Moral Schemas and Tacit Judgement or How the Defining Issues Test is Supported by Cognitive Science

Ideas from cognitive science are increasingly in uential and provide insight into the nature of moral judgement. Three core ideas are discussed: modern schema theory, the frequency of automatic decision-making and implicit processes as the default mode of human information processing. The DeŽ ning Issues Test (DIT) measures the beginnings of moral understanding, which are largely non-verbal and intuitive, in contrast to the Moral Judgement Interview (MJI), which measures the highest level of verbal understanding. The positive attributes of the DIT and its conceptualisation of moral judgement schemas are more apparent in a time of increasing respect for implicit knowledge and processing. The DIT offers a means of measuring moral judgement that Ž ts with current views in cognitive science. Although the MJI and interview techniques generally are worthwhile for measuring production competence, the DIT is better able to measure understanding at the level that drives most decisions for most people. Historically, philosophy has described moral judgement as conscious and deliberative decision-making. Consequently, studies of moral judgement usually focus on testing conscious, thoughtful reasoning about moral dilemmas (e.g. the Moral Judgement Interview [MJI], Colby, Kohlberg et al., 1987; Moral Competence, Lind, 1995; Sociomoral Re ection Measure [SRM], Gibbs & Widaman, 1982), distributive justice (Damon, 1975; Enright et al., 1981) and particular non-development distinctions in moral judgement (e.g. “domains,” Turiel, 1983; “culture,” Shweder, 1991; and “orientation,” Lyons, 1982). Common to all these methods is the need for the participant to give a verbal rationale for a decision (orally, as in the MJI, or in writing, as in the SRM). It is assumed that participants make their moral judgements re ectively, that they are able to articulate them, and that the method can be “error-free” (Kohlberg, 1976). Participant verbalisations are puriŽ ed of speciŽ c content and scored for evidence of underlying cognitive structures. In this context, the DeŽ ning Issues Test (DIT) has always stood out as a measure “apart”. Spawned from research in moral comprehension (Rest, 1973), the ISSN 0305-7240 print; ISSN 1465-3877 online/02/030297-18 Ó 2002 Journal of Moral Education Ltd DOI: 10.1080/0305724022000008124 298 D. Narvaez & T. Bock DIT originated as a cognitive-developmental tool that measures stage shifts in the upper half of Kohlberg’s stage hierarchy. From the outset there was concern that the DIT did not really measure moral judgement structures because it did not ask respondents to produce an answer in their own words and because it hopelessly wedded content with structure. Nevertheless, research with the DIT has been fruitful and has provided solid and broad evidence for its ability to measure moral judgement development (Rest et al., 1999) [1]. In fact, it provides more empirical support for the higher end of Kohlberg’s stage model than Kohlberg’s own measures. That is, unlike other measures of moral judgement, the DIT Ž nds widespread postconventional reasoning. In contrast, the MJI Ž nds little evidence for Stage 5 thinking and virtually none for Stage 6, threatening the validity of Kohlberg’s enterprise. What accounts for the evidence that the DIT is more supportive of the stage-theory than Kohlberg’s own measure? Our explanation is the following. The DIT does not measure the more competent end of the “zone of proximal development” in which verbal articulation of one’s perspective is required. To obtain a high score on a measure requiring verbal production such as the MJI one must be able to explain one’s reasoning logically and coherently, an ability that is facilitated by training in moral philosophy, but not necessarily by everyday life. The DIT tests the other, less competent, end of the “zone”, that which is apparent when assistance (such as words on a page) is available. Whereas the MJI measures explicit verbal knowledge, the DIT measures recognition knowledge, a type of tacit knowledge. Is one method better than the other? We believe that new constructs from cognitive science can direct our examination of this question and can help point out the merits of each approach. The study of human cognition in the 20th and 21st centuries offers many tools for measuring and understanding differences in morality and moral judgement. Three interwoven, core ideas emerging in cognitive science have particular implications for research in moral judgement. First, cognitive science makes the assumption that an individual processes or interprets experience according to organising, conceptual structures in the mind that have developed from and are in uenced by experience (e.g. Piaget, 1932/1965; Rumelhart, 1980; Fiske & Taylor, 1991). These conceptual structures, which are often called schemas, function as interpreters of stimuli. Although it may feel phenomenologically as if one experiences pure, raw “reality”, everything perceived is interpreted by pre-existing mental structures (Wenger & Wheatley, 1999; Hogarth, 2001). In fact, no one can understand experience without such mental structures or schemas (Neisser, 1976). Not only perception, but decision-making and reasoning are supervised by pre-existing schemas (Gigerenzer et al., 1999). Further, the vast majority of human responses, including moral judgements, are based on schemas (Bargh & Chartrand, 1999). We discuss schema theory in more detail later. Secondly, many cognitive processes, including decisions, occur automatically without awareness (Reber, 1993). For example, schemas are activated without intention or conscious control (Hasher & Zacks, 1984). Some schemas may be activated chronically due to several factors, including frequency of environmental cues (Higgins & King, 1981). For example, in the United States, local television Schemas and the DIT 299 news stations highlight the types of local crimes that are more likely to be committed by individuals with lower socio-economic status, showing photographs of the accused. Many of these individuals are black. Consequently, when a typical citizen is asked to think about a criminal, a black person “pops” into his or her mind, even though the statistical evidence shows that more crimes are committed by “whites”, and that white males tend to be more violent. Like it or not, the schema fostered by local news is activated beyond personal control. Many schemas operate automatically like this. Automatic processing is understood to be characterised by some combination of the following: involuntariness, autonomy, existence outside of awareness, unintentionality (initiated with or without intention), and effortlessness—requiring little if any working memory resources (Bargh, 1989; Uleman & Bargh, 1989). Many daily actions and responses have a majority of these characteristics. Thirdly, there is now a greater regard for the implicit processes and application of tacit knowledge in human decision-making that occur outside the awareness of the cogniser (e.g. Bargh, 1989) and beyond the participant’s ability to verbally articulate (e.g. Kihlstrom et al., 1996). Individuals have, use and are in uenced by a great deal of knowledge without awareness (Reber, 1993; Hogarth, 2001). Various psychologists have pointed this out throughout the history of psychology, even in its early days (Helmholtz, 1867; Carpenter, 1874; Ebbinghaus, 1885) as have philosophers (e.g. Polanyi, 1958; von Hayek, 1962). Polanyi often noted that people “know more than they can say”. A person may not be able to explain what she knows to another, but neither can the person explain it to herself. Polanyi’s point now has substantial psychological support (see Reber, 1993; Sternberg, 1999; Keil & Wilson, 2000). Evolutionary psychology has postulated that the human mind has evolved with numerous domain-speciŽ c modules (task-speciŽ c cognitive abilities) whose activation is automatic and does not require the conscious mind (e.g. Tooby & Cosmides, 1990). In fact, these broad-based implicit processes are considered to be primary—that is, they are the default operations of the mind—whereas consciousness and phenomenological awareness are relatively recent developments (Reber, 1993). Evidence is mounting for the heavy reliance on automatic, implicit cognitive processes and schemas in human behaviour (Reber, 1993). Research on implicit decision-making calls into question the privileged place of interview data (dependent on conscious understanding) over recognition data (dependent on implicit understanding, as in the DIT). In fact, the current emphasis on the importance of tacit knowledge and consciously inaccessible decision making places the DIT in a new light. Rest (1979) has always claimed that the DIT measures tacit knowledge. Sternberg and Horvath (1999) deŽ ne tacit knowledge as domain-speciŽ c “procedural knowledge that guides behaviour but is not readily available for introspection” (p. 231). The DIT all along has been tapping into the kind of knowledge that drives most of human behaviour (Bargh & Chartrand, 1999; Hogarth, 2001). In the following discussion we focus on the nature of this knowledge and its relation to research in moral judgement. 300 D. Narvaez & T. Bock The Nature of Schematic Processing The notion of schemas is one that has driven research in cognitive psychology for decades. Piaget described schemas as cognitive structures that organize an individual’s operational activities (Piaget, 1970). Classic schema theorists (e.g. Rummelhart, 1980; Taylor & Crocker, 1981) describe schemas as general knowledge structures residing in long-term memory. In order to determine how the human mind understands anything, including morally relevant stimuli, we must know the form and function of schemas generally. 1. What are Schemas? Schemas are sets of expectations, hypotheses and concepts that are formed as the individual notices similarities and recurrence in experience (Neisser, 1976; Rumelhart, 1980). For example, the toddler learns that certain behaviours agitate the parent, some of which result in discomfort (physical and/or emotional) for the child. Other behaviours elicit praise from the parent. With repeated experience, the child begins to expect the particular matching reaction of the parent when the behaviour is exhibited. If the parent, for example, shows the opposite reaction, the child will be very confused as this does not Ž t with the schema that has developed from previous experience. Modern schema theorists have provided more concrete descriptions of schemas. Derry’s (1996) Cognitive Schema Theory (CST) outlines a hierarchy of schemas: (a) memory objects (speciŽ c small units of related characteristics), (b) cognitive Ž elds (an activated set of memory objects) and (c) mental models (an overall meaning structure of a particular situation or experience). According to Cognitive Schema Theory, we might say that those with more complex moral judgement have a larger and better organised set of memory objects that can be activated within multiple cognitive Ž elds and form part of complex mental models. Although research would need to determine the architecture in actuality, we can speculate about a mental model for the dilemma, “Heinz and the drug”. Such a mental model might include cognitive Ž elds for marriage, stealing, human rights and so forth. Each of these cognitive Ž elds would be comprised of memory objects. For example, “marriage” might include memory objects like “marriage duties” and “love”. The memory object, “marriage duties” might consist of related characteristics such as “faithfulness” and “commitment”. There may be many layers of memory objects and cognitive Ž elds that comprise a mental model. An expert will have layer upon layer of interrelated schemas about the “Heinz and the drug” dilemma. The expert has more complex and elaborate mental models that can be activated in any number of ways because the architecture is so rich and interrelated. Those with lower levels of moral judgements have a more limited set of possible activations (fewer memory objects, cognitive Ž elds and mental models). In terms of mental architecture, the expert has castles of knowledge while a novice may have a bare foundation. Schemas involve multiple brain systems (e.g. visual, motor, language) and cognitive processes (Kesner, 1986; Hogarth, 2001). Schemas can involve one kind Schemas and the DIT 301 of system, for example, procedural knowledge (e.g. how to introduce one friend to another) or declarative knowledge (what morality means), or a combination of systems. Schema application can involve different types of reasoning (Ericsson & Smith, 1991), such as analogical and/or intuitive reasoning (Hogarth, 2001); different types of processing such as linear and/or parallel processing (McClelland, 1995); different levels of awareness such as subconscious and automatic or conscious and controlled (Uleman & Bargh, 1991); and different types of knowledge (declarative, procedural). Essentially, a schema is a goal-orientated cognitive mechanism that operates in one or more of these systems (Neisser, 1976). Much like expertise in parking a car, expertise in a moral judgement schema probably necessitates both kinds of knowledge (procedural: how do I think about this problem? and declarative: what codes do I apply?), dual forms of reasoning (analogical: what is an objective, logical response? and intuitive: this reminds me of ...), and processing (linear—what do I do next—and parallel). For example, applying a principle to a dilemma is an example of analogical and linear reasoning whereas making a decision “because it feels right” is based on automatic, parallel processing often described as intuition (Hogarth, 1999; Hammond, 2000). 2. How do Schemas Work? Schemas operate constantly in the mind, being evoked (or “activated”) by current stimulus conŽ gurations that resemble the stimuli that created the schema in the Ž rst place (Rock, 1997). To return to our previous example, the local news may discuss a criminal without using a picture, yet the report may evoke the picture of a black man in the perceiver because of previous similar stories which were presented with pictures of black men. The “criminals are black” schema is activated by familiar stimuli without need for conscious control. When the postconventional moral schema is activated, cognitive structures about equality and social justice and the procedures to promote them are also activated. The person who is concerned about being racist will probably have developed skill in deactivating the “criminals are black” schema and will consciously activate schematic knowledge that counters such a conclusion (Devine, 1989; Fiske, 1989). 3. How are Schema Elements Connected? Hierarchical level (i.e. mental model, cognitive Ž eld, memory object) is not the only thing that varies among schemas. Marshall (1995) describes schemas as basic storage devices represented by a tightly organised network structure of memory objects whose relations vary by types (positive or negative) and degrees (strong or weak) of relations. The degree of connectivity among constituents and subconstituents determines the strength and accessibility of a schema. The type of connectivity determines what group of concepts will be activated (positive relation) and which will be suppressed (negative relation). For example, when a person has an activated postconventional schema, she or he is less likely to have an activated 302 D. Narvaez & T. Bock personal interest schema (see below and Rest et al. (1999) for a complete description of the postconventional and personal interest schema). 4. Why are Schemas Important? Schemas are essential to human understanding because they serve so many functions (Neisser, 1976). A schema consists of a representation of some prior stimulus phenomenon that organises or guides the application of prior knowledge to new information (sometimes referred to as “top-down” processing) (Bower & Cirilo, 1985). Schema guidance is expectation-based processing that attends to the unusual (Mandler, 1984). Schema guidance is, in effect, goal-based concurrent processing; as the stimuli are sensed, they are Ž ltered through the activated schema kaleidoscope (Rock, 1997). Information is “chunked”, inter-related and evaluated (Fiske & Taylor, 1991). Missing or new information is quickly perceived and drives attention (Pashler, 1998). For example, consider the following text: “Harry walked in. He looked at the prices. He ordered”. It is difŽ cult to understand what is going on because several schemas compete to interpret an ambiguous situation. What schema should be in operation—visiting a restaurant, shopping at a warehouse appliance store, or patronising a barber shop? The reader needs further information such as “He enjoyed his mea”. This information informs us that he was at some kind of restaurant, but it is still unclear what kind of restaurant. Our schemas guide us in asking the right question: does he pay after his meal (sit-down restaurant) or before (fast-food restaurant)? Another sentence follows: “He cleared his table and left”. Now we can put all the parts together into a “fast-food-restaurant” schema and Ž ll in the blanks: he ordered and paid, sat down and ate, cleared his table and left. Schemas likely operate in important ways during moral behaviour, by inter-relating different stimuli, Ž lling in missing information, guiding attention and directing problem-solving. Moral schemas can be described as general knowledge structures used in social cooperation. Moral schemas are built from experience in social interaction. They are constructed automatically from the brain noticing the elements in the socially relevant environment that covary and the cause–consequence chains that obtain from particular actions. Schemas decrease the amount of processing needed for encountered stimuli and are considered to be part of every encounter with the environment (Mandler, 1984). 5. How are Schemas Activated? Schema structures that parse incoming sensory data are themselves unconscious and are activated automatically when their patterns match the pattern of incoming data (bottom-up activations) (Marcel, 1983; Mandler, 1984). The perceived regularities may or may not activate linguistic centres and, as a result, may or may not be accessible for verbal description (diSessa, 1982; McCloskey & Kohl, 1983). As Keil and Wilson (2000) point out, individuals are often able to understand something without being able to explain it to others. The inability to articulate understanding is not a matter of forgetting—because the correct explanation is recognisable— Schemas and the DIT 303 rather, the individual has not learned to put the understanding into words. This can be seen when a world-class athlete who is asked to coach a youth team fails miserably due to the inability to translate phenomenal execution skills into verbal instructional skills. Why does this happen? Should not an adult be able to explain what he or she knows how to do? Keil and Wilson distinguish between a basic explanatory set of schemas, present even in infants, and more advanced explanatory schemas that include statements of principles and are evident through verbal performance. Keil and Wilson’s theory can help explain the disparity in Ž ndings between the MJI and the DIT. The reason that individuals who display postconventional thinking on the DIT but may not do the same on the MJI may be that the person has not put their understanding into words. Understanding is “a cognitive state that remains largely implicit but that goes beyond merely being able to correlate variables” (Keil & Wilson, p. 97). The only way to move beyond shallow verbal explanations is to learn the intricacies of a theory (e.g. moral theory), as do experts. The schemas of an expert are more easily activated by features of a problem than the schemas of a person with intermediate knowledge (e.g. Patel & Kaufman, 1993). After the solution schema is activated, the expert will evaluate further information, searching for conŽ rmation of the initial hypothesis. In contrast, the person with intermediate knowledge generally requires much more information before generating a solution schema and generates multiple solution schemas (hypotheses) without the ability to evaluate them. The speciŽ c nature of what memory objects, cognitive Ž elds and mental models are activated, how they are activated (e.g. what kind of stimulus) and when (e.g. in what contexts) are areas ripe for research in moral judgement. 6. Can a Particular Schema Change? Schemas are noted for their  exibility and changeability (Neisser, 1976). No instantiation of a schema is identical to another (Hogarth, 2001). With each instantiation, the schema is altered through assimilation of and accommodation to new experience—integrating new information or modifying the strength of relations among memory objects (Derry, 1996). Schemas change in size, relation to other schemas and in the strength of internal relations. Studies of expertise such as Gijselaers and Woltjer (1997a) note that when solving domain problems novices have superŽ cial knowledge of problems (e.g. a label for the problem). Their representations of problems are stable in test–retest studies. After some initial study, beginners acquire bits and pieces of knowledge. Intermediates are able to appropriately structure and identify features of the problems. Only experts are able to identify effective ways to solve the problem. Yet because of their vast knowledge, experts are  exible in how they represent the problem—making slight changes or adding new subcategories when sorting domain problems in test–retest conditions (Gijselaers & Woltjer, 1997b). These patterns support Rumelhart and Norman’s (1988) view that schemas change with the accretion of new knowledge (e.g. the increased knowledge depth of the intermedi304 D. Narvaez & T. Bock ates), and the tuning and reconstruction of prior schemas (e.g. experts’ slightly changed representations of problems). In sum, schemas are powerful tools that aid information processing. They guide perception, attention, decisions, habits and behaviour. They operate in the moral domain via schemas for moral sensitivity, motivation and action. They operate during moral judgement and when a person takes the DIT. The Schemas of the DIT The DIT is designed to capture moral schema changes that are particularly visible throughout adolescence and early adulthood (Rest et al., 1999). Kohlberg contended that a critical social-cognitive advance in adolescence is the “discovery of society”—that is, understanding how people in society are related to each other through institutions, established practices and role-system (“the system”), not merely on a face-to-face basis, where all are kin, friends or well-known acquaintances (Colby et al., 1987). Others have drawn attention to the development of a sociocentric perspective in adolescents (e.g. Adelson, 1971; Youniss & Yates, 1997). Adelson and O’Neil (1966, p. 304) stated: With advancing age there is an increasing grasp of the nature and needs of the community. As the youngster begins to understand the structure and functioning of the social order as a whole, he begins to understand too the speciŽ c social institutions within it and their relations to the whole.... Thus the demands of the social order and its constituent institutions, as well as the needs of the public, become matters to be appraised in formulating political choices. The adolescent becomes aware that society is organised in terms of a system of rules, roles and institutions, raising simultaneously and necessarily questions of the morality of society and questions of moral authority. How does one organise a network of cooperation on a society-wide basis for mutual beneŽ t? How are power, wealth and opportunity to be distributed? What is the legitimate use of force? These questions are issues of “macro-morality” or society-wide cooperation, rather than issues of “micro-morality”—interactions with known others in everyday life and face-to-face. Rest et al. (1999) proposed that the DIT is particularly good at measuring change in the schemas individuals use to answer the “macro” question—how to get along with people who are not friends, kin or personal acquaintances; how to organise society-wide cooperation. Through various statistical analyses, three factors have been identiŽ ed in DIT scores. These have been named: Personal Interest Schema, Maintaining Norms Schema, and Postconventional Schema. (These were formerly categorized as Kohlberg Stage 2 and 3, Stage 4 and Stage 5 and 6, respectively.) Because a respondent needs at least a 12-year-old reading level when taking the DIT, respondents will have moved beyond the simpler forms of moral judgement from earlier childhood (e.g. Kohlberg’s Stage 1). As a result, in DIT data the simpler types of moral judgement appear collapsed together (i.e. Stage 2 and 3) Schemas and the DIT 305 or do not appear at all (i.e. Stage 1) (Rest, 1986). At this point, we would like to interpret the DIT according to schema theory. The three moral schemas that the DIT measures can be viewed as mental models—an integration of cognitive Ž elds and their memory objects—for reasoning about moral dilemmas. (Of course, there are other moral schemas that a person might use other than the three that the DIT measures, such as speciŽ c religious or cultural moral schemas.) We describe the content of each schema according to recent theory (Rest et al., 1999). The simplest schema measured is called the Personal Interests Schema (PIS). The PIS does not suppose a macro-morality perspective but includes the more Stage-2-like memory objects and cognitive Ž elds such as survival, personal advantage and impulsive cooperation. A fair world is one in which I can get what I want. At this level of thinking, the cogniser begins to consider the needs of others, but only in brief exchanges. The cogniser is able to sacriŽ ce momentarily as long as the result is more advantageous. That is, the cogniser learns to exchange cooperation, however brie y. Because the DIT requires a 12-year-old reading level, it is unable to distinguish carefully between Stage-2-like schemas from the more Stage-3-like schemas. And so both are merged in the PIS. We suppose that Stage-3-like schemas include cognitive Ž elds of caring for others, such as long-term negotiated cooperation and in-group reciprocity. PIS thinking enables the cogniser to apply reciprocity in their mental transformations of relationships with known others. Overall, the PIS answers the question of how to organise cooperation in society as if there were only “micro-moral” relationships to consider. The PIS attends to what each stakeholder in a moral dilemma has to gain or lose personally and in relation to the welfare of signiŽ cant others. PIS thinking is not concerned with organising cooperation on a society-wide basis, the issues of macro-morality. At the cusp of adolescent change, sensitivity to the issues of macro-morality begins to  ower. The PIS becomes inadequate for addressing issues that have become interesting and important, such as why we have laws or what duties I have towards other members of society. As a result, a more developmentally advanced type of thinking emerges that includes a wider societal perspective, the Maintaining Norms Schema (MNS). Within this schema one considers how people should cooperate generally with those who are not friends, kin or well-known acquaintances. The cogniser is able to coordinate personal/signiŽ cant other negation with reciprocity for the larger society through the interplay of cognitive Ž elds that describe established practices, rules, and codes and their de facto authorities. Other cognitive Ž elds connected to this schema include those relevant to issues of fair cooperation such as a uniform application of norms across individuals . Expectations of individuals are only partially reciprocal, however, because role and duty expectations are applied in an egalitarian but not equitable way. As development progresses, the Postconventional Schema (PCS) begins to make more sense. The individual has experiences, including the persuasive arguments of others, that necessitate thinking about a fair society more broadly—in terms of full reciprocity and equity across all groups within a society. If the individual has enough pluralistic (e.g. multicultural—see Endicott et al., in press) 306 D. Narvaez & T. Bock experiences to effect enough of a cognitive disequilibrium that is not easily denied or minimised, the individual’s thinking accommodates to and integrates a broader, postconventional perspective into the mental model of a moral society. The cogniser is able to suggest changes to the status quo, to negate extant societal laws and structures, for the sake of morality. He or she can apply moral ideals in a fully reciprocal manner in which each member of society has equivalent status. In particular, postconventional thinking considers macro-level cooperation in terms of advocating sharable ideals that are open to scrutiny and negotiated through the give and take of community life. Whereas Kohlberg focused on a particular postconventional orientation informed by Kant, Rawls and Frankena, the DIT postconventional schema has a broader scope. The PCS can be shaped by various combinations of moral and political philosophy. Individuals with the full use of postconventional tools are able to function at the highest levels of solving moral dilemmas within the community. The features for each schema are described in Table I (from Rest et al., 1999). We have constructed a diagram (see Fig. 1) of mental representations of the Heinz dilemma based on which moral schema is activated (please note that it is simplistic and partial, particularly in its representation of the complexity of the Postconventional Schema, and meant only as an example). It should be noted that network connections of features can be negative or positive. For example, a person who has reached postconventional reasoning will likely have a schema that contains a negative connection to absolute law, representing a perspective contrary to the MNS perspective. Some features of a problem representation will be activated under each schema, for example we have marked “love” and “personal advantage” although the activation may switch in valence. It should be noted that schemas may be embedded in or overlap with one another. For example, a religious fundamentalist perspective and a conservative political orientation often overlap with the Maintaining Norms Schema (see Narvaez et al., 1999). Although the three DIT schemas do not cover everything of importance in moral thinking, nor do they constitute a full model of moral-decision making (see Bebeau & Thoma, 1999), they are relevant to many issues of public policy controversy (e.g. abortion, religion in public schools, rights of homosexuals, women’s roles, euthanasia, due process rights of the accused, free speech and political demonstrations, etc.—see Narvaez et al., 1999). Moral schemas can illuminate how people formulate opinions about these hotly debated public policy issues, about the “culture wars” (Orthodoxy versus Progressivism—Hunter, 1991), and about the most important clash in ideology since the Cold War (religious fundamentalism versus secular modernism—e.g. Marty & Appleby, 1990). How does the DIT measure moral schemas?