THEORETICAL FRAMEWORK FOR EVALUATION OF CROSS-CULTURAL TRAINING EFFECTIVENESS
HARRY C. TR~A~~~~ University of Illinois, Urbana-Champaign The first part of this paper presents a theoretical framework which includes the independent and dependent variables that should be considered in studies designed to evaluate the effectiveness of cross-cultural training. Six kinds of training-general or specific, affective, cognitive or behavioral-plus self-insight training are considered. In addition, the quantity and timing of the training, and attributes of trainers and trainees are considered. The second part considers strategies for the evaluation of the effects of these inde~ndent variables. Controversies in evaluation research are reviewed to find out what can be learned from past experience that might be applicable to cross-cultural training. This review shows a trend toward the broadening of evaluation research, using many more dependent variables, with measurements obtained from many kinds of people. New methodologies that permit the incorporation of diverse points of view and a broad analysis of the effects of the training program are examined. Questions are also raised concerning who should have the information acquired during an evaluation study. Finally, the paper examines ethical problems that cannot be ignored by those evaluating crosscultural training.
In an increasingly shrinking, interdependent world, socialization for “cultural interdependence” will become an increasingly important problem. Future research will focus on such issues such as how to educate members of each culture (so as to maximize their effectiveness in each particular ecology), how to select diplomats, business leaders, union officers, religious leaders, educational representatives, legal advisors, and others whose job it will be to tie together the multinational corporations and international agencies, and how to train such persons for effective infraction with their counterparts. How to train the dependents of persons having to live in another culture, and particularly how to train the children who live in cultures other than their own, will soon become important research topics for developmental psychologists. How to constitute teams of experts from many cultures will soon become a major focus of the work of group psychologists.
20
International
Journal of Intercultural
Relations
A precondition of progress in such research is greater understanding of the effects of cultural training. There are already many kinds of cultural training (Brislin & Pedersen, 1976) but the weakest aspect of this work is the evaluation of its effectiveness. This paper outlines a framework for studies designed to evaluate such training. First, it examines the independent and dependent variables that should be considered in training studies. Second, it explores controversies within the field of evaluation research, which suggest different strategies of evaluation. The hope is that this paper will initiate some convergence toward a paradigm that might be used by many, repeatedly, to develop a cummulative knowledge of the effects of training on interpersonal relationships. THE INDEPENDENT
The Independent
AND DEPENDENT
VARIABLES
Variables
There are several kinds of training that are rather different in their emphasis. These will be described as affective, cognitive, behavioral, general, and specific training. Affective, cognitive, behavioral. Affective training is concerned with the training of the emotions. Emotions are elicited by particular situations, people, and environments. A person, as a result of early socialization, develops a pattern of emotional responses to specific situations, people, or environments. For example, one person may feel good being at parties, being with women who look like one’s mother, or when looking at the view from the top of a mountain; another may feel bad in the very same circumstances, reflecting excessive shyness, a poor relationship with a mother, or acrophobia. When a person is in a different culture he/she is exposed to different situations, people, and environments, some similar to those that produce positive and some similar to those that produce negative emotional responses. Using various kinds of desensitization techniques (e.g., Wolpe, 1958, 1969) one can ensure that the circumstances which frequently occur in the other culture and produce negative emotional responses no longer do so, and using frequent associations of pleasurable events with these circumstances one can increase the frequency of positive emotional responses to living in the other culture.
Triandis
21
Cognitive training concerns the change of thinking about the environment. A person often analyzes events in the environment. For example, he may ask: “Why did 0 send me an invitation to her wedding? To express friendship? To get another wedding present? To show off? To tell me she regrets our broken relationship?” Similarly, in a cross-cultural situation the number of inte~retations of the other’s behavior is large. Interpretation of the other’s behavior involves complex interactions of norms, roles, the self-concept of the actors, interpersonal agreements, past history of interaction, affective and expected future events (Triandis, 1975). To interpret the behavior of the other correctly one needs to know a great deal about the other culture. Triandis (1975) has analyzed this problem in detail, and concluded that a good relationship requires isomorphic attributions. That is, the causes that 0 attributes to his own behavior should be parallel to the causes that P attributes to0 s behavior; and vice versa, P should see his own behavior as caused by the same factors that 0 sees as causing P’s behavior. To achieve this, a form of training called a culture assimilator (Fiedler, Mitchell, & Triandis, 1971; Triandis, 1976) is helpful (Landis et al., 1975; Weldon et. al., 1975). Specifically, on tests that measure isomorphic attributions, assimilator trained subjects obtain higher scores. A person may know why the other acts and how it is appropriate to act, but may be unable to do so. For example, American males may know that holding hands with a Mexican male may be appropriate in a pa~icul~setting, as an indication of friendship, but normally have great difficulty in actually doing it. To do something very different from one’s habits one needs either to do role playing or experience a training program based on a behavior modification technique (Wolpe, 1958, 1969) or operant conditioning, based on Skinn~an principles.
General-specific. Training can be specific to a particular culture or general, exposing the trainee to a variety of cultural habits, norms, roles, values, and circumstances, to provide the trainee with a “sample of experiences” which reflect the variations that exist anywhere on earth. Obviously, the culture-specific training is more efficient if the person is to interact with people from only one other culture; however, many international situations involve interaction with persons from several cultures. Furthermore, diplomats and missionaries frequently are assigned to different cultures. Learning the full range of cultural variation may in fact prove more efficient in the long run. There is no
22
International
Journal of Intercultural
Relations
well-conducted, relevant research on this point, so we must leave this as a question. Note now that the three kinds of training (affective, cognitive, and behavioral) can be crossed with the two kinds (general, specific) to produce six kinds of training. While the idea of covering the world, as a part of general training, seems impossible, it is not so difficult if we train a person for interaction with six highly dissimilar cultures. Murdock ( 1957) has divided the world into six cultural areas. We could select one culture from each area and provide training for these six cultures. This would then be a good approximation of general training. The mechanics of general training, then, would be very similar to the mechanics of specific training. There are examples of cross-cultural training of each of the kinds outlined above, though some kinds of training combine the relatively “pure” types implied by the affective, cognitive, and behavioral trichotomy. For example, Trifonovitch (1973) has provided a form of experimental training that has implications for affective training, but will also produce some behavioral and cognitive changes. David ( 1972) has analyzed how behavior modification can be used in cross-cultural training, but his approach does have implications also for affective training. A seventh kind of training. Kraemer ( 1973) has used a form of training inspired by the work of Stewart (Stewart et al., 1969) which focuses on self-insight. The idea is to learn much about one’s own culture and how culture affects behavior and experience. This kind of training is really a cognitive-specific training, but it focuses on own rather than the other culture. Independent
Variables: Other Training Variations
Quantity. An independent variable that can be mixed with the seven kinds of training just outlined to produce a large number of different training programs is the amount of training. In language training “total immersion” programs, where the trainee spends several months in an environment where everyone speaks another language, have proven effective. There are few examples of such intensive courses for culture learning and there are no studies assessing quantitatively their effects relative to short courses. It would be most useful to have evidence on
Triandis
“marginal gains in cross-cultural spent learning about culture.
23
effectiveness”
relative to the time
Order effects. It is reasonable to hypothesize that some combination of the major kinds of training will produce better effects than any one of the major kinds of training alone. One can argue that self-insight should be the first kind of training to be given, to “open up” the trainee, creating an interest in the topic. Or some sort of experiential training, with the frustrations that are correlated with trying to relate to people from another culture, might be the best opener. Once the trainee is ready to learn, culture assimilator training should provide more information per unit of training time than any other method. However, this procedure does not allow the trainee to practice what was learned. Role playing and behavior modification kinds of approaches would be needed to accomplish behavioral changes. Thus the optimal order of raining may be: self-insight-culture assimilator-role playing, behavior modification, or experiential-culture assimilator-behavior modification. To test these hypotheses it would be necessary to compare the effectiveness of these sequences against random sequences of these kinds of training. Timing. When should training be given? Should it be given as part of the general orientation to the new assignment, in another culture, or at the time the trainee arrives in the other culture? Should we also give reentry training to reduce the shock of returning home? One can mix these questions with different quantities of different kinds of training and develop rather complex designs. Trainers. Who should give the training? Should one employ trainers from the other culture, from the trainee’s own culture, or some mixture of the two? What is the optimal mix? Trainees. Who should be trained? Should we train only the visitor or also his/her counterparts? What kinds of attributes of the trainee are most important to consider in “selection” so that one can obtain a population of trainees that is likely to be maximally successful? A number of variables appear relevant. These include: the trainee’s age and stage in career development; sex; level of cognitive complexity (probably the more complex would fit best in more complex cultures; the less complex in less complex cultures); culture (some cultures are
24
International
Journal of Intercultural
Relations
similar and transfer of training from own culture is easy; others are different and transfer is difficult; it is likely that the Skaggs-Robinson surface which relates positive and negative transfer of training to the similarity and differences of the stimuli [situations] and responses [behaviors] (Osgood, 1953) required in the other cultural setting, would provide a good theory for such studies); the status of the trainee and its relationship to the statuses of those in the other culture (probably status congruence would be helpful); functional preparation of the trainee; previous cultural experience; and finally, for some assignments, the trainee’s religion may be relevant. The Dependent
Variables
The major criteria of success can be classified into economic, and psychological.
social
Economic. The cost of training in relation to the effectiveness of the trainee abroad, taking into account such factors as whether the trainee and his/her dependents had to return too soon, and the effectiveness of the accomplishment of the assignment, are among the most important criteria in this category. Social. The extent to which the trainee developed successful interpersonal relationships, long lasting friendships, strong interdependence with hosts, and similar measures would be revelant. Some sociometric measures obtained from the hosts might be most revealing. Psychological. The extent to which the trainee feels “good” or “bad” about the period spent in the other culture, during and after the visit, would be important criteria. The trainee’s level of anxiety or depression can be measured with both physiological and subjective cndicators. indicators. An Example In a study that we are currently doing on the effects of training Spanish-speaking pupils and their teachers, we are measuring, with multiple indicators, the quality of the “interpersonal relationships.” We used both questionnaires and observation to obtain measures of this theoretical construct.
Triandis
From the teachers we obtained the following
25
measures:
1. Expectations of how Latin0 and Anglo pupils will behave. 2. Behavioral intentions about own behavior toward such pupils. 3. Ratings of the students on evaluative scales (well adjusted, pleasure to have in class). 4. Judgments on an attribution test, which measures the extent to which teacher attributions are insomorphic with the attributions that Latin0 pupils make about their own behavior. 5. Judgments of evaluation of their training (training was good, interesting, practical, a good use of my time). 6. Observations of classroom behavior.
From the pupils we obtained: 1. Expectations of how the teacher will behave toward them. 2. Behavioral intentions to behave toward the teacher. 3. Responses to a Thurstone scaled set of items reflecting “perceived ease of communication with the teacher” (e.g., Most of the time I understand clearly what my teacher is trying to do). 4. Responses to a questionnaire with smiling-frowning faces indicating how they think the teacher feels about Latin0 children. 5. School self-esteem. 6. Home self-esteem. 7. General self-esteem. This strategy of multimethod measurement is particularly appropriate for the evaluation of cross-cultural training, since any one measure may have special biases and may reflect method variance of no interest to the evaluators. In other kinds of cross-cultural training situations, it would be possible to use a similar strategy. For example, one might examine (1) the expectations of host nationals concerning how visitors, guests, advisers, or foreigners should behave, and how other relevant social roles are defined; (2) the behavioral intentions of host nationals concerning their own behavior toward visitors; and one can obtain (3) evaluations of visitors (well adjusted, pleasure to have in my country); (4) judgments about attributions concerning the causes of behavior of their own fellow countrymen and persons of different nationalities, in a variety of social
International
26
Journal of Intercultural
Relations
situations; also, (5) one can directly observe the behavior of host nationals in relation to visitors. Furthermore, one can study (6) the expectations of the visitors (trainees) concerning how host nationals will behave, (7) behavioral intentions concerning their own behavior toward host nationals, (8) perception of ease of communication with host nations, (9) perception of how host nationals perceive them and “welcome them,” and ( 10) perception of how well they are accomplishing their mission. Discrepancies between expectations of how the other will behave, and the other’s behavioral intentions to behave, or discrepancies between intentions to behave in a certain way and the other’s expectations of what is appropriate behavior, are further measures that can be used to assess the quality of the interpersonal relationship. CONTROVERSIES
IN EVALUATION
RESEARCH
Cross-cultural training can be viewed as resulting in a special kind of social change. People (trainees) are changed before going abroad, or they change after they arrive in their host country. The effects of social change, whether specific, as in the case of cross-cultural training, or more diffuse, as when President Lyndon Johnson attempted to eliminate poverty, are assessed through “evaluation research.” In the sections that follow we will examine controversies in evaluation research and attempt to learn from them, to improve the assessment of the effects of cross-cultural training. Historical
Buckground
Social experiments have been going on for thousands of years, but only recently have social scientists developed methodologies that give hope that such experiments can be evaluated. Educational systems, legal systems, tax structures, and hundreds of social programs have been changing from time to time, and have resulted in different solutions to the same problem in different countries or in different states. But which programs work better ? This question was asked most selfconsciously during the Johnson administration-the War on Poverty and the Great Society. At that point in time, also, social scientists such as Don Campbell were ready to provide the rationale and methodology for conceptualizing reforms as experiments (Campbell, 1969) and to show how quasi-experiments (Campbell &Stanley, 1963) can be done. The interaction of need and possibility was explosive. Soon all kinds
Triandis
27
of approaches to social innovations were being evaluated. The future is likely to bring even more evaluation research, as politicians move toward zero budgeting. Thousands of studies are now being published. Between 1963 and 197 1 the federal government spent 6.8 billion dollars for training 6.1 million people. It is not difficult to see why people wanted to know whether this program worked. Gilbert, Light, and Mosteller ( 1975) provide an excellent overview of the range of programs that have been evaluated during that period, and review the results of many programs, including the emergency school assistance program, the negative income tax experiment, the social workers for multiple families program, the cottage life intervention program of the bureau of prisons, the Manhattan bail project, the Los Angeles sheriff academy, and a large number of medical innovations. They conclude that only one-fifth of these programs was successful. Since experimentally induced social change is very difficult, they argue that program innovators should be rewarded for doing accurate evaluations rather than for the development of successful programs, since learning that something does not work may be as socially valuable as learning that something does work, and the important thing is to learn what works so that we can develop better programs in the future. Social ills have developed because of the operation of forces that have been active for hundreds of years, hence it is difficult to wipe them out with a new program, no matter how much money we spend, Furthermore, even very small effects, if reliable, can be socially very valuable, as the banking and insurance businesses have shown. In a different field, Hilton and Lumsdaine ( 1975) presented a review of evaluations of fertility planning programs around the world. A major dependent variable was the acceptance of particular contraceptive methods. The review showed that there have been numerous good evaluative studies in that field, including several true experiments with randomization. Guttentag (1976) saw a “paradigm shift” in evaluation research in the very near future. She argued that evaluation research is a natural outgrowth of current social changes. In preindustrial society people played a game against nature. Muscle power was the major form of energy. In industrial society machines replaced muscle, and the emphasis was on saving time, being productive and efficient. In postindust~al societies the emphasis will be on service. The major life games will be games among persons. Participation and satisfaction with services will be the important criteria.
28
International
Journal of Intercultural
Relations
Thus as we move from one kind of society to another, the central criterionfor the evaluation of any socialprogram shifts. In preindustrial society it was survival; in industrial, efficiency; in postindustrial, satisfaction with services received. These shifts are also mirrored in the kind of science that each society will emphasize. In the preindustrial, it was the military arts and sciences; in the industrial, the physical sciences and engineering, rationality, cause and effects relationships, invarient laws, mind over matter; in postindustrial society, the social and behavioral sciences, the creation of innovative social settings, information systems, services to all, and mastering “games” among persons. The diversity of social groups is an essential element of postindustrial society. Hence the needs and points of view of different people will have to be considered. Subjectivity will be studied as an essential element of the new society. Hence evaluation research will become a central activity of this society. Moving away from the notion that there is one best way to do things (Fred Taylor’s central concept of “scientific management” at the turn of the century), we will move to the many ways to do what different kinds of people need. Participation in decisions will become more widespread. The central question will no longer be “Is this a good program ?” but “Is this a program that meets the diverse objectives of all the relevant groups?” Furthermore, evaluation research will be used as a means of changing programs, as a constant feedback is established between program development and change on the one hand, and evaluation on the other hand. Consistent with this analysis is Maruyama’s (1976) contrast among four ways of conceptualizing social systems: The hierarchical admits of only one “truth” developed by experts. What is good for one is good for all members of the society. Programs are good if they meet the criteria stated as goals by the experts. The individualistic assumes that what is good for one individual is unrelated to what is good for another. There is a random distribution of events. The homeostutic assumes that balance is a virtue, and social harmony the greatest good. Change is seen to be balanced by other change, so that the system corrects any deviations. Finally, the morphogenetic social system assumes that there is constant differentiation, increased structure, and increased information flow within the social system (with positive ‘feedback) which corrects changes that are too extreme. In the first system, planning is done by experts and evaluation should be by experts. In the second, planning should allow everyone to
Triandis
29
do his/her own thing. Evaluation might employ a self-anchoring scale (e.g., Cantril, 1965) and people asked to state their personal goals. In the third, maximum participation, arranged so that all agree on a course of action (new program) which does not disturb the balance, is visualized. Finally, in the last, new programs are seen as welcome, and accommodation of new elements is seen as natural. Maruyama states that the West, Islam, and India use a hierarchical way of conceptualizing social systems; the Chinese village is a homeostatic system, and the Japanese use both homeostatic and morphogenetic systems. The correspondence between Guttentag and Maruyama is clear. The industrial system is hierarchical; the postindustrial requires some of the viewpoints of the other three Maruyama systems, particularly the morphogenetic. More Recent Developments A paper by House (1976) explores ethical issues in evaluation research. This paper advocates moving away from the utilitarian ethics of the past (presumably characteristic of industrial society) to a new ethic dominated by the concept of fairness. The ethics advocated by House are based on Rawls’s theory of justice-as-fairness. It moves away from accountability and the measurement of specific utilitarian attributes of programs such as verbal ability or performance in schools. The argument is that there are more important variables that usually remain unmeasured in single-variable evaluation research such as a person’s self-esteem. House’s viewpoint. The argument is developed from the point of A Theory ofJustice by John Rawls. According to Rawls, the good cannot be established prior to and independently from what is right, and the right is not equivalent to maximizing what is good. Rather, the right must be established before the good. Each person is assumed to have nonnegotiable rights which cannot be bargained away no matter how it affects the goods. Thus the first step is the development of agreements on how the good is to be distributed, with each individual being free to determine his/her own good, provided it conforms with what is right. Rawls’s theory includes the conception that “all social valuesliberty and opportunity, income and wealth, and the bases of selfrespect-are to be distributed equally unless an unequal distribution of
30
International
Journal of Intercultural
Relations
any, or all, of these values is to everyone’s advantage” (1971, p. 62). The first principle of Rawls’s system concerns the equality of basic liberties. People‘are said to have a right to the most extensive system of basic liberties compatible with a similar system of liberties to all. The second argues that social and economic inequalities are acceptable only if they benefit the least advantaged, and if they are open to all under conditions of fair equal opportunity. For example, it is acceptable for some people to have high incomes if such incomes are necessary to attract people to positions that would benefit the least advantaged. Basic liberties include the right to self-esteem. Thus in program evaluation it is essential to ask how the self-esteems of those touched by the program are affected. If information would injure the self-esteem of persons it should not be included in an evaluation report, says House. He goes on to criticize evaluation efforts based on simple criteria, such as scores on a vocabulary test, which ignore the more “important” criteria, such as self-esteem. He further criticizes evaluation efforts that do not consider explicitly the role of values in deciding what to study, how to study it, and how to analyze the data, and gives several examples of the influence of values. Thus it is essential to make values explicit and to consider the fact that different people have different values. Evaluation research should consider the meaning of any criterion in the context of the values of those affected by the program. One of the frequent problems is that those affected may not be represented, or may not have a voice, in the design or the data interpretation phases of the evaluation project. The evaluator must make an effort to correct this deficiency, particularly when those that are not represented are the least advantaged. The final product of an evaluation should be in the form: program P is good on dimension D, for sample S. A perspective on shifting emphases. A common thread in the Guttentag, Maruyama, and House viewpoints just outlined is the shift away from single or simple measurements, representing the decisions of the evaluator concerning what is important, and the development of a strategy that reflects the viewpoints of multiple “clients,” particularly the people who are affected by the evaluation. This represents a major shift from the classic procedures outlined by Tyler (1950). Recall that Tyler evaluated educational experiments by means of student test scores. Gains in such scores were considered “good” while evidence of “no gain” suggested that the educational procedure should be rejected. The approach seems to be inadequate both from the point of view of
Triandis
31
ethics and methodology. From the point of view of ethics one can question whether the procedure is designed to improve the lot of the least advantaged. Clearly this was not the case. On the contrary, it provided a way to justify inequalities and the economic exploitation of minorities as reflected in wage differentials. Me~odologically the procedure has been criticized because of the frequent use of inappropriate control groups that were prone to results due to regression artifacts (Campbell & Erlebacher, 1975). The conceptions of evaluation research that prevailed up to the 1950s were replaced by conceptions developed by CampbelI and his associates in the 1960s (e.g., Campbell&Stanley, 1963; Campbell, 1969). In their turn, these conceptions apparently are being challenged in the 1970s. The Handhook of Evaluation Research apparently reflects more the 1960s than the 1970s so that an effort has been made to summarize the new trends in Review Annuals, such as the one edited by Glass (I 976). The critics of the mid- 1970s are arguing that earlier conceptions were inadequate because their dependent variables were inappropriate and their experimental designs too narrow to reflect the multiplicity of viewpoints that should be represented in evaluation research. The critics charge, among other things, that evaluation situations are characterized by numerous relevant parameters (e.g., attributes of the teacher, students, curriculum, learning situation, and institution) which interact in complex ways (Cronbach, 1975), and do not lend themselves to the neat experimental designs with manipulation of key variables of the Campbell-Stanley (1963) variety. Furthermore, such designs assume that innovations remain more or less invariant during the evaluation, and that educational programs given the same labels are in fact similar. In reality, say the critics, programs with labels, or the same program at different times, may be highly heterogeneous. A rigid design does not get “the feel” for the effects of such variations. In addition, traditional evaluation procedures are too restricted in scope, obtaining measurements on only a few dependent variables and ignoring the important variables such as the students’ self-esteem, parental satisfaction, success of the student in life, or staff satisfaction with the program. Moreover, the traditional approach, while looking at major effects, misses the smaller perturbations and unusual effects. A lot of sources of variance are relegated to error terms, a lot of interactions are ignored or artificially eliminated, and the complexity of reality is distorted to the point that what emerges is a caricature without validity. Given the fluid state of evaluation research, the resulting situation is
32
International
Journal of Intercultural
Relations
confusing and complex. Numerous approaches are being proposed, each with a half-life of a few years. Among the main types of approaches Stake (1976) has identified the following: 1. Institutional self-study-committees set standards for staff performance, discuss how much has been accomplished to meet goals (e.g., Dressel, 197 1). 2. Blue-ribbon committee visit-outstanding experts visit, review existing data, interview key people, write reports (e.g., Havinghurst, 1964). 3. Utilitarian-economic data, lists of options, estimates of costs and benefits, efficiency computations (e.g., Alkin, 1967). 4. Social policy analysis-measures of social conditions with several instruments (e.g., Coleman, 1972). 5. Goal-free evaluation-to avoid undue influence by advocates of the program (e.g., Striven, 1974). 6. Adversary evaluation-to consider simultaneously the viewpoints of advocates and critics (e.g., Levine, 1973). 7. Classic-Campbell and associates. 8. Student-gain by testing-Tyler ( 1950). 9. Transaction-observation-case studies, classroom observations, interviews with persons with heterogeneous perspectives and informal methods (Parlett & Hamilton, 1972; Stake, 1977). Each of these approaches should be useful for some situation. Ideally, the best evaluation strategy is a reasoned balance utilizing a little of each of these approaches. There are no good ways to prescribe how one is to arrive at this balance. However, considering the political and ethical problems associated with evaluation research should help one arrive at a balanced conceptualization. Political Problems Conflict is at the heart of evaluation research. Basically, the program administrator does not want to be evaluated, and if he accepts the evaluation it is because politically she/he cannot afford to reject it. Gurel (1975) has discussed some of the conflicts associated with evaluation. Most organization members are not used to being evaluated and get upset at the very thought of evaluation. The program manager wants to prove the program successful, but the evaluator often does not care.
Triandis
33
Their objectives are often in conflict: the manager has many secondary goals which the evaluator does not even consider. But the evaluator needs a more or less constant program if he is to arrive at any conclusions. Another area of conflict is over the relative importance of different kinds of goals or criteria. Many evaluators rely on economic measures. For example, Rothenberg (1975) and Levin (1975) outlined cost-benefit strategies. Zeckhauser (1976) discussed how dollars can be used to measure the value of human life. He quoted a source which assigned $200,000 to the value of the life of a worker (based on labor market considerations). Levin (1975) distinguished different kinds of analyses: cost-benefit (which attempts conversion of everything to dollars), costeffectiveness (which considers outcomes that are not convertible to dollars, such as clean air or mental health), and cost-utility (which inco~orates the decision maker’s own values, and hence may result in different results for different decision makers). Costs are social sacrifices (e.g., value of time of program participants, cost of books, days away from work). Opportunity costs are losses of social benefits that could have been obtained if funds had not been expended on a particular program. Sjoberg ( 1975) is concerned about the identification of the evaluator with the dominant power structure of a given society. He raises issues similar to those raised by House and reviewed above. A key issue is of course what dependent variables will be selected by the evaluator. Are these the variables that will show a particular program in the best light? The worse light? Who will benefit if the results are positive? Negative? Such questions must be asked and answered by those planning an evaluation well before they decide on a research design. One approach that is advocated by many evaluators now is that one do “formative” rather than “summative” evaluation. In formative evaluation one evaluates a constantly changing program. in summative one evaluates the sum of several efforts-a total program. Summative evaluation is not very different from ordinary research, while formative makes evaluation an aspect of program development. The distinction between evaluation and ordinary research is made sharply by Moore (1976). Research, he argues, attempts to simulate real life. In evaluation research one tries to simulate the laboratory-to gain the kinds of control one has in the laboratory. In most research one determines what the independent and dependent variables are to be and
34
International
Journal of Intercultural
Relations
has much control. In evaluation the independent variable is often a program and one has little control over it. Variables are not clean and one cannot be sure that contextual variables remain constant. Determining the goals of a program is an essential step before one starts an evaluation. Obviously, this is largely a political action. Furthermore, says Moore, one needs to get commitments from program participants that they will change their actions, depending on the results of the evaluation. One needs also to get agreement about the major features of the research design (e.g., control groups) and about the dependent variables (self-report, behavioral, unobtrusive, etc.). Hastings (1976) complained about the extent to which evaluation is conceptualized as something that funding agencies want done and not enough as something that people want. He was particularly critical of evaluators who do not consider problems of timing. Evaluation should be done only after the program has reached some degree of stability. Evaluators should not decide about goals before they know much about the program. One should not decide about goals and freeze. Rather, one should look at the program from many angles and then decide. Such a view is compatible with responsive evaluation as advocated by Stake (1976). Conceptualizing
Evaluation
Evaluation is not just learning about a program; it is not just “knowing” a program. Rather, it is changing one’s attitude about a program. There is a danger that the evaluator will become coopted by program staff and managers. Hence, some sort of adversary procedure is needed. One must aim at getting and giving information to those persons who are most concerned about the program (Stake, 1976). Thus in conceptualizing an evaluation one needs to consider ( 1) what activities are involved, (2) what are the requirements for information about these activities by different groups of people, and (3) what values, goals, or criteria are to be used so that different groups of people will find the evaluation most helpful. One should not limit oneself to performance measures. Changes in opinions or feelings are equally important. Aesthetic experiences can be as legitimate as dollars, and criteria might reflect aesthetic judgments just as legitimately as costs and benefits. In short, the evaluator is a facilitator of decision making, giving to different groups information needed to change their programs. To accomplish these goals the evaluator should start by talking with the
Triads
35
clients, then identifying the scope of the program, looking at the activities that are part of the program, inductively discovering what is accomplished by such activities, and only then conceptualizing the program. Once this is done the evaluator would be in a position to identify the data needs of different groups. Then observers and measurements can be selected and observations made that center around three foci: antecedents, transactions, and outcomes of different activities. One can then ask whether these activities conform to standards, or are confirming/disconfirming the conceptualization of the program. This step is a kind of validation of the program. Now the information can be assembled into a final report which is presented to the clients. Thus, both at the start and at the end of the evaluation, the evaluator talks with clients. Stakes’s conceptualization of evaluation research, while richer and broader, is similar to the suggestions of Lumsdaine and Bennett (1975). They focused on five steps: 1. 2. 3. 4. 5.
Determine the goals and indices. Have they been reached? Determine the processes-how does one reach these goals? Compare the processes with the original plans. What are the impacts of the program? What is responsible for the effects?
Stake takes this further and shows how to ask these questions repeatedly and ask them of different groups of concerned persons. Similarly, Guttentag and Edwards advocate that we make explicit the values of different groups before we design an evaluation. Furthermore, we need some sort of primitive theory relating activities to outcomes. Groves (1976) presented a theory of classroom behavior arguing that teachers behave one way when they like a child and a different way when they do not. One can then use multiple observations of teachers’ behavior (e.g., how frequently teachers reprimand or ask a child a question) to obtain an indirect measure of liking. Then we can see if a particular program increases or decreases teacher liking of their pupils. Methodological
Issues
Is a social program constant enough to be considered a basis for constructing an independent variable? Have the objectives of the various groups been incorporated in the research design? Who will control the information? How is it to be used? How are the data to be aggregated?
36
International
Journal of Intercultural
Relations
How does one put together different perspectives? What kinds of experimental designs should be used? Random assignment? Such questions are frequently asked in connection with evaluation research. The answers are given from different perspectives and there is little agreement among the experts. Specifically, the experimental and quasi-experimental design advocates see programs as sufficiently constant to allow for the use of their approach, but their critics, and what appears to be the majority of evaluation researchers in 1976, appear to doubt that programs are sufficiently constant. Most of the mid- 1970s approaches attempt to take into account the viewpoints of various groups concerned with the results of an evaluation. Typical of this approach is Edwards and Guttentag (1975) who advocate the following steps: 1. Identify the person or organization whose utilities are to be maximized. If several persons or organizations are concerned, all of their utilities must be maximized. 2. Identify the issues and decisions to which the utilities needed are relevant. 3. Identify the entities to be evaluated. 4. Identify the relevant dimensions of value-what dimensions are important for the evaluation of the entities, from the perspectives of different people (groups) to arrive at different decisions. 5. Rank the dimensions in order of importance. The ranking can be done by an individual or by representatives of different groups acting as a group. 6. Rate dimensions in importance, preserving ratios. Start by assigning to the least important dimension a value of 10. Now take the next least important dimension. How many more times is it more important? For instance, if one dimension is 10 and another 80, it means that the latter is eight times more important than the former. 7. Sum the importance ratings divide by the sum and multiply by 100. This results in weights that are like probabilities. 8. Measure the location of each entity on each dimension. “Measure” here can be subjective, or objective. Transform all measures to a 0 to 100 scale. 9. Now compute the utilities for each entity, taking the weights from step seven and the output from step eight. The sum of the weighted utilities is the aggregate utility. Ui = I;jwjuij.
Triads
10. Decide. For a single act maximize needed, maximize XU.
37
U. If several decisions
are
Who will control the information? It is important to think of the least advantaged. Are they going to be hurt by the info~~tion? How can the information be channeled to prevent this? In presenting the data to different audiences one can ensure that individuals are protected, i.e., anonymity guaranteed, and group data are aggregated to the point which is essential for sound decisions, but not so as to harm the least advantaged participants.
Ra~du~ assignment. The hottest issue in this literature is around the question of random assignment. There are all kinds of arguments favoring random assignment. The most convincing is the paper by Campbell and Boruch (1975) which suggests that six ways that have been suggested to do quasi-experiments, in compensatory education, tend to underestimate the effects. Boruch ( 1976) reviews arguments given by advocates of nonrandomization and finds no merit in them. Randomization is not impossible, he claims. It requires insistence, but there are many examples where it has been used. Nor is it unethical. When one has limited resources, not all can get the benefits of a program. There is no better way to decide who is to benefit than by randomness. Wortman, Hend~cks, and Hillis (1976) examine factors affecting participant reactions to random assignment. In an experiment in which they assigned subjects to an “aware” condition (subjects told that there would be experimental and control groups and that they must volunteer first and assignment to groups will be done later), an “unaware” condition (subject told nothing), and a “becoming unaware” condition (subjects found out about experimental and control groups “accidentally” from another subject) they showed that the “aware” felt about the experiment the same way as the “unaware,” but the “becoming aware” fett quite negatively. An essential element according to Wortman et al., is that the subjects had a choice about participation since in a replication which lacked this element the results were different. In spite of the emphasis on randomization, there are situations when it is not possible. When an evaluator is asked to come in to study an ongoing program it is difficult to use randomization. There are many other situations in which it cannot be used (Striven, 1976). Sherwood, Morris, and Sherwood (1975) propose a procedure of
38
International
Journal of Intercultural
Relations
matching groups on variables that are not correlated with the dependent variables. This approach avoids the obvious regression artifacts that are associated with usual matching designs. However, it does not control for the possible unrepresentativeness of the experimental and matching groups (a selection bias). While one can sympathize with the practical realities of evaluation research which may make randomization impossible, it seems that one must insist on it. Analyses. The emphasis in evaluation research is on multivariate designs (Eber, 1975) and regression analysis (Cohen, 1975). However, the most controversial issue is whether one should utilize Baysian or traditional statistical perspectives. Guttentag and others have emphasized the Baysian perspectives, on the grounds that evaluation research concerns making decisions, to maximize subjective expected utility. If one is absolutely sure that a program works, one does not need data to “prove it,” and if the data suggest that the program does not work, the convinced advocate can find enough “weaknesses” in the data, the analysis, or the interpretation to reject the data. Thus, the key question is: how much evidence does the evaluator need to change his or her perception of the program? To answer this question Baysian statistics are completely appropriate. Consider a set of actions which have different utilities, depending on the state of the world (Lewis, 1976). Suppose the utility of a given action a when the state of the world is 8 is u (a,0). We want to know what is the state of the world (8) given that we have observations or data (x). In short, what is the probability that the state of the world is 8 given data x? It is p (0 1x). The subjective expected utility for action a depends on the utility of a when there are different states of the world 8i. We want to maximize this utility which means we want to maximize 2 u (a, Oi) p (Oil x) So, we would look at different as and pick the one that gives us the maximum utility. Before we start our investigation we have a prior distribution of 8, which reflects our conception of the state of the world. Now we collect data. We consider the probability that we will obtain data like x, if the state of the world is 8. This is p (x 18 ). The Baysian theorem gives us
39
Triandis
the posterior probability that the state of the world is indeed 0, if we know the p(0), p (x), and p (xl 0) quantities. Namely, P@IX)
=
PC@>
x p (XIW
p(x) Novick and Jackson (1974) give details of how to do this, and provide tables necessary to do the job. Applications TheHandbook, as well as the books by Bennett and Lumsdain ( 1975) and Glass ( 1976)) provide broad coverage of applications of evaluation research. Examples from the areas of mental health, family intervention, career planning, population research, educational evaluation, welfare and social services, and crime and justice are presented. Numerous ingenious procedures are advocated, that appear to solve particular practical problems. The major difficulty that it identified is the resistence of administrators to the utilization of evaluation research results, even when these results are sound. David and Salasin (1975) discussed this problem in detail and presented a model of program attainment. According to this model: B = E + T + S + (P + H) (Dx C) - I Where B stands for behavior or program attainment; E stands for values that give purpose-such as “this approach is just what our situation calls for” vs “approach is a poor fit with our program”; T stands for timing and involves a judgment of how timely it is to introduce the innovation; S stands for stimulus conditions and reflects the circumstances that prevail at the time-such as “things that are happening make it essential to adopt this approach right now” vs. “this is a poor time to adopt the approach”; P stands for patterns of ideas for proposed action-such as “courses of action are clear,” “ program appears to be logical and based on solid material”; H stands for habit strength and reflects rewards that are anticipated by the change-such as “I’ll feel personal benefits from this change,” “ much to be gained”; D stands for drive or felt needsuch as “program seems necessary” vs. “optional,” “I feel motivated to participate”; C stands for capacity or ability-such as “I’ll have no trouble doing it, ” “it can be done for negligible time investment,” “approach is easy to carry out”, . I stands for inhibitors or resistance-
40
International
Journal of Intercultural
such as “I would be reluctant to adopt,” consequences.”
Relations
“I see a number of negative
WHAT CAN BE LEARNED FROM EVALUATION RESEARCH FOR USE IN CROSS-CULTURAL TRAINING?
The major point that emerges from a consideration of the controversies within the field of evaluation research is that current trends are toward broadening this type of research. One is to ask more questions in many more forms. One important way of broadening this research is to ask whether all the relevant points of view are represented. For example, if an American corporation has a branch in Thailand, and trains Americans to live in Bangkok, there are a number of populations whose points of view might be considered: (1) the top management of the corporation, (2) the immediate supervisors of the trainees, (3) the trainees themselves, (4) the friends of the trainees, both at work and in the city or country, (5) the relatives of the trainees, both close, such as nuclear family, and remote, (6) the Thai counterparts, (7) the supervisors of the Thai counterparts, (8) the Thai clients, (9) the Thai government, and (10) samples of Thai society. Another important conclusion is that a program is not just “good” or “bad,” but it may be good or bad for different purposes and for different populations. Criteria will shift with time and place. One must measure all important variables that may be affected by a program. Fundamental questions have to be faced: for example, what if the training makes the American “too effective,” and the Thai clients lose their money and are very unhappy? Still another question is the management of information. How is one to tell a person that he is not trainable and should not go abroad? How is one to incorporate diverse values in designing a program? If Rawls’s viewpoint is taken seriously, it raises some very important ethical issues for the cross-cultural trainer. Rawls would argue that if there is an inequality between the visitor and the host, the inequality must be for the benefit of the least powerful. For example, if an advisor and a less well-informed host national are to interact, Rawls would have them be equal in income, status, and so on, though they can be unequal on the amount of information each has because the expert’s additional information can benefit the host national. Furthermore, the information that the visitor from the developed
Triads
41
countries gives to the host national can be examined from within Rawls’s viewpoint. If the information results in the host national having a lower self-esteem, without benefiting him in some other important way, ethical considerations dictate that the information should not be given. Thus within that perspective, the behavior of missionaries is unethical, unless one accepts the assumption that the natives improve their status in heaven! This raises thorny ethical problems that crosscultural trainers must not ignore. Cross-cultural training benefits some persons. But we can also ask: Whose utilities are to be maximized? Should the agency that pays for the cross-cultural training set the criteria? Should we consider the viewpoints of many types of host nationals? Should we consider the viewpoints of many types of persons from the trainee’s country? What ethical principles are to be used in deciding whose utilities are to be given priority? Finally, the recent change in viewpoint, among evaluation researchers, from summative toformative evaluation has implications for crosscultural training. It suggests that we should not set up our evaluations of such training on a one-shot basis, but rather should develop a continuous feedback of information concerning such training. Such feedback should be used in changing the training. The input of both the trainees and the host nationals with whom the trainees will interact should be included in the feedback. The costs of training must also be considered. These include the time of the program participants, expenses for trainers, materials, books, days away from work, and so on. Such costs must be balanced against benefits, e.g., increases in self-esteem. Many of the procedures outlined in the methodology section above permit incorporation of diverse points of view and a broad analysis of the effects of the program. Thus in this paper the major independent variables of cross-cultural training were explored, some of the dependent variables were mentioned, and then, after examination of the controversies in the field of evaluation research, it was concluded that a much broader perspective on the evaluation of cross-cultural training is in order. REFERENCES ALKIN, M. C. Toward an evaluation model: A systems approach. Unpublished script, 1967, UCLA Center for the Study of Evaluation.
manu-
42
International
Journal of Intercultural
Relations
BENNETT, C. A. & LUMSDAINE, A. A. Evaluation and experiment. New York: Academic Press, 1975. BORUCH, R. F. On common contentions about randomized field experiments. In G. V. Glass (Ed.) Evaluation studies: Review annual, vol. I, Beverly Hills: Sage, 1976, 158-194. BRISLIN, R. & PEDERSEN, P. Cross-cultural orientation programs. New York: Gardner Press, 1976. CAMPBELL, D. T. Reforms as experiments. In E. L. Struening and M. Guttentag (Eds.)Handbookufeva/uation research, vol. 1. Beverly Hills: Sage, 1976,7 I- 100. CAMPBELL, D. T., & BORUCH, R. F. Making the case for randomized assignment to treatments by considering the alternatives: Six ways in which quasi-experimental evaluations in compensatory education tend to underestimate effects. In C. A. Bennett and A. A. Lumsdaine (Eds.), Evaluation and experiment. New York: Academic Press, 1975, 195-296. CAMPBELL, D. T., & ERLEBACHER, A. How regression artifacts in quasiexperimental evaluations can mistakenly make compensatory education look harmful. In E. L. Struening and M. Guttentag (Eds.), Handbook ofevaluation research, vol. 1. Beverly Hills: Sage, 1976, 597-617. CAMPBELL, D. T., & STANLEY, S. C. Experimental and quasi-experimental designs for research. Chicago: Rand McNally, 1963. CANTRIL, H. Thepatternofhuman concerns. New Brunswick, N.J.: Rutgers University Press, 1965. COHEN, J. Multiple regression as a general data-analytic system. In E. L. Struening and M. Guttentag (Eds.), Handbook of evaluation research, vol. 1. Beverly Hills: Sage, 1976, 57 l-595. COLEMAN, J. S. Policy research in the social sicences. Morristown, N.J.: General Learning Press, 1972. CRONBACH, L. J. Beyond the two disciplines of scientific psychology. In G. V. Glass (Ed.), Evaluation studies: Review annual, vol. 1. Beverly Hills: Sage, 1976, 15-27. DAVID, K. Intercultural adjustment and applications for reinforcement theory to problems of “culture shock.” Trends, 1972, 4, l-64. DAVIS, H. R.,-& SALASIN, S. E. The utilization of evaluation. In E. L. Struening and M. Guttentag (Eds.) Handbook of evaluation research, vol. I. Beverly Hills: Sage, 1976, 621-666. DRESSEL, P. L. Accreditation and institutional self-study. North Central Association Quarterly, 197 I, 46, 277-287. EBER, H. W. Multivariate methodologies for evaluation research. In E. L. Struening and M. Guttentag (Eds.), Handbook of evaluation research, vol. I. Beverly Hills: Sage, 1976, 553-570. EDWARDS. W., & GUTTENTAG, M. ExperimentsLand evaluations: A reexamination. In C. A. Bennett and A. A. Lumsdaine (Eds.), Evaluation and experiment. New York: Academic Press, 1975, 409-463. FIEDLER, F. E., MITCHELL, T., & TRIANDIS, H. C. The culture assimilator: An approach to cross-cultural training. Journal of Applied Psychology, 1971, 55, 95-102. GILBERT, J. P., LIGHT, R. J., & MOSTELLER, F. Assessing social innovations: An
Triandis
43
empirical base for policy. In C. A. Bennett and A. A. Lumsdaine (Eds.),Evaluarion rind e.xperiment. New York: Academic Press, 1975, 39-193. GLASS, G. V. ~val~t~on studies: Review annuals, vol. 1. Beverly Hills: Sage, 1976. GROVES, C. Lecture to evaluation seminar, 1976. GUREL, L. The human side of evaluating human serivces programs: Problems and prospects. In E. L. Struening and M. Guttentag (Eds.), Handbook of evaluation research, vol. 2. Beverly Hills: Sage, 1976, f I-28. GUTTENTAG, M. Presidentiai address to the American Psychological Association Division 8, Washington, D.C., September 1976. HASTINGS, T. Lecture to evaluation seminar, 1976. HAVINGHURST, R. J. i%r public schools of Chicago. Board of Education: City of Chicago. 1964. HILTON, E. T., & LUMSDAINE, A. A. Field trial designs in gauging the impact of fertility planning programs. In C. A. Bennett and A. A. Lumsdaine (Eds.), Evaluation and experiment. New York: Academic Press, 1975, 3 19-408. HOUSE,’ E. R. Justice in evaluation. In G. V. Glass (Ed.), Evaluation studies: Review annual, vol. 1. Beverly Hilts: Sage, 1976, 75-100. KRAEMER, A. A cultural self-awareness approach to imp~vjng intercul~r~ communication skills. Alexandria, Va.: Human Resources Research Organization, April, 1973. LANDIS, D., DAY, H. R., MCGREW, P. L., THOMAS, J. A., & MILLER, A. B. Can a Black “culture assimilator” increase racial understanding?Journal ofSociu~ Issues. 1976, 32, 169-183. LEVIN, H. M. Cost-effectiveness analysis in evaluation research. In E. L. Struening and M. Guttentag (Eds.), Handbook of evaluation research, vol. 2. Beverly Hills: Sage, 1976. LEVINE, M. Scientific method and the adversary model: Some preliminary suggestions. Eval~tion Comment, 1973, 4, l-3. LEWIS, C. Lecture to evaluation seminar, 1976. LUMSDAINE, A. A., & BENNETT, C. A. Assessing alternative conceptions of evaluation. In C. A. Bennett and A. A. Lumsdaine (Eds.), Evaluation and experiment. New York: Academic Press, 1975, 525-553. MARUYAMA, M. Lecture to evaluation seminar, 1976. MOORE, T. Lecture to evaluation seminar, 1976. MURDOCK, G. P. World ethnographic sample. American Anthropologist, 1957, 59, 664-687. NOVICK, M. R., &JACKSON, P. H. Statistical methodsfor educational andpsychological research. New York: McGraw-Hill, 1974. OSGOOD, C. E. Method and theory in experimental psychology. New York: Oxford Press, 1953. PARLETT, M., & HAMILTON, D. Evaluation as illumination: A new approach to the study of innovatory progress. In G. V. Glass (Ed.), Evaluation studies: Review annual, voI. 1. Beverly Hills: Sage, 1976, 140-157. RAWLS, J. A theory ofjustice. Cambridge, Mass.: Balknap, 1971. SCRIVEN, M. Pros and cons about goal free evaluation. In W. J. Popham (Ed.), Evaluation in education: Current applications. Berkeley: McCutchan, 1974. SCRIVEN, M. Maximizing the power of causa1 investigations: The modus operandi
44
International
Journal of Intercultural
Relations
method. In G. V. Glass (Ed.), Evuluarion studies: Revliew annuul, vol. 1. Beverly Hills: Sage, 1976, 101-l 18. SHERWOOD, C. C., MORRIS, J. N., & SHERWOOD, S. A multivariate nonrandomized matching technique for studying the impact of social interventions. In E. L. Struening and M. Guttentag (Eds.), Handbook of evaluation research, vol. 1. Beverly Hills: Sage, 1975, 183-224. SJOBERG, G. Politics, ethics and evaluation research. In E. L. Struening and M. Guttentag (Eds.), Handbook of evaluation research, vol. 2. Beverly Hills: Sage, 1975, 29-51. STEWART, E., DANIELIAN, J., & FOSTER, R. Simulating intercultural communication through role playing. Humrro, technical report no. 69-7, pay, 1969. STRUENING, E. L., & GUTTENTAG, M. Handbook of evaluation reseurch. Beverly Hills: Sage, 1976. TRIANDIS, H. C. Culture training, cognitive complexity and interpersonal attitudes. In R. Brislin, S. Bochner, and W. Lonner (Eds.), Cross-cultural perspectives on learning. Beverly Hills and New York: Sage and Wiley/Halsted, 1975, 39-77. TRIANDIS, H. C. (Ed.). Variations in black und white perceptions of rhe social environmenr. Urbana: University of Illinois Press, 1976. TRIFONOVITCH, G. On cross-cultural orientation techniques. Topics in cultural learning, 1973, I, 38-47. TYLER, R. W. Basicprinciples of curriculum and instruction. Chicago: University of Chicago Press, 1950. WELDON, D., CARLSTON, D. C., RISSMAN, A. K., SLOBODIN, L., & TRIANDIS, H. C. A laboratory test of effects of culture assimilator training. Journal of Personality and Social Psychology, 1975, 32, 300-310. WOLPE, J. Psychotherapy of reciprocal inhibition. Stanford: Stanford Press, 1958. WOLPE, J. The practice of behavior therapy. New York: Pergamon Press, 1969. WORTMAN, C. B., HENRICKS, M., & HILLIS, J. W. Factors affecting participant reactions to random assignment in ameliorative social programs. Journal of Personality and Social Psychology, 1976, 33, 256-266. ZECKHOUSER, R. Procedures for valuing lives. In G. V. Glass (Ed.), Evaluation studies: Review annual, vol. I Beverly Hills: Sage, 1976, 28-74.
ABSTRACT
TRANSLATIONS
MARCO TEORICO PARA EVALUAR LA EFICACLA DE LA CAPACITACION TRANSCULTURAL La primera parte presenta un esquema tedrico que incluye las variables independientes y dependientes que deben ser tenidas en cuenta en 10s estudios proyectados para valorar la eficacia de la capacitacidn transcultural. Se consideran seis tipos de capacitacidn, a saber, general o especifica, afectiva, cognoscitiva o de comportamiento, ademas de auto “insight.” Se toman en cuenta ademis la cantidad y oportunidad de1 moment0 de1 capacitador, asi coma las caracteristicas de adiestradores y adiestrados. En la segunda parte se estudian las Ias estrategias para valorar 10s efectos de estas variables independientes. Se estudian las controversias existentes sobre la investigacidn sobre la valoracidn con el fin de aprender lo que puede lograrse de las experiencias pasadas que puedan ser aplicadas a la capacitacidn en temas transculturales. Este anilisis nos muestra que existe una ampliacidn en este tipo de investigacidn que lleva a usar un mayor ntimero de
Triandis
45
variables dependientes, con medidas obtenidas de poblaciones muy diversas. Las nuevas metodologias que permiten incorporar nuevos puntos de vista son examinadas asi coma 10s efectos de 10s programas de capacitacidn. Se plantea tambien el problema de quie’n debe tener acceso a la informacidn adquirida mediante un estudio de evaluacidn. Finalmente, el trabajo examina 10s problemas e’ticos que no pueden ser ignorados por quienes hacen valoraciones de la capacitacidn transcultural. DONNEES ~EORIQUES POUR EVALUER L’ENTRAINEMENT TRANSCULTUREL
L’EFFICACITE
DE
La premiere partie presente les donnees theoriques incluant les variables inde’pendantes et dependantes qui devraient 2tre conside’rees dans les etudes ayant pour but d’e’valuer l’efficacite de l’entramement transculturel. Six sortes d’entrainement, general ou specifique, affectif, cognitif ou behavioral, plus l’entrainement de “retour-sur-soi” sont consider&. En plus, la quantite’ et le temps de I’entralnement, les attributs des entraineurs et entrain& sont consider&. La seconde partie considere les strategies pour l’e’valuation des effets de ces variables ind~~nd~tes. Les discussions dans la recherche d’~valuation sont amenees afin de trouverce qui peut etre appris de l’experience pass&e qui pourrait 2tre applique’ i l’entrainement transculturel. Cette revue montre une direction pour un eiargissement de la recherche d’&aluation, utilisant plusieurs autres variables dipendantes, avec des mcsures provenant de plusieurs sortes de personnes. Les nouvelles methodologies qui permettent l’incorporation de divers points de vue et une large analyse des effets du programme d’entrainement sont examines. Des questions sont aussi soulevees concemant la ou les personnes qui devraient avoir I’information durant une etude evaluative. Finalement, ce texte examine les problemes ithiques qui ne peuvent etre ignores par ces entrainements e’valuatifs transculturels.