Requests for copies should beaddressed to:TGroenewald,Technikon SA, Private Bag X6, Florida,1710 * For the purpose of this study no distinction is made between morality and ethics; and these concepts are treated as synonymous (Hugo & VanVuuren (1995) THE CONSTRUCTION OF A NORMATIVE INSTRUMENT FOR THE MEASUREMENT OF MORAL REASONING JC WHITE ACRAFFORD JM SCHEPERS Programme in Industrial Psychology Department of Human Resource Management Rand Afrikaans University ABSTRACT The main aimof this study was to construct a normative instrument for the measurement ofdi¡erent levels of moral reasoning.The sample comprised 426 undergraduate students in Industrial Psychology and Personnel Management from twoAfrikaans speaking universities. Aquestionnaire was developed to measure di¡erent levels of moral reaso- ning on a normative scale. A factor analysis on 90 items yielded two factors.These factors were interpreted as princi- pled moral reasoning and prescriptive moral reasoning.The two scales were subjected to an item analysis andyielded reliability coe⁄cients of 0,936 and 0,937.The implications of these ¢ndings are discussed. OPSOMMING Die hoofdoel van die studie was om’n normatiewe meetinstrument te konstrueer vir die meting van verskillende vlakke van morele redenering. Die steekproef het uit 426 voorgraadse students in Bedryfsielkunde en Personeelbe- stuur aan twee Afrikaanstalige universiteite bestaan.’n Normatiewe skaal is ontwikkel om verskillende vlakke van morele redenering te meet.’n Faktorontleding vandie 90 items in die skaal het twee faktore totgevolg gehad, naamlik principie« le morele redenering en voorskriftelike morele redenering. Die skale is vervolgens aan’n itemontleding on- derwerp en het betroubaarhede van 0,936 en 0,947, onderskeidelik, opgelewer. Die implikasies van die bevindinge word bespreek. The question is raised, time and again, whether morality should not be the life-artery of good business. Nevertheless, world wide crime, and white collar crime in particular, are the order of the day (Stead,Worrell & Stead,1990). It appears that morality is con- sidered in theory only.White-collar crime is rife in South Africa and continually a¥icts local organisations and institutions. The problems accompanying tax collection, for instance, illustrate the grave lack of morality in South Africa (Rossouw,1997).The problem has reached such proportions that organisations have found it necessary to launch a project like‘‘Business Against Cri- me’’, which among other things focuses on establishing strong moral values within organisations. Furthermore, since 1994, the King report on corporate management has made it compulsory for organisations listed on the Johannesburg Stock Exchange to present an annual report onthe state of their ethical codes of con- duct and moral culture (Rossouw,1997). In addition, there is a tendency for international organisations to introduce £atter structures. This will have far-reaching ef- fects since, until recently, organisational behaviour was regula- ted by formal, multi-level structures and corresponding modes of authority. This involved attempts to induce moral behaviour by means of policy-making. In the absence of this formal regulation of behaviour, the responsibility for moral behaviour now lies with the individual (Rossouw,1997). There will always be moral and ethical questions to deal with, especially where there is con£ict between the interests of share- holders and the personal values of employees. In this respectTre- vino (1986) points out that managers on all levels tend to subject their personal values to their own advance and success in the or- ganisation. It is a cause for concern that, in spite of all the above- mentioned considerations, organisations generally regard busin- ess ethics as low on the list of strategic priorities (Robertson, 1996). Research in this area is therefore important. The study of moral decision-making in organisations is cer- tainly nothing new, and has, for some time, been a focus of study in organisational psychology (Cadbury, 1987; Reilly & Myroslaw, 1990; Stead et al., 1990;Trevino, 1986). There is ho- wever still ample scope for research on this subject, and the aim of this study was to make a contribution to organisational psychology through developing a normative instrument for the measurement of moral reasoning. AsJordaan andJordaan (1992) explain, moral learning involves the acquisition of ethical codes of behaviour. These codes of behaviour are either universal or are accepted and prescribed by a certain society and culture. In terms of these codes, people judge behaviour as morally justi¢ed or unjusti¢ed. This ‘self- judgement’ based on ethical codes of behaviour is related to moral reasoning as described by Derry (1989) who regards moral reasoning as a way of coping with moral con£icts.The process includes a personal de¢nition and framework of the moral con£ict, as well as evaluation and resolution of the con- £ict. So di¡erent people experience moral con£ict in di¡erent ways, even when faced with the same situation. A moral decision is, of course, made with speci¢c reasons in mind, thus a comprehensive de¢nition of morality should in- clude both behavioural and cognitive factors (Taylor,1977). Mo- rality therefore involves moral reasoning or the reasons behind moral decisions, as well as the behaviour accompanying these decisions. The nature of moral development will now be considered from three points of view that have their roots in three of the schools of Psychology. First, morality is approached from the perspective of social learning theory, which adopts many of the concepts employed by learning theory, to explain how people develop and are socialised (Sieber,1980). In social learn- ing theory, behaviour is seen to be acquired by direct learning, modelling and imitation, and is maintained by positive rein- forcement. Moral behaviour is viewed as the result of learned responses conditioned by the social environment (Sieber, 1980). The process of acquiring moral behaviour depends on what is considered right or wrong by the group, and is subject to social sanctioning by the group (Windmiller, 1980). Thus, the group or community determines which behaviours are right or wrong, and a child learns these rules. In this way, a child raised in a cannibalistic society will internalise canniba- lism as an acceptable form of behaviour. A second view of moral development is presented by psy- choanalytical theory. This perspective, which is primarily in- teractionist, considers the in£uences of interpersonal, familial and socio-cultural factors on individual development includ- ing his or her moral development (Tice,1980). Society is seen to dictate which behaviours are acceptable and which are not Journal of Industrial Psychology, 2001, 27(3),61-67 Tydskrif vir Bedryfsielkunde, 2001, 27(3),61-67 61 TABLE 1 LEVELS AND STAGES OF MORAL REASONING ACCORDING TO KOHLBERG PRECONVENTIONAL LEVEL At this level the individual is sensitive to cultural rules regarding what is considered right and wrong.The individual will, however, interpret right and wrong in terms of the physical consequences of his or her actions (punishment, reward, exchange of favours) or in terms of the physical power of those who enunciate the rules.This level is divided into two stages: Stage 1: Heteronomous Morality This stage is characterised by na|«ve moral realism, with the moral importance of an action regarded as an inherent feature of the action. At this stage a person regards stealing as wrong because‘‘one is not supposed to steal.’’There is very little or a complete absence of moral reasoning. Punishment plays an important role at this stage because it is associated with a‘wrong’action rather than the person pragmatically avoiding the negative consequences of his or her actions. Stage 2: Individualistic, Instrumental Morality This stage is characterised by a concrete, individualistic perspective, implying that what is regarded as right or wrong is determined by the reward value of the behaviour. Other people’s needs do play a role, but only in terms of an exchange; only when a favour has been shown, will it be returned. CONVENTIONAL LEVEL At this stage the individual regards it as important to meet the expectations of his or her family, group or community, regardless of the immediate consequences.The attitude is one of conformity and loyalty to the expectations of the social order.The social order is actively maintained, supported and justi¢ed by those responsible.This level encompasses two stages. Stage 3: Interpersonally Normative Morality Kohlberg believes that this stage represents ‘correct’ behaviour which pleases people and of which they approve.The emphasis falls on conformity and stereotypical ideas accepted by the majority of people. Stage 4: Social System Morality Stage 4 is characterised bydoing one’s duty, the exercising of authority and the maintenance of the social order as part of society. At this stage the viewpoint is no longer restricted to the individual’s own familyand the common good plays an important role in moral decisions. Consequently the rules of society are accepted in a fairly uncritical way and observed as instruments of good order. POSTCONVENTIONAL LEVEL At this stage a clear e¡ort is made to de¢ne the values and principles of a moral society.These values and principles are upheld without necessarily enjoying the approval of the groups or persons who advocate these principles and apart from the individual’s own identi¢cation with these groups.This level also has two stages. Stage 5: Human Rights and Social Welfare Morality Kohlberg believes that, in terms of this stage, right and wrong are determined democratically.They are matters of personal values and opinion based on universal values and rights that any person would want to incorporate into a moral society.The validity of existing laws and social systems can be evaluated in the light of these human rights and values.This can be done because, even though laws and rules guarantee the rights of the individual and society, individual rights at times transcend laws and rules if the laws and rules seem destructive and unjust. Such unenforceable rules and laws should therefore not be obeyed at all costs, but can be changed in a democratic manner. Stage 6: Morality of Universalisable, Reversible, and Prescriptive General Ethical Principles Stage 6 describes the ideal relationship between free, autonomous individuals. It involves an impartial consideration of the circumstances or case of each person a¡ected by a moral decision. Justice, impartiality and reciprocity are guaranteed during the decision-making. (from Daniels,1984; Jordaan and Jordaan,1992; Kohlberg and Kramer,1969; Kohlberg,1973; Kohlberg,1976; Kohlberg,1980; Kohlberg,1988). (Jordaan & Jordaan, 1992). Although this approach is seem- ingly similar to that of social learning theory, Hugo and Van Vuuren (1995) highlight the di¡erence between them. In the psychoanalytic view, learning is regarded as static: but accord- ing to social learning theory it is an ongoing process. Thirdly, moral reasoning is considered from a cognitive point of view. According to this framework, the individual’s internal pro- cesses are just as important for development as the environ- mental context (Glassman, 1979). The emphasis no longer falls on the process of social conditioning, but on rational consi- derations as to what is right or wrong (Penn & Collier, 1985). These rational considerations are seen as successive, clearly dis- tinguishable patterns of reasoning which form a developmental sequence, fromconcrete and egocentric to abstract and universal. The cognitive approach has its origins in the work of Piaget (Penn & Collier,1985). Piaget proposed that cognitive processes formcoherent systems that adapt to changingenvironmental sti- muli.Thus, the human mind does not merelyabsorb discrete da- ta with which it comes into contact, but endeavours to organise it (Piaget,1950). As a result, individuals develop a need to obtain relevant information fromthe environment to create a meaning- ful system that will foster interaction with the world (Penn & Collier, 1985). The type of information an individual obtains from the environment is according to Damon (1980) a function of the individual’s current knowledge and abilities. As a result of the individual’s continous organisation of changing environ- mental stimuli, new forms of behaviour emerge. Piaget never proceeded with his early work on morality and his stage-theory of moral development was later extended by Lawrence Kohlberg. Like other structuralists, Kohlberg (1976) was not interested in moral behaviour, as such, but rather in the way moral decision-making and moral reasoning take pla- ce. According to Kohlberg, an individual’s level of cognitive development will determine the constraints on his or her pro- gression through the stages of moral development (Kohl- berg,1976; Kohlberg,1980; Rest,1986; Rest,1988;Windmiller, 1980). In the light of the relationship between cognition and morality Windmiller (1980) de¢nes morality as the di¡erence between that which is right and wrong, where‘right’refers to a universal truth. Consequently moral development is not pe- culiar to a speci¢c culture, but inherent in all cultures (Ey- senck, 1994). From this point of view the supposition that moral development constitutes the internalisation of cultural norms is rejected. By far the most in£uential theory of the three already men- tioned is the cognitive theory of learning or structuralist ap- proach (Jordaan & Jordaan, 1992). According to Turiel (1980), Piaget initially suggested that moral development consists moral development is not peculiar to a speci¢c culture, but in- herent in all cultures (Eysenck,1994). From this point of view the supposition that moral development constitutes the inter- nalisation of cultural norms is rejected. By far the most in£uential theory of the three already men- tioned is the cognitive theory of learning or structuralist ap- proach (Jordaan & Jordaan, 1992). According to Turiel (1980), Piaget initially suggested that moral development consists lte- rable physical laws rooted in nature. As the child develops, his or her morality will change from a form of one-sided respect to a morality of co-operation and mutual respect (the autono- mous level).Turiel (1980) believes that this level is founded on concepts like reciprocity and equality. Rules are no longer re- garded as mysterious or unalterable but rather as the outcome of a mutual agreement supporting the objectives of co-opera- tion and, consequently, as negotiable and adaptable. Central to the above-mentioned, is the development of morality along a continuum, ranging from an external orientation to an inter- nal orientation (Nichols & Day,1982). Kohlberg revised Piaget’s‘two-stage theory’and extended it to six stages of moral development (Turiel, 1980). An outline of the stages is contained inTable 1. WHITE, CRAFFORD, SCHEPERS62 The de¢nite sequence of six stages is characterised by distinct di¡erences in structure, and a hierarchical integration of lower stages into higher stages of reasoning. Stages of moral devel- opment are quanti¢able, and a few measuring instruments ba- sed on Kohlberg’s theory will now be discussed. Instruments designed to measure moral reasoning In developing his theory of moral reasoning Kohlberg (1980) used a speci¢c research procedure. He presented boys, aged ten to sixteen years, from di¡erent cultures, with stories posing moral dilemmas. These boys had to resolve the dilemmas supplying reasons for their decisions and answering certain questions. On the basis of the type of reasoning used by the boys, Kohlberg (1973) postulated his pattern of six quali- tatively di¡erent stages of moral reasoning. He found that no stage is omitted and that approximately 50% of an individual’s moral decisions can be classi¢ed in a particular stage, which indicates the individual’s current level of moral reasoning (Kohlberg,1973; McGeorge,1974).The remaining moral deci- sions made by the individual fall in the previous and subse- quent stages, which indicates that there is a continuous process of moral development (McGeorge, 1974). If, for in- stance, an individual uses Stage 2 reasoning 70% of the time, and Stage1reasoning 30% of the time, his or her level of moral reasoning would be classi¢ed as Stage 2. The dominant stage inwhich an individual’s responses are classi¢ed, is known as his or her global score. An obvious disadvantage of this procedure is that an individual is con¢ned to one stage of moral reaso- ning. On account of this restriction Kohlberg expanded his theory so that an individual could function in more than one stage simultaneously (McGeorge,1974). A second measuring instrument, designed by McGeorge, is called the‘‘Moral Maturity Scale’’(McGeorge,1974). A system of weights is used in scoring: the ordinal positions of the sta- ges reached are multiplied with the percentage of time the stage is used.The products are then summed to give the ¢nal score. According to this method of scoring the individual in the above-mentioned example would obtain a score of 170 if stage two reasoning was used 70% of the time and stage one reasoning 30% of the time (2 x 70 + 1 x 30 = 170). If the in- dividual used stage two reasoning100% of the time, the score would amount to 200. It is clear that the Moral Maturity Sca- le yields a score that is more sensitive than Kohlberg’s global score. According to Nichols and Day (1982) the development of the ‘‘De¢ning Issues Test’’ (‘‘DIT’’) by Rest (1974) ^ also based on the work of Kohlberg ^ constitutes a further milestone in mea- suring levels of moral reasoning. In the light of the complexity of the criteria used by Kohlberg, the‘‘DIT’’was seen as an impor- tant breakthrough. Respondents ¢nd recognition exercises such as the‘‘DIT’’easier to complete than, forexample, the production exercises used by Kohlberg (Rest,1986).The questions are pres- ented in multiple-choice format which rules out the problem of moral reasoning being restricted by an individual’s inability to verbalise his or her thoughts. Rest (1986) explains that the ‘‘DIT’’requires the respondent to select answers from three given possibilities relating to moral dilemmas. Next the respondent has to select twelve possible reasons for each answer and rank the four most important rea- sons (Davidson & Robbins,1978; Nichols & Day,1982). A further study undertaken to obtain a more valid repre- sentation of an individual’s level of moral reasoning was that of Taylor (1977). In an attempt to overcome the shortcomings of Kohlberg’s theoryTaylor developed a new questionnaire, the Reasons for Actions Questionnaire. It was designed in such a way that moral arguments relating to four stages, (namely Stages 2 5 of Kohlberg’s stages of moral reasoning), could be formulated.This meant that an individual could re- spond to questions relating to any of these four stages with equal ease. Taylor (1977), however, departed from open-en- ded questions like those used by Kohlberg, because they are unsatisfactory for measuring moral reasoning. Open-ended responses are usually so short and ambiguous that they make evaluation all but impossible. Consequently, Taylor (1977) made use of a ¢xed response format. Taylor (1977) refuted the supposition that a ¢xed response for- mat would lead to respondents arguing on higher levels of moral reasoning. He found that lower levels of moral reason- ing elicited the same proportion of responses. However, in Hugo and Van Vuuren’s (1995) research, where Taylor’s ¢xed response format was used, a large number of the respondents were categorised in the ¢fth stage of moral reasoning. Furthermore,Taylor’s instrument is ipsative in nature. It is thus impossible to use the Reasons forActions Questionnaire tocom- pare di¡erent individuals at the same stage of moral reasoning. The reason being that ipsative scores are systematically in£uen- ced byother scores and each individual’s scores produce the same mean and standard deviation, thus providing no standard against which to compare them.The most important limitation of ipsa- tive measurement is that parametric statistics cannot be used be- cause ipsative measurements are not independent of one another and they lead to a preponderance of negative intercorrelations. Kerlinger (1986) points out that most statistical tests are based on the assumption that the elements used in statistical tests are independentofone another. If this is not the case, aswith ipsative measurements, an analysis of correlations, for example factor analysis, could lead to misleading results on account of the spu- rious negative intercorrelations. In the light of the above-men- tioned it must be pointed out that all the available instruments for the measurement of moral reasoning are at least partly ipsati- ve in nature.The aim of this study was therefore to construct a scale of moral reasoning with normative rather than ipsative pro- perties. In the construction of scales, questions can be used, or positive or negative statements can be made. Likert scales involve state- ments, and the respondents have to indicate the extent to which they agree or disagree with the statements. According to Schepers (1992) a seven-point scale is less restrictive on the variance of a scale, yet easy to respond to. Consequently, in the construction of the new measuring instrument a seven-point scale was chosen. Every ‘‘reason for action’’ is linked to a seven- point scale varying from‘‘do not agree at all’’to‘‘fully agree’’. METHOD Sample A sample of 426 was drawn from undergraduate students in Industrial Psychology and Personnel Management, at two universities in the Gauteng area. An analysis of the sample indicated that 54% were Afrikaans-speaking and 29% Eng- lish-speaking, while 13% recorded an African language and 4% another language as their mother tongue. All the stu- dents in the relevant classes (apart from those who were ab- sent) were selected in order to ensure the randomness of the sample. Measuring instruments The Moral Reasoning Questionnaire (MRQ) Since only ipsative and semi-ipsative instruments are available for the measurement of moral reasoning, a normative scale was constructed and relevant data collected. In preparation for the construction of the scale the most recent literature was studied in order to evaluate the available measuring instruments and to identify typical behaviour associated with each level of rea- soning. Kohlberg’s theory was used as the foundation because it is by far the most comprehensive theory of moral reasoning. Each of the six stages as described by Kohlberg was analysed with the intent of identifying concepts associated with each speci¢c stage.The next step involved identifying dilemmas af- fecting under-graduate students. The dilemmas addressed the following issues: theft, adultery, dishonesty, fraud and integri- ty. A normative scale was decided upon, after which items we- re designed and a questionnaire developed. A moral dilemma, for instance a potential theft, was described and six reasons for conduct (based on Kohlberg’s six stages of moral 63MORAL REASONING TABLE 2 SAMPLE ITEM You work in the marketing section of a large banking group and attend a seminar with other marketing experts. One of the people attending the seminar is the marketing director of your greatest competitor. You overhear him discussing his bank’s marketing strategy with his col- leagues and notice that he has left his briefcase open in which lies a copy of the marketing strategy. Since his back is turned, you have an opportunity to remove the document that will give you important information about your competitor’s marketing strategy.You do not take the copy because: 1. You would not want him to steal your marketing plan and what would everyone think of you, if you did. strongly disagree strongly agree 2. You consider the possible harmful e¡ects of your action on the marketing director and ot- her people in the company concerned and decide for this reason to leave the document untouched. strongly disagree strongly agree 3. If you take the copy there is a very good chance that you will be caught and upset the mar- keting manager so badly, that it could cost you your career. strongly disagree strongly agree 4. You risk possible prosecution because such behaviour is against the law. strongly disagree strongly agree 5. If the marketing director sees you, he will expose you as a thief. strongly disagree strongly agree 6. You believe that you, as an individual, have no right to take information belonging to anot- her individual or business, without their consent. strongly disagree strongly agree 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 TABLE 3 MATRIX OF INTERCORRELATIONS OF SUBTESTS OF THE MORAL REASONING QUESTIONNAIRE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 SUBTEST 1 1,000 SUBTEST 2 0,083 1,000 SUBTEST 3 0,321 0,089 1,000 SUBTEST 4 -0,229 -0,133 -0,087 1,000 SUBTEST 5 0,229 0,038 0,235 -0,179 1,000 SUBTEST 6 0,214 0,284 0,173 -0,156 0,181 1,000 SUBTEST 7 0,234 0,014 0,248 -0,127 0,337 0,222 1,000 SUBTEST 8 -0,201 -0,143 -0,238 0,121 -0,258 -0,241 -0,265 1,000 SUBTEST 9 0,139 0,045 0,100 -0,163 0,218 0,141 0,183 -0,246 1,000 SUBTEST 10 0,237 0,092 0,173 -0,246 0,248 0,143 0,191 -0,221 0,294 1,000 SUBTEST 11 -0,252 -0,113 -0,198 0,264 -0,269 -0,243 -0,269 0,301 -0,345 -0,345 1,000 SUBTEST 12 -0,070 -0,179 0,019 0,121 0,075 -0,077 -0,053 0,090 -0,008 -0,018 0,049 1,000 SUBTEST 13 -0,267 -0,071 -0,201 0,354 -0,255 -0,229 -0,160 0,152 -0,249 -0,274 0,298 -0,016 1,000 SUBTEST 14 0,188 0,116 0,049 -0,279 0,216 0,170 0,120 -0,179 0,238 0,322 -0,303 -0,005 -0,283 1,000 SUBTEST 15 0,067 -0,199 0,106 -0,017 0,122 -0,023 0,130 -0,080 0,141 0,102 -0,093 0,099 -0,146 0,008 1,000 SUBTEST 16 0,103 0,154 0,043 -0,059 0,092 0,134 0,082 -0,086 0,110 0,038 -0,067 0,080 -0,050 0,139 -0,151 1,000 SUBTEST 17 -0,289 0,117 -0,206 0,192 -0,190 -0,002 -0,144 0,187 -0,126 -0,212 0,255 0,114 0,224 -0,194 -0,182 -0,001 1,000 SUBTEST 18 -0,158 -0,235 -0,128 0,139 -0,015 -0,162 -0,071 0,192 -0,027 -0,056 0,096 0,198 0,114 -0,035 0,117 -0,114 -0,021 1,000 SUBTEST 19 0,120 0,276 0,005 -0,18 0,116 0,185 0,111 -0,-58 0,198 0,174 -0,163 -0,043 0,090 0,253 -0,200 0,282 0,036 -0,091 1,000 reasoning) were provided. SeeTable 2 for an example of one of the dilemmas.The respondents were expected to evaluate each of the six reasons on a seven-point scale ranging from ‘‘fully agree’’to‘‘don’t agree at all’’. As the questionnaire con- sisted of ¢fteen moral dilemmas each involving six possible reasons for action it contained a total of 90 items.The dilem- mas are described brie£y and various reasons (each re£ecting a particular stage) are provided to explain why the actor cho- se not to act improperly.This format was speci¢cally chosen to diminish the likelihood of socially desirable responses. The questionnaire was then given to three experts with the re- quest that they evaluate the given responses in terms of Kohl- berg’s (1985) di¡erent levels of moral reasoning i.e. they were asked to indicate to what extent they agreed that a speci¢c item actually represented a speci¢c level of moral reasoning.Ten other people were then requested to identify all the items that were di⁄cult to respond to and to assist in editing them. Social desirability The newly constructed instrument uses a ¢xed response for- mat. A disadvantage of this format is that it can lead to the se- lection of higher-level responses that seem socially desirable rather than truthful (Taylor, 1977; Hugo & VanVuuren, 1995). To control for the possibility of obtaining socially desirable responses, the ‘‘Marlowe-Crowne Social Desirability Scale’’ was used. This instrument was developed by Crowne and Marlowe, and consists of 33 items, each coupled to a true or false option (Crowne & Marlowe,1960).The reliability of the instrument is 0,88 and the test-retest reliability is 0,89. Procedure The Moral Reasoning Questionnaire and the‘‘Marlowe-Crow- ne Social Desirability Scale’’were administered to 426 undergra- duate students. The goal of the study was explained to the students before the questionnaires were handed out. There was no time restriction on the completion of the questionnaires. RESULTS Factor analysis To counteract the e¡ect of di¡erential skewness (of the test items), the following procedure was followed in the factor analysis of the Moral Reasoning Questionnaire (MRQ). First the 90 items of the MRQ were inter-correlated and the matrix of intercorrelations was subjected to a principal factor analysis. The unreduced intercorrelation matrix yielded 19 eigenvalues greater than unity. Accordingly 19 factors were extracted (Kaiser,1961). Next, the obtained factor matrix was rotated to simple structure with the aid of aVarimax rotation. Following this, subtests were formed, by summing all the scores of items with high loadings on a factor. In this way, 16 subtests were formed.Three of the factors were eliminated because the item loadings were very low. Next the 16 sub-tests were intercor- related and the matrix of intercorrelations was subjected to a principal factor analysis. The unreduced intercorrelation matrix of the subtests yielded two eigenvalues greater than unity. Accordingly two factors were extracted.The obtained factor matrix was rotated to sim- ple structure with the aid of a Direct Oblimin rotation. The matrix of intercorrelations is given inTable 3 and the eigenva- lues inTable 4.The rotated factor matrix is given inTable 5. WHITE, CRAFFORD, SCHEPERS64 TABLE 4 EIGENVALUES OF UNREDUCED INTERCORRELATION MATRIX ROOT EIGENVALUES 1 *7,340 2 *1,290 3 *0,922 4 *0,815 5 *0,728 6 *0,676 7 *0,565 8 *0,554 9 *0,524 10 *0,482 11 *0,410 12 *0,395 13 *0,387 14 *0,339 15 *0,305 16 *0,270 TRACE 16,000 TABLE 5 ROTATED FACTOR MATRIX (DIRECT OBLIMIN ROTATION) ITEMS FACTOR I FACTOR II h2j SUBTEST 1 2.4, 4.6,5.1,6.5,7.4,7.5, 8.3, 8.4, 8.5, 9.6, 12.1, 12.2, 12.3 13.5,14.3,14.4,14.6,15.2,15.3,15.6 0,099 0,650 0,518 SUBTEST 2 10.1,10.2,10.3,10.4,10.5,10.6 0,118 0,538 0,388 SUBTEST 3 3.1, 3.2, 3.3, 3.4, 3.5, 3.6 0,190 0,420 0,318 SUBTEST 4 8.6,12.6,13.1,13.2,13.4,13.6,15.5 0,759 0,014 0,591 SUBTEST 5 5.2, 5.3, 5.4, 5.5, 5.6 0,602 0,060 0,414 SUBTEST 6 6.1, 6.2, 6.3, 6.4, 6.6 0,422 0,406 0,571 SUBTEST 7 2.1, 2.2, 2.3, 2.5, 2.6 0,719 -0,107 0,426 SUBTEST 8 4.1, 4.2, 4.3, 4.4, 4.5 0,658 -0,012 0,422 SUBTEST 9 9.1, 9.2, 9.3, 9.4, 9.5 0,006 0,800 0,647 SUBTEST 10 11.1,11.2,11.4,11.5,11.6 0,318 0,466 0,515 SUBTEST 11 7.1, 7.2, 7.3, 7.6 0,243 0,511 0,485 SUBTEST 12 12.4,12.5,15.1,15.4 0,493 0,322 0,558 SUBTEST 13 1.1,1.2,1.3,1.4,1.5,1.6 0,420 0,181 0,310 SUBTEST 14 11.3,13.3 -0,176 0,782 0,459 SUBTEST 15 14.1,14.2,14.5 0,426 0,374 0,533 SUBTEST 16 8.1, 8.2 0,642 0,037 0,445 Items per factor 42 48 TABLE 6 INTERCORRELATIONS OF FACTORS FACTOR I FACTOR II FACTOR I 1,000 FACTOR II 0,665 1,000 TABLE 7 ITEM STATISTICS IN RESPECT OF SCALE I Item Xg Sg rgxSg rgx Q1.1 5,514 1,693 0,857 0,506 Q1.2 5,355 1,663 0,786 0,472 Q1.3 5,730 1,572 0,616 0,392 Q1.4 5,889 1,436 0,598 0,417 Q1.5 5,686 1,551 0,550 0,355 Q1.6 6,007 1,606 0,808 0,503 Q2.1 5,916 1,383 0,700 0,506 Q2.2 6,200 1,156 0,587 0,508 Q2.3 5,784 1,505 0,837 0,556 Q2.5 6,125 1,318 0,799 0,606 Q2.6 5,425 1,525 0,597 0,391 Q4.1 5,714 1,488 0,632 0,425 Q4.2 5,875 1,298 0,572 0,444 Q4.3 6,025 1,267 0,684 0,540 Q4.4 5,911 1,452 0,792 0,545 Q4.5 6,193 1,230 0,721 0,586 Q5.2 5,536 1,552 0,704 0,545 Q5.3 6,082 1,368 0,780 0,570 Q5.4 5,484 0,655 1,068 0,645 Q5.5 5,118 1,781 0,947 0,532 Q5.6 6,002 1,322 0,796 0,602 Q6.1 5,125 1,698 0,963 0,567 Q6.2 5,464 1,466 0,881 0,601 Q6.3 5,770 1,309 0,738 0,564 Q6.4 5,409 1,559 0,922 0,592 Q6.6 5,711 1,511 0,956 0,633 Q8.1 6,186 1,166 0,691 0,593 Q8.2 5,941 1,440 0,760 0,528 Q8.6 6,227 1,222 0,610 0,499 Q12.4 6,141 1,430 0,818 0,572 Q12.5 6,382 1,190 0,736 0,618 Q12.6 6,509 1,052 0,501 0,476 Q13.1 6,568 1,024 0,586 0,572 Q13.2 6,245 1,466 0,703 0,480 Q13.4 5,916 1,541 0,842 0,547 Q13.6 6,505 1,001 0,554 0,554 Q14.1 5,509 1,604 0,916 0,571 Q14.2 5,716 1,509 0,960 0,636 Q14.5 6,141 1,378 0,872 0,633 Q15.1 5,352 1,805 0,835 0,463 Q15.4 5,970 1,537 0,873 0,568 Q15.5 6,495 1,080 0,662 0,613 rgxSg = Index of reliability of item g rgx=Correlation of item g with total Sg= Standard deviation of item g 8g=Mean of item g TABLE 8 ITEMS OF SCALES I AND II CATEGORISEDACCORDING TO LEVEL OF MORAL REASONING Level of Items of Items of moral reasoning Scale I Scale II 1 10 5 2 1 14 3 5 10 4 6 9 5 10 5 6 10 5 42 48 From an inspection of Table 5 it appears that both factors are well determined, with 42 items loading on Factor I and 48 on Factor II. The inter-correlation between the two factors (r=0,665) is given inTable 6. Item analysis Next, two scales were formed by assigning all the items with high loadings on Factor I to Scale I and all the items with high loadings on Factor II to Scale II. Following this, the two scales were subjected to item analysis.The NP50 programme of the National Institute for Personnel Research was used for this purpose. None of the items of Scale I were rejected, and a reliability coe⁄cient of 0,936 was obtained (according to Cronbach’s coe⁄cient alpha).The item statistics of Scale I are given in Ta- ble 7.The item-means range from 5,118 to 6,568 and the stan- dard deviations of the items from1,001to 1,805.The item-test correlations range from 0,355 to 0,645 and the indices of relia- bility vary from 0,501to1,068. A further analysis of the Scale I items reveals that Stages1, 5 and 6 of Kohlberg’s moral-judge- ment theory each appears ten times in Scale I, and that Stages 2, 3 and 4 appear considerablyless frequently (seeTable 8).The content of the items thus mainly concern Stages 1, 5 and 6 of moral reasoning. Scale 1 is characterised by principled reaso- ning (Stages 5 and 6) as well as intuitively principled reactions (principled intuition) (Stage 1), which probably involves very little moral reasoning in comparison with Stages 5 and 6. De- cisions concerning correct behaviour in Scale I are arrived at by intuitive knowledge of what is right (Stage 1) and an un- derstanding of the principles of human rights, justice and im- partiality (Stages 5 & 6). Consequently, Scale I was identi¢ed as Principled Moral Reasoning. None of the Scale II items were rejected during the iteration process, and a reliability coe⁄cient of 0,947 (according to Cronbach’s coe⁄cient alpha) was obtained.The item statistics regarding Scale II are given inTable 9.The item-means range from 3,714 to 6,270 and the standard deviations of the items from 1,186 to 2,011. The item-test correlations range from 0,375 to 0,657 and the indices of reliability vary from 0,587 to 1,144. A further analysis of the Scale II items reveals that Stage 2 of Kohlberg’s moral-judgement theory appears fourteen ti- mes and Stages 3 and 4 ten and nine times respectively, and that Stages 1, 5 and 6 appear ¢ve times each (see Table 8). The 65MORAL REASONING TABLE 9 ITEM STATISTICS IN RESPECT OF SCALE II Items Xg Sg rgxSg rgx Q2.4 5,109 1,661 0,624 0,375 Q3.1 5,045 1,790 0,793 0,443 Q3.2 4,934 1,793 0,802 0,447 Q3.3 5,395 1,691 0,905 0,535 Q3.4 3,714 1,873 1,020 0,544 Q3.5 4,811 1,957 0,860 0,440 Q3.6 5,102 1,890 0,807 0,427 Q4.6 5,491 1,693 0,721 0,426 Q5.1 4,580 1,807 0,784 0,434 Q6.5 5,068 1,719 0,769 0,447 Q7.1 5,011 1,981 1,021 0,516 Q7.2 5,698 1,740 0,841 0,484 Q7.3 5,350 1,700 0,986 0,580 Q7.4 6,014 1,462 0,712 0,487 Q7.5 5,270 1,770 1,039 0,587 Q7.6 5,875 1,488 0,928 0,624 Q8.3 5,409 1,596 0,899 0,563 Q8.4 4,898 1,665 0,988 0,594 Q8.5 5,173 1,636 0,969 0,592 Q9.1 5,650 1,638 0,872 0,532 Q9.2 4,011 1,867 1,144 0,613 Q9.3 4,650 1,785 1,050 0,588 Q9.4 5,123 1,730 1,128 0,652 Q9.5 5,302 1,588 0,909 0,573 Q9.6 5,143 1,772 1,141 0,644 Q10.1 4,573 2,010 0,991 0,493 Q10.2 4,745 2,011 1,033 0,514 Q10.3 5,964 1,595 0,793 0,497 Q10.4 5,773 1,683 0,898 0,534 Q10.5 5,855 1,611 0,998 0,620 Q10.6 6,045 1,475 0,800 0,543 Q11.1 5,345 1,660 0,939 0,566 Q11.2 5,530 1,468 0,723 0,493 Q11.3 4,395 1,847 1,112 0,602 Q11.4 5,359 1,591 0,791 0,497 Q11.5 5,302 1,724 0,892 0,518 Q11.6 6,134 1,186 0,587 0,495 Q12.1 6,077 1,393 0,777 0,558 Q12.2 5,893 1,500 0,986 0,657 Q12.3 5,518 1,761 0,974 0,553 Q13.3 4,898 1,966 1,026 0,522 Q13.5 5,759 1,677 0,856 0,510 Q14.3 4,811 1,802 1,133 0,629 Q14.4 5,905 1,343 0,871 0,648 Q14.6 5,507 1,704 1,066 0,626 Q15.2 5,695 1,653 0,887 0,537 Q15.3 6,270 1,225 0,731 0,597 Q15.6 6,025 1,459 0,791 0,542 rgxSg = Index of reliability of item g rgx = Correlation of item g with total Sg = Standard deviation of item g 8g = Mean of item g TABLE 10 CORRELATIONS OF THE MARLOWE-CROWNE SOCIAL DESIRABILITY SCALE WITH SCALES I AND II OF THE MRQ (MORAL REASONING QUESTIONNAIRE) Scale I Scale 5% 6% MARLOWE-CROWNE 0,227 0,247 5% 6% content of the items therefore relates mainly to Stages 2, 3 and 4 of moral reasoning. It seems that people who reason on these levels do so in accordance with prescriptions from authority ¢- gures, legislation, social expectations or to avoid punishment. Their moral reasoning and associated behaviour are thus regula- ted and prescribed by external sources; consequently, the scale was identi¢ed as Prescriptive Moral Reasoning. The two scales of the Moral Reasoning Questionnaire were furthermore correlated with the scores of the ‘‘Marlowe- Crowne Social Desirability Scale’’. From the matrix in Table 10 it is evident that there are statistically signi¢cant positive correlations between Principled Moral Reasoning and the ‘‘Marlowe-Crowne Social Desirability Scale’’. The magnitude of the correlations is, however, small (0,227 to 0,247). Social desirability thus seems to play a minor role in the Moral Rea- soning Questionnaire. DISCUSSION The results obtained in the present study yielded two scales of moral reasoning. These have been called Principled Moral Reasoning and Prescriptive Moral Reasoning and conform to Piaget’s two-stage theory rather than Kohlberg’s six-stage theory of moral reasoning. As has been pointed out, Piaget distinguished between a morality of co-operation and a mora- lityof limitation. A few points of comparison between Piaget’s theory and the scales of the Moral reasoning Questionnaire will now be discussed. Principled Moral Reasoning can be said to represent Stages1,5 and 6 of Kohlberg’s theory of moral reasoning and also bears resemblance to Piaget’s morality of co-operation. Piaget’s mo- rality of co-operation is similar to Kohlberg’s Stages 5 and 6 in the sense that the concepts of personal values and opinions, justice, impartiality and reciprocity are involved. Of interest, however, is that naive moral realism (Stage 1) forms part of principled moral reasoning, while its description resembles the morality of limitation. (According to the description of Stage1, the moral foundation of an action is regarded as an in- herent, unchangeable and absolute characteristic of the action.) A possible explanation for this deviation from Piaget’s theory is that naive moral realism basically involves intuition rather than moral reasoning or following instructions. Prescriptive moral reasoning (Stages 2, 3 and 4) corresponds to Piaget’s morality of limitation, where morality is regulated and dictated by the reward value of an action, the expectations of the family, group or community as well as rules laid down by society. Furthermore, in Piaget’s theory, the two stages fol- low one another but the scope of this study does not allow one to determine whether prescriptive and principled moral reaso- ning follow a hierarchical course or not. Principled moral reasoning seems to take place in two ways, namely in a naive-intuitive manner (Stage 1) and a principled, well-reasoned manner (Stages 5 and 6). The question arises whether principled moral reasoning, which rests on certain principles and intuition instead of external instruction, leads to a greater amount of consistency between the individual’s public and secret moral behaviour. It can also be asked whether a person whose moral behaviour is dictated by external in- struction, such as legislation or group norms, will behave in a consistent way, especially in the absence of instruction. It will require further research to answer questions of this nature.This kind of research would be important in the light of the intro- ductory discussion regarding £atter structures in organisations and the accompanying individual responsibility for moral be- haviour in the absence of formal regulation. Furthermore, the correlation between scores on the Principled Moral Reasoning Scale and the Prescriptive Moral Reasoning Scale is 0,665, making it possible that an individual might attain moderate to high scores on both of the scales. A possible reason for the overlap can be attributed to the fact that respondents are presentedwith reasons foraction associatedwith all the stages of moral reasoning.Thus, if a respondent reasoned at level ¢ve, this does not preclude reasoning at lower levels as well. For example a person may choose not cheat because of reasons of fairness (Stage 5) but would also choose not to violate the rules of the university (Stage 4). Norms were calculated and made available to determine the position of an individual relative to his or her norm group. The limited scope of the study does not make it possible, ho- wever, to determine the hierarchical course of development of the di¡erent stages. Using the Moral Reasoning Questionnaire to determine dif- ferent levels of moral reasoning of managers in organisations may be regarded as simplistic. In reality, moral behaviour is complex and one must take into account the relative nature thereof (De Klerk, 1991) as well as the fact that people behave in di¡erent ways without deliberately being unethical. WHITE, CRAFFORD, SCHEPERS66 This does not imply that a lack of moral behaviour should be accepted without question but that such instances require an attitude of tolerance rather than blind condemnation. In this regard, the Moral Reasoning Questionnaire provides a useful means forexploring the underlying reasons forbehaviour so as to promote greater understanding in the work place. For, des- pite the complex nature of moral behaviour, it is only those who work in organisations who can allow ethical behaviour to amount to more than a message on paper. REFERENCES Cadbury, A. (1987). Ethical managers make their own rules. Harvard Business Review,69-73. Crowne, D.P. & Marlowe, D. (1960). A new scale of social de- sirability independent of psychology. Journal of Consulting Psychology, 24(2), 349- 354. Damon, W. (1980). Structural-developmental theory and the study of moral development. In M.Windmiller, N. Lam- bert, E. Turiel (Eds.). Moral development and socialization. Boston, MA: Allyn & Bacon. Daniels, M. (1984).The relationship between moral development and self actualization. Journal of Moral Education,13(1), 25-30. Davidson, M.L. & Robbins,S. (1978). The reliability and vali- dity of objective indices of moral development. Applied Psychological Measurement, 2(3), 389-401. Dawson, L.M. (1995).Women and men, morality and ethics. Business Horizons, 38(4), 61-68. De Klerk, D. (1991). Ethical relativity. Entrepeneur,17. Derry, R. (1989). An empirical study of moral reasoning among managers. Journal of Business Ethics, 8, 855-862. Eysenck, H.J.(1994). How we acquire a sense of morality. Cri- minalJustice Ethics (13),53-61 Glassman,W.E.(1979). The Cognitive Approach. In J. Medcof & J. Roth (Eds.). to Psychology. Milton Keynes: Open Ap- proaches University Press. Hugo, A. & vanVuuren, L. J. (1995). Across-cultural analysis of work values and moral reasoning. Journal of Industrial Psy- chology, 22(1),12-18. Johns, T. (1995). Don’t be afraid of the moral maze. People Management, 1(20), 32-34. Jordaan, W. & Jordaan J. (1992). Mens in konteks. Johannes- burg: Lexicon. Kaiser, H.F. (1961). A note on Guttman’s lower bound for the number of common factors. BritishJournal for Statistical Psy- chology,14(1),1. Kerlinger, F.N. (1986). Foundations of behavioral research. Ja- pan: CBS Publishing. Kohlberg, L. & Kramer, R. (1969). Continuities and disconti- nuities in childhood and adult moral development. Hu- man Development,12, 93-120. Kohlberg, L. (1973). Stages and aging in moral development ^ some speculations.The Gerontologist,13(1), 497-502. Kohlberg, L.(1976). Moral stages and moralization. InT. Licko- na (Ed.). Moral development and behaviour. NewYork: Holt, Rinehart andWinston. Kohlberg, L.(1980).Stagesofdevelopment as abasis foreducation. In B. Munsey(Ed.). Moral development, moral education, and Kohlberg. Birmingham, Alabama: Religious Education Press. Kohlberg, L.(1988). Acurrent statement on some theoretical is- sues. In S. Modgil and C. Modgil (Eds.). Lawrence Kohl- berg : Controversy and consensus. Philadelphia, PA: The Falmer Press. McDonald, J.(1993). Thin examples of moral dilemmas. Social theory and practice,19(2), 225-235. McGeorge, C. (1974). Situational variation in level of moral judgement.The British Journal of Educational Psychology, 44, 116-122. Murphy, J.M. & Gilligan,C. (1980). Moral development in late adolescence and adulthood : a critique and reconstruction of Kohlberg’s theory. Human Development, 23, 77-104. Nichols, M.L. & Day,V.E. (1982). A comparison of moral rea- soning of groups and individuals on the ‘‘De¢ning Issues Test’’. Academy of Management Journal, 25(1), p.201-208. Penn,W.Y. & Collier, B.D. (1985). Current research in moral development as a decision support system. Journal of Busin- ess Ethics,4,131-136. Piaget, J. (1950). The moral judgement of the child. London: Routledge & Kegan Paul. Reilly, B.J. & Myroslaw, J.K. (1990). Ethical business and the ethical person. Business Horizons, 23-27. Rest, J.R. (1974). Major concepts in moral judgement develop- ment. Reportbased ondissertation bysameauthor. Minneapolis: University of Minnesota. Rest, J.R. (1986). Moral development: Advances in research and theory. NewYork: Praeger. Rest, J. R. (1988).Why does college promote development in moral judgement? Journal of Moral Education,17(3),183-194. Robertson, D. (1996).Workingat keeping up in the global race. SundayTimes BusinessTimes. 3 November. Rossouw, D. (1997). Etiese kodes van belang vir maatskappye. RAU Rapport. Sawyer, J.C.(1977). A factor analytical studyof the dimensional structure within stages of moral development and that structure’s relationship to speci¢c personality traits. Unpu- blished doctoral dissertation. University of Massachusetts. Schepers, J.M. (1992).Toetskonstruksie:Teorie en praktyk. Jo- hannesburg: RAU Press. Sieber, J.E. (1980). A social learning approach to morality. In M.Windmiller, N. Lambert & E.Turiel (Eds.). Moral deve- lopment and socialization. Boston, MA: Allyn & Bacon. Stead,W.E.,Worrell, D.L., & Stead, J.G. (1990). An integrative model for understanding and managing ethical behavior in business organizations. Journal of Business Ethics, 9, 233- 242. Stoker, D.J. (1990). Statistiese tabelle. Pretoria: Academica. Taylor, T.R. (1977). The assessment of morality. Unpublished masters’thesis. University of South Africa. Tice,T.N. (1980). A psychoanalytical perspective. In M.Wind- miller, N. Lambert & E.Turiel (Eds.). Moral development and socialization. Boston, MA: Allyn & Bacon. Trevino,L.K. (1986). Ethical decision making in organizations: A person-situation interactionist model. Academy of Mana- gement Review,11(3), 601-617. Tudin, P., Straker,G., & Mendolsohn, M. (1994). Social and political complexity and moral development. SouthAfrican Journal of Psychology, 24(3),163-168. Turiel, E. (1980). The development of social-convential and moral concepts. In M.Windmiller, N. Lambert & E.Turiel (Eds.). Moral development and socialization. Boston, MA: Al- lyn & Bacon. Windmiller, M. (1980). Introduction. In M. Windmiller, N. Lambert & E. Turiel (Eds.). Moral development and socializa- tion. Boston, MA: Allyn & Bacon. 67MORAL REASONING