Journal of Youth and Adolescence, Vol. 7, No. 4, 1978
Intrapersonal Variations in Stage of Moral Reasoning Among Adolescents as a Function of Situational Context James S. Leming 1
Received March 27, 1978
On a set o f classical moral dilemmas and on a set o f practical moral dilemmas 60 randomly selected eighth- and twelfth-grade students were asked to (a) judge the actions o f others (]udgrnent) and (b ) reach a personal decision and give reasons for what they would do (deliberation). The interviews were transcribed and scored for stage o f moral reasoning. Stage o f moral reasoning associated with deliberation on practical moral dilemmas was found to be significantly lower than their moral judgment on classical moral dilemmas. A n explanation was presented which combined social learning and cognitive-developmental moralization theory.
~TRODUCTION The recent review by Kurtines and Grief (1974) of the empirical basis of the cognitive developmental approach to the study of moral judgment has raised many serious questions concerning the reliability and validity of Kohlberg's moral reasoning typology. In their review Kurtines and Grief analyzed the construct validity of the Kohlberg instrument from two perspectives: the claim for the invariance of sequence of stages and the claim for moral development as representing a hierarchy of complexity. Also involved in the analysis of the construct validity of the Kohlbergian system, and not discussed in the Kurtines and
1Assistant Professor, Department of Curriculum, Instruction and Media, Southern Illinois University at Carbondale. Receivedhis Ph.D. from the University of Wisconsin at Madison. Research interests include the nature of moral character and the implications involved for the moral education curriculum. 405 0047-2891/78/1200-0405 $05.00/0 9 1978 Plenum Publishing Corporation
406
Leming
Grief (1974) review, is the assumption that moral reasoning represents a generalized cognitive orientation which is applied consistently across all situations. Kohlberg assumes that moral reasoning is not situationally specific behavior. Kohlberg (1971) discusses this dimension of his theory in the following manner: Stages def'me "structured wholes," total ways of thinking, not attitudes toward particular situations. (p. 169)... Moral stages constitute "principles" in the sense that they represent the major consistencies of moral evaluation within the individual not directly due to factual beliefs (Our evidence for this is that factor analysis indicates a single stage factor cutting across all moral situations, and all aspects of morality on which the individual is assessed) (p. 177). In other words a given stage response does not represent a specific response determined by knowledge and familiarity with that task or similar ones; rather, it represents an underlying thought organization. One would expect, therefore, that at a given point in developmental time, individuals will consistently use their highest attained level of moral reasoning to resolve moral conflict regardless of any variations in situational context. This expectation, entailed by the cognitivedevelopment theory, will form the basis of the research reported here. In order to inquire into this aspect of the Kohlberg theory it is necessary to first refer to some salient distinctions within the realm of moral reasoning. The first distinction is between the judgment mode and the deliberation mode of moral reasoning. Moral judgment will be defined as occurring where one tries to determine after someone else has acted whether that person has taken the best course open to him, for the purpose of praising or condemning him. Moral deliberation will be defined as moral reasoning where one tries, before acting, to determine which is the best course open to oneself with a view to entering on it. In other words, in the judgment mode of moral reasoning the individual is presented with a fait accompli and is asked to evaluate the actions involved. I n the deliberation mode of moral reasoning the individual is faced with an incomplete situation and his task is to choose the morally correct course of action for himself. A second crucial distinction centers on the differences between classical moral dilemmas and practical moral dilemmas. Classical moral dilemmas refer to situations which are removed from the life space of the individual and involve characters with whom the subject has trouble identifying. Practical moral dilemmas refer to situations within the life space of the individual and involving people and issues familiar to him. An analysis of Kohlberg's method of assessing stage of moral reasoning reveals that he is primarily measuring the individual's moral judgment applied to classical moral dilemmas. For example, in a typical dilemma a man (Heinz)has stolen a drug which will save his wife's life. The subject is then asked such questions as, Should Heinz have stolen the drug? Should Heinz be punished? Clearly, the moral reasoning measured under these conditions does not deal with issues or situations familiar to many people. On the basis of the conception that stage
Moral Reasoning and Situational Context
407
of moral reasoning represents "structured wholes," the cognitive developmentalists assume that moral judgment about classical dilemmas is similar to deliberation about practical dilemmas. This research will examine the empirical basis of this assumption. Specifically, this research attempts to answer the following two questions: (1) Is there a difference between the moral reasoning people use to justify actions they claim they would take and the moral reasoning they use in judging the actions of others? (2) Is there a difference between the moral reasoning people use in dealing with situations unfamiliar to their life space and the moral reasoning they use in situations familiar to their life space? METHOD
Instrument
The Kohlberg method of assessing stage of moral reasoning involves interviewing the subjects on moral dilemmas using a semistructured interview schedule, tape-recording the subject's responses, transcribing the tape recording, and finally scoring the transcript according to procedures developed by Kohlberg (1972). As a result of the scoring, the subjects are assigned to one of the six stages of moral development. In order to assess differences in moral reasoning caused by mode of reasoning and type of situation it was necessary to develop alternative moral dilemmas and interview schedules. The classical dilemmas used were Heinz, Joe and his father, and Alexander. A series of practical moral dilemmas was developed which contained moral conflict situations likely to be found within the life space of the prospective subjects. A questionnaire was given to 186 seventh- and twelfthgrade students asking them to identify or suggest moral conflict situations with which they were familiar. From the situations most frequently identified, a set of six pilot dilemmas was created. A final selection of three practical moral dilemmas was made on the basis of the results of a pilot study. A typical practical moral dilemma was the Party dilemma. In this dilemma a girl's parents have denied her permission to go to a friend's party. The girl's best friends are going to be at the party, so she tells her parents that she is going to a movie and goes to the party anyway. The other two practical moral dilemmas deal with the issues of cheating (the Assignment dilemma) and peer group conflict (the Group dilemma). An interview schedule was developed for the practical dilemmas to assess the subjects' stage of moral reasoning in the judgment mode. After each dilemma the subjects were asked to respond on a 5-point scale to the statement, "In my life situations like this one are familiar." Significantly more subjects (p < 0.05) were found to agree with this statement after discussing practical moral dilemmas than agreed with it after discussing classical moral dilemmas.
408
Leming
Care was taken to ensure that the practical moral dilemmas and interview protocols contained many of the same issues as the classical moral dilemmas and their interview protocols. The issues of affectional role obligations and contact and promise were present in both the classical and the practical moral dilemmas. Thus, any differences found in mean level of moral reasoning should reflect inconsistencies on the same moral issues across situations and should not be attributed solely to horizontal decalage, that is, individual differences resulting from one set of dilemmas including one set o f moral issues and the other set o f dilemmas containing a different set o f issues. In addition to measuring the subjects' moral reasoning in the judgment mode on the classical and practical dilemmas, it was also necessary to measure their moral reasoning in the deliberation mode on the same sets of dilemmas. To accomplish this it was necessary to reword both the classical and practical dilemmas, so that they were now stated in the present tense and the actual choice in the dilemmas was still open and unstated. For example, the Heinz dilemma had to be rewritten in such a way that the subject is asked to consider a situation where someone he loves is dying of cancer, he cannot raise the money, and the choice offered is whether or not he will steal the drug to save his loved one's life. The interviewer carefully avoided incorporating " y o u " questions into the judgment interview protocols. In sum, there were four sets o f dilemmas and interview schedules on which the subject's stage o f moral reasoning was assessed (see Table I). Subjects The sample consisted of 60 public school students randomly selected from two middle schools and two high schools in the Madison, Wisconsin, area. At the time o f the interviews 30 of the subjects had just completed seventh grade and 30 of the subjects had just completed eleventh grade. Equal numbers of boys and girls were present in the sample. Procedure
The interviews took place at two o f the local schools. The subjects were interviewed on 12 different dilemmas, 3 within each o f the four forms of moral Table I. Forms of Moral Reasoning
Components of form Form
Mode of moralreasoning
Type of dilemma
MJCMD MDCMD MJPMD MDPMD
Moral judgment (MJ) Moraldeliberation (MD) Moral judgment (MJ) Moraldeliberation (MD)
Classicalmoral dilemma (CMD) Classicalmoral dilemma (CMD) Practical moral dilemma (PMD) Practical moraldilemma (PMD)
Moral Reasoning and Situational Context
409
reasoning. The interviews took between two and three hours for each subject. Two 10-minute breaks were given, and the order in which the dilemmas was presented to each subject was randomly determined in order to eliminate any fatigue effect. Each interview was tape-recorded and then transcribed. Scoring The transcripts of the interviews were scored according to procedures outlined by Kohlberg (1972). Scoring was organized by form, which resulted in four stage scores for each subject. Stage of moral reasoning scores can be reported as either global scores or as mean moral maturity scores. The global score and the subject is classified as either a pure stage or a mixed stage. For ease of statistical computation, this study used the subject's mean moral maturity scores. The mean moral maturity score (MMS) is ascertained by identifying stage scorable responses by issues within the transcripts of individual dilemmas, Issues are defined by Kohlberg (1972) as "defining the concrete objects of concern or value to the subject in the situation. Secondly they are the things to be defined and chosen between in the situation, they define the moral conflict" (p. 18). For every issue stage score within a form a point value is then assigned. The point value is based on a ratio of 3:2:1, depending upon whether the stage score was circled (most salient issue for resolving the dilemma), uncircled (ascertained with a high degree of certainty), or question-marked (ascertained with a low degree of certainty). Next a percent score was figured for each stage present in the subject's reasoning based on the total points assigned. The percent score was then multiplied by the number representing the stage. When summed the results can yield scores ranging from 100 (100% at Stage 1) to 600 (100% at stage 6). In order to obtain a reliability score it was necessary to hire and train a graduate student in education. This reliability scorer evaluated the response of 10 randomly selected transcripts. A product-moment correlation coefficient was computed between the two scorers' MMS on the individual forms. Using this procedure the correlation coefficients for the 10 subjects on the separate forms 2 was MJCMD (0.88), MDCMD (0.79), MJPMD (0~ and MDPMD (0.80).
Statistical Analysis
A 2 • 2 • 2 X 2 analysis of variance permits examination of the data for significant differences by sex, year in school, mode of reasoning, and type of 2See Table I for explanation of the forms.
410
Leming
dilemma. Standard fixed effects analysis of variance statistical design requires independent assignment of scores to each cell. This research does not meet this condition, since individual subjects' scores were assigned repeatedly across the four factors identified. That is, there is a score for each subject in all the four forms of moral reasoning. Therefore, Kirk's multiple factors (1968) repeated measures split-plot design ( S P F - pr.gu) was used. This method of analysis allows one to answer the primary research questions concerning differences in moral reasoning between modes of reasoning and types of situations.
RESULTS
The analysis of variance found that two of the four main effects were statistically significant beyond the 0.05 level. Twelfth-grade subjects were found to be significantly higher in mean moral maturity scores than eighth-grade subjects, and mean moral maturity scores for all subjects were found to be significantly higher in the judgment mode than in the deliberation mode. Two of the interactions in the analysis of variance were statistically significant. First, the Table 1I. MMS (All Subjects) for Mode • Dilemma Interaction D i l e m ~ / Mode
Across CMD PMD dilemma
MJ MD Across mode
264.4 263.6 260.3 240.9 262.3 252.2
263.9 250.6 257.3
28O 270 260 MMS 250 240 230 CMD
PMD
Type of Dilemma
Fig. 1. Graph of interaction: Mode X Dilemma. Open triangle =M J; circle = MD.
Moral Reasoning and Situational Context
411
Table IIL MMS (All Subjects) for Age X Mode • Dilemma Mode (dilemma) Grade 8 12 Across grade
MJ
MD
CMD PMD
CMD PMD
Across mode (dilemma)
246.8 249.0 281.9 278.1 264.4 263.6
245.6 235.0 275.1 246.9 260.3 240.9
244.1 270.5 257.3
interaction between mode of reasoning and type of dilemma was found to be significant beyond the 0.001 level. An examination of Table II reveals that the interaction is due to substantially lower MMS for deliberation on practical moral dilemmas. Figure 1 presents this interaction graphically and demonstrates that the type of dilemma is only an important factor in one's moral reasoning within the deliberation mode. The interaction between mode o f moral reasoning and type o f dilemma can be further understood through an examination of the second significant interaction: Grade X Mode X Dilemma. The means associated with this interaction are reported in Table III. These means indicate that the most marked change in MMS occurs for both eighth- and twelfth-grade subjects engaged in the MDPMD form of moral reasoning. Across grade this drop in MMS is around 20-24 points. These means and hence the interaction can be more fully understood graphically. Figure 2 presents the graph of the Grade • Mode X Dilemma interaction. This graph shows that the Mode X Dilemma interaction is due to the twelfth-grade subjects' drop in MMS in moral deliberation within practical
290 280 270 MMS 260 250 240 230 MJCMD
MJPMD
MDCMD
MDPMD
Mode & Dilemma
Fig. 2. Graph of interaction: Grade • Mode • Dilemma. Open triangle = 8th grade; circle = 12th grade.
412
Leming
Table IV. Percent of Stage Usage on the Four Forms of Moral Reasoning by Twelfth-Grade Subjects Stage Form
1
2
3
4
5
MJCMD MJPMD MDCMD MDPMD
0 0 1 0
45 49 43 68
34 35 43 22
15 10 10 6
6 6 5 4
moral dilemmas, or MDPMD. The significant difference found between eighthand twelfth-grade subjects can now be interpreted as due predominantly to age differences on the MJCMD, MDCMD, and MDPMD forms of moral reasoning alone. A post hoc analysis utilizing the Sheff6 procedure showed that the difference between eighth- and twelfth-grade subjects on the MDPMD form of moral reasoning was not statistically significant (p < 0.05). The statistically significant difference between MMS on modes of moral reasoning is accounted for by the low mean scores found on moral deliberation within practical moral dilemmas. A further clarification of this significant drop in level of moral reasoning can be gained by examining subjects' percentage stage usage across all four forms of moral reasoning. It is evident from the data presented in Table IV that the drop in MMS on the MDPMP form of moral reasoning for twelfth-grade subjects is due largely to a decrease in Stage 3 usage and a concomitant increase of Stage 2 responses. In addition, an intercorrelation matrix between the-subjects moral reasoning on all four forms shows that the standard conception of moral reasoning as defined by Kohlberg (MJCMD) explains only 31% of the variance on deliberation on practical dilemmas (MDPMD). In summary, the analysis of variance reveals that (1) older subjects are significantly higher in mean stage of moral reasoning than younger subjects; (2) across all subjects, reasoning in the judgment mode is significantly higher than reasoning in the deliberation mode; (3) the interactions of Mode • Dilemma and Grade X Mode X Dilemma are statistically signifiant; and (4) the interaction effects are the result of the twelfth-grade subjects' lower MMS on the MDPMD form of moral reasoning.
DISCUSSION In order to fully understand the significance of these findings it is necessary to clarify more precisely the nature of the moral reasoning assessed in this study. Previous studies on the comprehension of moral statements by Rest (1973) and
Moral Reasoning and Situational Context
413
Rest et al. (1969) have suggested that individuals understand and therefore can communicate in the language of all stages below their own modal stage of moral reasoning as assessed by standard Kohlbergian procedures. When assessing an individual's stage of moral reasoning with the Kohlberg methodology, one is measuring the highest stage attained by the individual. Even though a subject may verbalize lower stages during the course of an interview, the key statements for assessing stage of moral reasoning are those in which he discusses what he sees as the best reasons for a given decision. In the process of being interviewed, many subjects supersede initial lower stage scorable responses with higher stage scorable responses. This happens as a result of careful probing by the interviewer which highlights the difficulties and complexities inherent in the intial responses and causes the subject to generate alternatives. Often the interviewer must ask the subject to choose between a number of alternatives discussed in the course of the interview and to select the best reason for a given moral decision. The reasons identified by the subject as best are the responses which are eventually stage scored because they are the statements which the subject sees as ultimately defining right or wrong. Although subjects may see the usefulness of stages lower than their own modal stage in resolving moral dilemmas, under skillful probing eventually these lower stage reasons are rejected for what are seen as preferable or better reasons. In this study, however, the twelfth-grade subjects spontaneously used and defended as best a lower mean level of moral reasoning in deliberation on practical moral dilemmas than they used in judging the right thing to do on classical moral dilemmas. A detailed analysis of the data revealed that this finding was not due to differences associated with intrapersonal variation from moral issue to moral issue. Few subjects ever reason consistently across all issues at the same stage and can be scored 100% at a given stage. That subjects in this study tended to use lower stages in the resolution of more familiar moral conflict situations is not an isolated finding. Gilligan et al. (1971) compared moral judgment of adolescents on sexual dilemmas with moral judgment on the standard Kohlberg dilemmas. They found that "Considering the sample as a whole, the level of moral reasoning was higher on the standard dilemmas than on the sexual dilemmas. The mean MMS for the standard stories was 345 compared with 315 on the sex stories" (p. 145). Gilligan etal. (1971) found that 50% used the same major stage on both sets of dilemmas, 40% used lower stages and 10% higher stages on the sex dilemmas. Haan's reexamination (1975) of the Haan e t al. (1968) data on college activism also found significant intrapersonal variation as a function of situational context. Haan (1975) found that two thirds of the college sample used a different modal stage of moral reasoning - 46% higher and 20% lower - when reasoning about the morality of actual campus unrest than they did when reasoning about Kohlberg's hypothetical (classical) moral dilemmas. Since all of the subjects in the Haan et al. (1968) study had recently been exposed to or had participated in
414
Leming
campus unrest, their higher scores on reasoning about the actual situation can be considered as a form of growth, since the experience was potentially disequiliberating. However, the fact that 20% in the Haan (1975) study, 31% of the subjects of the present study, and 40% of the subjects in the Gilligan et. al. (1971) study reasoned at a lower modal stage about a realistic situation than about a classical (hypothetical) situation calls for a more careful examination of the "structured wholes" concept. Specifically, an attempt must be made to explore why reasoning about familiar situations affects some individuals' ability to use their highest attained level of moral development in the resolution of moral conflict. This phenomenon of situationally specific intrapersonal variation in stage of moral reasoning suggests an interpretation of moralization that involves social learning dynamics in addition to cognitive development. Aronfreed (1968, 1976) has argued that the social environment plays a much more specific and determinant role in the moralization process than the cognitive developmentalist would have us believe. The discussion below builds upon Aronfreed's insights. Kohlberg has assumed that the child carries an intrinsically ordered incipient program of moral development which unfolds under the stimulation of social experience. Kohlberg has not viewed this social experience as the transmitter of the stages, for the stages are an underlying structure and not a set of social values. Rather he has considered the social experience as a generalized and indeterminant nutriment. The research presented above suggests that this view of moralization needs to be broadened to include the possibility of situationally specific stage responses to moral conflict which are resistant to broader shifts in cognitive maturity. The following explanation attempts to retain a developmental view of cognitive growth while accounting for the operation of social learning dynamics. The explanation summarized below is an attempt to account for why a percentage of a population (ranging from 20%-40%, depending upon the study cited) when reasoning about more practical and immediate situations use reasoning at a lower level of moral development- than they demonstrate when reasoning about more hypothetical moral situations. The following statements present this explanation: (a) Structural changes in moral thought are predicated upon more general shifts in cognitive development. That is, structural changes are not intrinsically moral; their moral status resides in the substantive base of the values upon which they operate. This substantive base is learned by the individual from the social environment. (b) A structural change in an individual's general cognitive reasoning does not always entail a parallel structural change in moral reasoning. General structural cognitive change is a necessary but not sufficient condition for structural change in moral reasoning. (c) New acquisitions in moral thought displace rather than transform the moral structures previously held.
Moral Reasoning and Situational Context
415
(d) Displacement of previous forms of moral thought by new forms of moral thought is a function of the context of the moral reasoning and the set of expectancies of reinforcement associated with the old and new forms of thought within that context. (e) Individuals operate in many different moral environments. Each of these environments contains different moral structures (reasons) and contents (choices). (f) The more hypothetical (removed from the life space of the individual) the context of the moral reasoning, the less likely the individual is to anticipate reinforcement (either negative or positive) as a result of adopting new forms of reasoning. As a result, anticipated negative reinforcements are less likely to serve as obstacles to the transition of new structures. (g) The more affectively proximate (within the individual's life space) the context of the moral reasoning, the more likely the moral structures are to exist within a set of expectancies of reinforcement. The individual adopts a particular pattern for resolving moral conflict because it has shown itself to be valuable in coping with the existing environment in a satisfying manner. Also, in affectively proximate situations, moral context (choice) is likely to be tied to moral structure (reasons). Because this pairing between the choice and structure is so strong and routinized, when new structures are introduced, they may be rejected as offering only uncertainty. (h) What is affectively salient for one person is often affectively neutral for another. The role of moral perception - w h e t h e r or not a real-life situation is perceived to have affective salience - i s a key factor in whether an individual will be open to cognitive growth in that context or whether he will stay with more established patterns of thought. The above picture of moralization is consistent with the view that individuals' moral cognition undergoes structural and qualitative transformations. However, the interactions between the moral content of the social environment, the individual's cognitive structures, and the affective salience of moral situations suggest that the explanation of moral thought deserves to be painted in broader brushstrokes. Broader and more comprehensive conceptualizations of the moralization process hold the key to the value of future inquiry and research. Narrow and unidimensional views of the learning of moral standards can never fully explain this most complex and diverse aspect of individual growth and development. REFERENCES
Aronfreed, J. (1968). Conduct and Conscience, Academic Press, New York. Aronfreed, J. (1976). Moral development from the standpoint of a general psychological theory. In Lickona, T. (ed.), Moral Development and Behavior, Holt, Rinehart and Winston, 1976. GiUigan, C. F., Kohlberg, L., Lerner, J., and Belenky, M. (1971). Moral reasoning about sexual dilemmas: The development of an interview and scoring system. In Technical Report of the Commission on Obscenity and Pornography. Vol. 1 (No. 52560010), Government Printing Office, Washington, D.C.
416
Leming
Haan, N. (1975). Hypothetical and actual moral reasoning in a situation of civil disobedience. J. Personal. Soc. Psychol. 32: 255-270. Haan, N., Smith, M. B., and Block, J. (1968). Moral reasoning of young adults: Political social behavior, family background and personality correlates. J. Personal Soc. Psychol. 10: 181-201. Kirk, R. E. (1968). Experimental Design: Procedures for the Behavioral Sciences, Brooks/ Cole, Belmont, Calif. Kohlberg, L. (1971). From is to ought: How to commit the naturalistic fallacy and get away with it in the study of moral development. In Michel, T. (ed.), Cognitive Development and Epistemology, Academic Press, New York. Kohlberg, L. (1972). lssue Scoring Guide, Laboratory of Human Development, Cambridge, Mass. Kurtines, W., and Grief, E. (1974). The development of moral thought: Review and evaluation of Kohlberg's approach. Psychol. Bull 81: 453-470. Rest, J. (1973). The hierarchial nature of moral judgment: A study of patterns of comprehension and preference of moral stages. J. Personal. 41 : 86-109. Rest, J., Turiel, and Kohlberg, L. (1969). Relations between level of moral judgment and preference and comprehension of the moral judgment of others. J. Personal. 37: 225-252.
Intrapersonal Variations in Stage of Moral Reasoning Among Adolescents as a Function of Situational Context James S. Leming 1
Received March 27, 1978
On a set o f classical moral dilemmas and on a set o f practical moral dilemmas 60 randomly selected eighth- and twelfth-grade students were asked to (a) judge the actions o f others (]udgrnent) and (b ) reach a personal decision and give reasons for what they would do (deliberation). The interviews were transcribed and scored for stage o f moral reasoning. Stage o f moral reasoning associated with deliberation on practical moral dilemmas was found to be significantly lower than their moral judgment on classical moral dilemmas. A n explanation was presented which combined social learning and cognitive-developmental moralization theory.
~TRODUCTION The recent review by Kurtines and Grief (1974) of the empirical basis of the cognitive developmental approach to the study of moral judgment has raised many serious questions concerning the reliability and validity of Kohlberg's moral reasoning typology. In their review Kurtines and Grief analyzed the construct validity of the Kohlberg instrument from two perspectives: the claim for the invariance of sequence of stages and the claim for moral development as representing a hierarchy of complexity. Also involved in the analysis of the construct validity of the Kohlbergian system, and not discussed in the Kurtines and
1Assistant Professor, Department of Curriculum, Instruction and Media, Southern Illinois University at Carbondale. Receivedhis Ph.D. from the University of Wisconsin at Madison. Research interests include the nature of moral character and the implications involved for the moral education curriculum. 405 0047-2891/78/1200-0405 $05.00/0 9 1978 Plenum Publishing Corporation
406
Leming
Grief (1974) review, is the assumption that moral reasoning represents a generalized cognitive orientation which is applied consistently across all situations. Kohlberg assumes that moral reasoning is not situationally specific behavior. Kohlberg (1971) discusses this dimension of his theory in the following manner: Stages def'me "structured wholes," total ways of thinking, not attitudes toward particular situations. (p. 169)... Moral stages constitute "principles" in the sense that they represent the major consistencies of moral evaluation within the individual not directly due to factual beliefs (Our evidence for this is that factor analysis indicates a single stage factor cutting across all moral situations, and all aspects of morality on which the individual is assessed) (p. 177). In other words a given stage response does not represent a specific response determined by knowledge and familiarity with that task or similar ones; rather, it represents an underlying thought organization. One would expect, therefore, that at a given point in developmental time, individuals will consistently use their highest attained level of moral reasoning to resolve moral conflict regardless of any variations in situational context. This expectation, entailed by the cognitivedevelopment theory, will form the basis of the research reported here. In order to inquire into this aspect of the Kohlberg theory it is necessary to first refer to some salient distinctions within the realm of moral reasoning. The first distinction is between the judgment mode and the deliberation mode of moral reasoning. Moral judgment will be defined as occurring where one tries to determine after someone else has acted whether that person has taken the best course open to him, for the purpose of praising or condemning him. Moral deliberation will be defined as moral reasoning where one tries, before acting, to determine which is the best course open to oneself with a view to entering on it. In other words, in the judgment mode of moral reasoning the individual is presented with a fait accompli and is asked to evaluate the actions involved. I n the deliberation mode of moral reasoning the individual is faced with an incomplete situation and his task is to choose the morally correct course of action for himself. A second crucial distinction centers on the differences between classical moral dilemmas and practical moral dilemmas. Classical moral dilemmas refer to situations which are removed from the life space of the individual and involve characters with whom the subject has trouble identifying. Practical moral dilemmas refer to situations within the life space of the individual and involving people and issues familiar to him. An analysis of Kohlberg's method of assessing stage of moral reasoning reveals that he is primarily measuring the individual's moral judgment applied to classical moral dilemmas. For example, in a typical dilemma a man (Heinz)has stolen a drug which will save his wife's life. The subject is then asked such questions as, Should Heinz have stolen the drug? Should Heinz be punished? Clearly, the moral reasoning measured under these conditions does not deal with issues or situations familiar to many people. On the basis of the conception that stage
Moral Reasoning and Situational Context
407
of moral reasoning represents "structured wholes," the cognitive developmentalists assume that moral judgment about classical dilemmas is similar to deliberation about practical dilemmas. This research will examine the empirical basis of this assumption. Specifically, this research attempts to answer the following two questions: (1) Is there a difference between the moral reasoning people use to justify actions they claim they would take and the moral reasoning they use in judging the actions of others? (2) Is there a difference between the moral reasoning people use in dealing with situations unfamiliar to their life space and the moral reasoning they use in situations familiar to their life space? METHOD
Instrument
The Kohlberg method of assessing stage of moral reasoning involves interviewing the subjects on moral dilemmas using a semistructured interview schedule, tape-recording the subject's responses, transcribing the tape recording, and finally scoring the transcript according to procedures developed by Kohlberg (1972). As a result of the scoring, the subjects are assigned to one of the six stages of moral development. In order to assess differences in moral reasoning caused by mode of reasoning and type of situation it was necessary to develop alternative moral dilemmas and interview schedules. The classical dilemmas used were Heinz, Joe and his father, and Alexander. A series of practical moral dilemmas was developed which contained moral conflict situations likely to be found within the life space of the prospective subjects. A questionnaire was given to 186 seventh- and twelfthgrade students asking them to identify or suggest moral conflict situations with which they were familiar. From the situations most frequently identified, a set of six pilot dilemmas was created. A final selection of three practical moral dilemmas was made on the basis of the results of a pilot study. A typical practical moral dilemma was the Party dilemma. In this dilemma a girl's parents have denied her permission to go to a friend's party. The girl's best friends are going to be at the party, so she tells her parents that she is going to a movie and goes to the party anyway. The other two practical moral dilemmas deal with the issues of cheating (the Assignment dilemma) and peer group conflict (the Group dilemma). An interview schedule was developed for the practical dilemmas to assess the subjects' stage of moral reasoning in the judgment mode. After each dilemma the subjects were asked to respond on a 5-point scale to the statement, "In my life situations like this one are familiar." Significantly more subjects (p < 0.05) were found to agree with this statement after discussing practical moral dilemmas than agreed with it after discussing classical moral dilemmas.
408
Leming
Care was taken to ensure that the practical moral dilemmas and interview protocols contained many of the same issues as the classical moral dilemmas and their interview protocols. The issues of affectional role obligations and contact and promise were present in both the classical and the practical moral dilemmas. Thus, any differences found in mean level of moral reasoning should reflect inconsistencies on the same moral issues across situations and should not be attributed solely to horizontal decalage, that is, individual differences resulting from one set of dilemmas including one set o f moral issues and the other set o f dilemmas containing a different set o f issues. In addition to measuring the subjects' moral reasoning in the judgment mode on the classical and practical dilemmas, it was also necessary to measure their moral reasoning in the deliberation mode on the same sets of dilemmas. To accomplish this it was necessary to reword both the classical and practical dilemmas, so that they were now stated in the present tense and the actual choice in the dilemmas was still open and unstated. For example, the Heinz dilemma had to be rewritten in such a way that the subject is asked to consider a situation where someone he loves is dying of cancer, he cannot raise the money, and the choice offered is whether or not he will steal the drug to save his loved one's life. The interviewer carefully avoided incorporating " y o u " questions into the judgment interview protocols. In sum, there were four sets o f dilemmas and interview schedules on which the subject's stage o f moral reasoning was assessed (see Table I). Subjects The sample consisted of 60 public school students randomly selected from two middle schools and two high schools in the Madison, Wisconsin, area. At the time o f the interviews 30 of the subjects had just completed seventh grade and 30 of the subjects had just completed eleventh grade. Equal numbers of boys and girls were present in the sample. Procedure
The interviews took place at two o f the local schools. The subjects were interviewed on 12 different dilemmas, 3 within each o f the four forms of moral Table I. Forms of Moral Reasoning
Components of form Form
Mode of moralreasoning
Type of dilemma
MJCMD MDCMD MJPMD MDPMD
Moral judgment (MJ) Moraldeliberation (MD) Moral judgment (MJ) Moraldeliberation (MD)
Classicalmoral dilemma (CMD) Classicalmoral dilemma (CMD) Practical moral dilemma (PMD) Practical moraldilemma (PMD)
Moral Reasoning and Situational Context
409
reasoning. The interviews took between two and three hours for each subject. Two 10-minute breaks were given, and the order in which the dilemmas was presented to each subject was randomly determined in order to eliminate any fatigue effect. Each interview was tape-recorded and then transcribed. Scoring The transcripts of the interviews were scored according to procedures outlined by Kohlberg (1972). Scoring was organized by form, which resulted in four stage scores for each subject. Stage of moral reasoning scores can be reported as either global scores or as mean moral maturity scores. The global score and the subject is classified as either a pure stage or a mixed stage. For ease of statistical computation, this study used the subject's mean moral maturity scores. The mean moral maturity score (MMS) is ascertained by identifying stage scorable responses by issues within the transcripts of individual dilemmas, Issues are defined by Kohlberg (1972) as "defining the concrete objects of concern or value to the subject in the situation. Secondly they are the things to be defined and chosen between in the situation, they define the moral conflict" (p. 18). For every issue stage score within a form a point value is then assigned. The point value is based on a ratio of 3:2:1, depending upon whether the stage score was circled (most salient issue for resolving the dilemma), uncircled (ascertained with a high degree of certainty), or question-marked (ascertained with a low degree of certainty). Next a percent score was figured for each stage present in the subject's reasoning based on the total points assigned. The percent score was then multiplied by the number representing the stage. When summed the results can yield scores ranging from 100 (100% at Stage 1) to 600 (100% at stage 6). In order to obtain a reliability score it was necessary to hire and train a graduate student in education. This reliability scorer evaluated the response of 10 randomly selected transcripts. A product-moment correlation coefficient was computed between the two scorers' MMS on the individual forms. Using this procedure the correlation coefficients for the 10 subjects on the separate forms 2 was MJCMD (0.88), MDCMD (0.79), MJPMD (0~ and MDPMD (0.80).
Statistical Analysis
A 2 • 2 • 2 X 2 analysis of variance permits examination of the data for significant differences by sex, year in school, mode of reasoning, and type of 2See Table I for explanation of the forms.
410
Leming
dilemma. Standard fixed effects analysis of variance statistical design requires independent assignment of scores to each cell. This research does not meet this condition, since individual subjects' scores were assigned repeatedly across the four factors identified. That is, there is a score for each subject in all the four forms of moral reasoning. Therefore, Kirk's multiple factors (1968) repeated measures split-plot design ( S P F - pr.gu) was used. This method of analysis allows one to answer the primary research questions concerning differences in moral reasoning between modes of reasoning and types of situations.
RESULTS
The analysis of variance found that two of the four main effects were statistically significant beyond the 0.05 level. Twelfth-grade subjects were found to be significantly higher in mean moral maturity scores than eighth-grade subjects, and mean moral maturity scores for all subjects were found to be significantly higher in the judgment mode than in the deliberation mode. Two of the interactions in the analysis of variance were statistically significant. First, the Table 1I. MMS (All Subjects) for Mode • Dilemma Interaction D i l e m ~ / Mode
Across CMD PMD dilemma
MJ MD Across mode
264.4 263.6 260.3 240.9 262.3 252.2
263.9 250.6 257.3
28O 270 260 MMS 250 240 230 CMD
PMD
Type of Dilemma
Fig. 1. Graph of interaction: Mode X Dilemma. Open triangle =M J; circle = MD.
Moral Reasoning and Situational Context
411
Table IIL MMS (All Subjects) for Age X Mode • Dilemma Mode (dilemma) Grade 8 12 Across grade
MJ
MD
CMD PMD
CMD PMD
Across mode (dilemma)
246.8 249.0 281.9 278.1 264.4 263.6
245.6 235.0 275.1 246.9 260.3 240.9
244.1 270.5 257.3
interaction between mode of reasoning and type of dilemma was found to be significant beyond the 0.001 level. An examination of Table II reveals that the interaction is due to substantially lower MMS for deliberation on practical moral dilemmas. Figure 1 presents this interaction graphically and demonstrates that the type of dilemma is only an important factor in one's moral reasoning within the deliberation mode. The interaction between mode o f moral reasoning and type o f dilemma can be further understood through an examination of the second significant interaction: Grade X Mode X Dilemma. The means associated with this interaction are reported in Table III. These means indicate that the most marked change in MMS occurs for both eighth- and twelfth-grade subjects engaged in the MDPMD form of moral reasoning. Across grade this drop in MMS is around 20-24 points. These means and hence the interaction can be more fully understood graphically. Figure 2 presents the graph of the Grade • Mode X Dilemma interaction. This graph shows that the Mode X Dilemma interaction is due to the twelfth-grade subjects' drop in MMS in moral deliberation within practical
290 280 270 MMS 260 250 240 230 MJCMD
MJPMD
MDCMD
MDPMD
Mode & Dilemma
Fig. 2. Graph of interaction: Grade • Mode • Dilemma. Open triangle = 8th grade; circle = 12th grade.
412
Leming
Table IV. Percent of Stage Usage on the Four Forms of Moral Reasoning by Twelfth-Grade Subjects Stage Form
1
2
3
4
5
MJCMD MJPMD MDCMD MDPMD
0 0 1 0
45 49 43 68
34 35 43 22
15 10 10 6
6 6 5 4
moral dilemmas, or MDPMD. The significant difference found between eighthand twelfth-grade subjects can now be interpreted as due predominantly to age differences on the MJCMD, MDCMD, and MDPMD forms of moral reasoning alone. A post hoc analysis utilizing the Sheff6 procedure showed that the difference between eighth- and twelfth-grade subjects on the MDPMD form of moral reasoning was not statistically significant (p < 0.05). The statistically significant difference between MMS on modes of moral reasoning is accounted for by the low mean scores found on moral deliberation within practical moral dilemmas. A further clarification of this significant drop in level of moral reasoning can be gained by examining subjects' percentage stage usage across all four forms of moral reasoning. It is evident from the data presented in Table IV that the drop in MMS on the MDPMP form of moral reasoning for twelfth-grade subjects is due largely to a decrease in Stage 3 usage and a concomitant increase of Stage 2 responses. In addition, an intercorrelation matrix between the-subjects moral reasoning on all four forms shows that the standard conception of moral reasoning as defined by Kohlberg (MJCMD) explains only 31% of the variance on deliberation on practical dilemmas (MDPMD). In summary, the analysis of variance reveals that (1) older subjects are significantly higher in mean stage of moral reasoning than younger subjects; (2) across all subjects, reasoning in the judgment mode is significantly higher than reasoning in the deliberation mode; (3) the interactions of Mode • Dilemma and Grade X Mode X Dilemma are statistically signifiant; and (4) the interaction effects are the result of the twelfth-grade subjects' lower MMS on the MDPMD form of moral reasoning.
DISCUSSION In order to fully understand the significance of these findings it is necessary to clarify more precisely the nature of the moral reasoning assessed in this study. Previous studies on the comprehension of moral statements by Rest (1973) and
Moral Reasoning and Situational Context
413
Rest et al. (1969) have suggested that individuals understand and therefore can communicate in the language of all stages below their own modal stage of moral reasoning as assessed by standard Kohlbergian procedures. When assessing an individual's stage of moral reasoning with the Kohlberg methodology, one is measuring the highest stage attained by the individual. Even though a subject may verbalize lower stages during the course of an interview, the key statements for assessing stage of moral reasoning are those in which he discusses what he sees as the best reasons for a given decision. In the process of being interviewed, many subjects supersede initial lower stage scorable responses with higher stage scorable responses. This happens as a result of careful probing by the interviewer which highlights the difficulties and complexities inherent in the intial responses and causes the subject to generate alternatives. Often the interviewer must ask the subject to choose between a number of alternatives discussed in the course of the interview and to select the best reason for a given moral decision. The reasons identified by the subject as best are the responses which are eventually stage scored because they are the statements which the subject sees as ultimately defining right or wrong. Although subjects may see the usefulness of stages lower than their own modal stage in resolving moral dilemmas, under skillful probing eventually these lower stage reasons are rejected for what are seen as preferable or better reasons. In this study, however, the twelfth-grade subjects spontaneously used and defended as best a lower mean level of moral reasoning in deliberation on practical moral dilemmas than they used in judging the right thing to do on classical moral dilemmas. A detailed analysis of the data revealed that this finding was not due to differences associated with intrapersonal variation from moral issue to moral issue. Few subjects ever reason consistently across all issues at the same stage and can be scored 100% at a given stage. That subjects in this study tended to use lower stages in the resolution of more familiar moral conflict situations is not an isolated finding. Gilligan et al. (1971) compared moral judgment of adolescents on sexual dilemmas with moral judgment on the standard Kohlberg dilemmas. They found that "Considering the sample as a whole, the level of moral reasoning was higher on the standard dilemmas than on the sexual dilemmas. The mean MMS for the standard stories was 345 compared with 315 on the sex stories" (p. 145). Gilligan etal. (1971) found that 50% used the same major stage on both sets of dilemmas, 40% used lower stages and 10% higher stages on the sex dilemmas. Haan's reexamination (1975) of the Haan e t al. (1968) data on college activism also found significant intrapersonal variation as a function of situational context. Haan (1975) found that two thirds of the college sample used a different modal stage of moral reasoning - 46% higher and 20% lower - when reasoning about the morality of actual campus unrest than they did when reasoning about Kohlberg's hypothetical (classical) moral dilemmas. Since all of the subjects in the Haan et al. (1968) study had recently been exposed to or had participated in
414
Leming
campus unrest, their higher scores on reasoning about the actual situation can be considered as a form of growth, since the experience was potentially disequiliberating. However, the fact that 20% in the Haan (1975) study, 31% of the subjects of the present study, and 40% of the subjects in the Gilligan et. al. (1971) study reasoned at a lower modal stage about a realistic situation than about a classical (hypothetical) situation calls for a more careful examination of the "structured wholes" concept. Specifically, an attempt must be made to explore why reasoning about familiar situations affects some individuals' ability to use their highest attained level of moral development in the resolution of moral conflict. This phenomenon of situationally specific intrapersonal variation in stage of moral reasoning suggests an interpretation of moralization that involves social learning dynamics in addition to cognitive development. Aronfreed (1968, 1976) has argued that the social environment plays a much more specific and determinant role in the moralization process than the cognitive developmentalist would have us believe. The discussion below builds upon Aronfreed's insights. Kohlberg has assumed that the child carries an intrinsically ordered incipient program of moral development which unfolds under the stimulation of social experience. Kohlberg has not viewed this social experience as the transmitter of the stages, for the stages are an underlying structure and not a set of social values. Rather he has considered the social experience as a generalized and indeterminant nutriment. The research presented above suggests that this view of moralization needs to be broadened to include the possibility of situationally specific stage responses to moral conflict which are resistant to broader shifts in cognitive maturity. The following explanation attempts to retain a developmental view of cognitive growth while accounting for the operation of social learning dynamics. The explanation summarized below is an attempt to account for why a percentage of a population (ranging from 20%-40%, depending upon the study cited) when reasoning about more practical and immediate situations use reasoning at a lower level of moral development- than they demonstrate when reasoning about more hypothetical moral situations. The following statements present this explanation: (a) Structural changes in moral thought are predicated upon more general shifts in cognitive development. That is, structural changes are not intrinsically moral; their moral status resides in the substantive base of the values upon which they operate. This substantive base is learned by the individual from the social environment. (b) A structural change in an individual's general cognitive reasoning does not always entail a parallel structural change in moral reasoning. General structural cognitive change is a necessary but not sufficient condition for structural change in moral reasoning. (c) New acquisitions in moral thought displace rather than transform the moral structures previously held.
Moral Reasoning and Situational Context
415
(d) Displacement of previous forms of moral thought by new forms of moral thought is a function of the context of the moral reasoning and the set of expectancies of reinforcement associated with the old and new forms of thought within that context. (e) Individuals operate in many different moral environments. Each of these environments contains different moral structures (reasons) and contents (choices). (f) The more hypothetical (removed from the life space of the individual) the context of the moral reasoning, the less likely the individual is to anticipate reinforcement (either negative or positive) as a result of adopting new forms of reasoning. As a result, anticipated negative reinforcements are less likely to serve as obstacles to the transition of new structures. (g) The more affectively proximate (within the individual's life space) the context of the moral reasoning, the more likely the moral structures are to exist within a set of expectancies of reinforcement. The individual adopts a particular pattern for resolving moral conflict because it has shown itself to be valuable in coping with the existing environment in a satisfying manner. Also, in affectively proximate situations, moral context (choice) is likely to be tied to moral structure (reasons). Because this pairing between the choice and structure is so strong and routinized, when new structures are introduced, they may be rejected as offering only uncertainty. (h) What is affectively salient for one person is often affectively neutral for another. The role of moral perception - w h e t h e r or not a real-life situation is perceived to have affective salience - i s a key factor in whether an individual will be open to cognitive growth in that context or whether he will stay with more established patterns of thought. The above picture of moralization is consistent with the view that individuals' moral cognition undergoes structural and qualitative transformations. However, the interactions between the moral content of the social environment, the individual's cognitive structures, and the affective salience of moral situations suggest that the explanation of moral thought deserves to be painted in broader brushstrokes. Broader and more comprehensive conceptualizations of the moralization process hold the key to the value of future inquiry and research. Narrow and unidimensional views of the learning of moral standards can never fully explain this most complex and diverse aspect of individual growth and development. REFERENCES
Aronfreed, J. (1968). Conduct and Conscience, Academic Press, New York. Aronfreed, J. (1976). Moral development from the standpoint of a general psychological theory. In Lickona, T. (ed.), Moral Development and Behavior, Holt, Rinehart and Winston, 1976. GiUigan, C. F., Kohlberg, L., Lerner, J., and Belenky, M. (1971). Moral reasoning about sexual dilemmas: The development of an interview and scoring system. In Technical Report of the Commission on Obscenity and Pornography. Vol. 1 (No. 52560010), Government Printing Office, Washington, D.C.
416
Leming
Haan, N. (1975). Hypothetical and actual moral reasoning in a situation of civil disobedience. J. Personal. Soc. Psychol. 32: 255-270. Haan, N., Smith, M. B., and Block, J. (1968). Moral reasoning of young adults: Political social behavior, family background and personality correlates. J. Personal Soc. Psychol. 10: 181-201. Kirk, R. E. (1968). Experimental Design: Procedures for the Behavioral Sciences, Brooks/ Cole, Belmont, Calif. Kohlberg, L. (1971). From is to ought: How to commit the naturalistic fallacy and get away with it in the study of moral development. In Michel, T. (ed.), Cognitive Development and Epistemology, Academic Press, New York. Kohlberg, L. (1972). lssue Scoring Guide, Laboratory of Human Development, Cambridge, Mass. Kurtines, W., and Grief, E. (1974). The development of moral thought: Review and evaluation of Kohlberg's approach. Psychol. Bull 81: 453-470. Rest, J. (1973). The hierarchial nature of moral judgment: A study of patterns of comprehension and preference of moral stages. J. Personal. 41 : 86-109. Rest, J., Turiel, and Kohlberg, L. (1969). Relations between level of moral judgment and preference and comprehension of the moral judgment of others. J. Personal. 37: 225-252.
