest, J. R. (1983). Morality. In P. H. Mussen, J. Flavell, & E. Markman (Eds.), Handbook of child psychology: Volume 3, Cognitive development (pp. 556-629). New York: Wiley.
Rest, J. R. ( 1986). Moral development: Advances in research and theory. New York: Praeger.
Rest, J., Copper. D., Coder, R., Masantz. J., & Anderson, D. (1974). Judging the important issues in moral dilemmas: An objective test of development. Developmental Psychology, 10, 491-501.
Rest, J. R., Narvaez, D., Bebeau, M. J., & Thoma, S. J. (1999). Postconventional moral thinking. A neo-Kohlbergian approach. Mahwah, NJ: Lawrence Erlbaum Ass.
Rest, J. R., Narvaez, D., Thoma, S. J., & Bebeau, M. J. ( 1999). DIT2: Devising and testing a revised instrument of moral judgment. Journal of Educational Pxvchollogy, 91, 644-659.
Rest, J., Thoma. S.. & Edwards. L. ( 1997). Designing and validating a measure of moral judgment: Stage preference and stage consistency approaches. Journal of Educational Psychology, 89, 5-28.
Rest, J. R., Thoma, S. J., Narvaez, D.. & Bebeau, M. J. (1997). Alchemy and beyond: Indexing the Defining Issues Test. Journal of Educational Psychology, 89, 498-507.
Rokeach, M. (1973). The nature of human values. New York: The Free Press.
Scott, W. A. ( 1966). Brief report: Measures of cognitive structure. Multivariate Behavioral Research, 1, 391-395.
Shweder, R. A. (1982). Review of Lawrence Kohlberg's essays on moral development. Volume I: The philosophy of moral development. Contemporary Psychology, 421-424.
Simpson, E. L. ( 1974). Moral development research: A case study of scientific cultural bias. Human Development, 17, 81-106.
Sparks, P., & Durkin, K. (1987). Moral reasoning and political orientation: The context sensitivity of individual rights and democratic principles. Journal of Personality and Social Psychology, 52,931-936.
Spruijt, E. (1993). Relaties: feiten, opvattingen en problemen. In W. Meeus & H. 't Hart (eds.), Jongeren in Nederland [Youth in the Netherlands] (pp. 56-78). Amersfoort (The Netherlands): Acco.
Thoma, S. J. ( 1993). The realtionship between political preference and moral judgment development in late adolescence. Merill-Palmer Quarterly, 39, 359-374.
Thoma, S. J. (1994). Moral judgment and moral action. In J. Rest & D. Narvaez (Eds.), Moral development in the professions: Psychology and applied ethics (pp. 199-211). Hillsdale. NJ: Lawrence Erlbaum Ass.
Thoma, S. J., Narvaez, D., Rest, J., & Derryberry, P. (1999). Does moral judgment development reduce to political attitudes or verbal ability? Evidence using the Defining Issues Test. Educational Psychology Review, 11, 325-341.
Van der Avon, A. (1988). Vrijinnigheid in relatiex, hedendaagxe attitudes in de primaire levenssfeer [Freedom in relations, recent attitudes on intimate relationships]. Nijmegen (The Netherlands): ITS.
Van Hoof, A., & Raaijmakers, Q. A. W. (2002). The spatial integration of adolescent identity: Its relation to age, education, and subjective well-being. Scandinavian Journal of Psychology, 43, 201-212.
Vollebergh, W. (1986). De politisering van seksisme. In M. Matthijssen, W. Meeus, & K van Wel (Eds.). Beelden van jeugd [Images of youth] (pp. 68-89). Groningen (The Netherlands): Wolters Noordhof.
Vollebergh, W. (1991). The limits of tolerance. Unpublished doctoral dissertation. Utrecht University, The Netherlands.
QUINTEN A. W. RAAIJMAKERS AND ANNE VAN HOOF
Utrecht University, The Netherlands
Quinten A. W. Raaijmakers and Anne Van Hoof. Utrecht University, The Netherlands.
Appreciation is due to anonymous reviewers.
Please address correspondence and reprint requests to: Quinten A. W. Raaijmakers, Department of Child and Adolescent Studies, Utrecht University, Postbox 80140. 3508 TC Utrecht. The Netherlands.
Phone: 31-30-2534650; Fax: 31-30-2532352; Email: Q.A.W.Raaijmakers@fss.uu.nl…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Does Moral Reasoning Represent Sociomoral Structure or Political Ideology? a Further Exploration of the Relations between Moral Reasoning, Political Attitudes, Consistency of Moral Thought, and the Evaluation of Human Rights in Dutch Young Adults. Contributors: Raaijmakers, Quinten A. W. - Author, Van Hoof, Anne - Author. Journal title: Social Behavior and Personality: an international journal. Volume: 34. Issue: 6 Publication date: July 1, 2006. Page number: 617+. © 2009 Scientific Journal Publishers, Ltd. Provided by ProQuest LLC. All Rights Reserved.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.