INFORMAL LOGIC XIII. 1, Winter 1991 Testing Fairmindedness ALEC FISHER University of East Anglia 1. Introduction Richard Paul is well-known for his advocacy of "strong" critical thinking, that complex of practices and virtues which includes fairmindedness as one of its most important elements. There are many places in his writings where Paul discusses fair- mindedness; here is just one example, taken from the Critical Thinking Handbook:K-3 (p.7): To think critically about issues we must be able to consider the strengths and weaknesses of opposing points of view. Since critical thinkers value fairmindedness, they feel that it is especially important that they entertain positions with which they disagree. They realize that it is unfair either to judge the ideas of another until they fully understand them, or act on their own beliefs without giving due consideration to relevant criticisms. The process of considering an opposing point of view aids critical thinkers in recognizing the logical com- ponents of their beliefs, (e.g. key concepts, assumptions, implications, etc.) and puts them in a better position to amend those beliefs. This passage is quite characteristic of Paul's writings on the subject, and neatly articulates the basic idea and the value attaching to it. Given the importance of fairrnindedness for Paul, it may come as no surprise to learn that he recently constructed a multiple-choice test of fairmindedness. If the test was success- ful, the intention was to include it as a sub- test within a revised version of the Watson- Glaser Critical Thinking Appraisal (hence its name, Critical Thinking Appraisal; Fairmindedness Subtest); however there were problems with Paul's test. These problems are instructive, and this paper explains them in the hope that this will con- tribute to their solution, since a test of fair- mindedness would be of great value to many teachers of critical thinking, both in itself and as a teaching instrument. Paul's test was intended to help discriminate what he calls uncritical thinkers from closedminded critical thinkers and fairminded critical thinkers. Uncritical thinkers are simply poor at reasoning things through (and would be expected to score relatively poorly on any well-designed test of critical thinking). Closedminded critical thinkers (often called weak critical thinkers by Paul) are good at reasoning up to a point (they are "clever in argument" and would be expected to score relatively well on existing tests of critical thinking). However, they use this skill narrowly, in particular they use it only in pursuit of their own interests and in defence of their own point of view and they do not question these. Fairminded critical thinkers, on the other hand, are skilful in argument (and would score well on standard tests) but they apply that skill just as readily when their own beliefs are challenged or when their own interests are at risk. That is to say, even when their own position is threatened, fair- minded critical thinkers will take seriously viewpoints and perspectives other than their own and will argue as sympathetically, and as powerfully as possible, from those other perspectives, when weighing the pros and cons in the case. Given this conception, how should fair- mindedness be tested? 32 Alec Fisher 2. The Design of the Paul's Subtest The Subtest contains the following Preamble. which explains the design of the test: There will be two parts to the test. First students will be given sets of opposite or contradictory beliefs and asked to identify which they consider more justified or cor- rect (the purpose is to determine which of the sets they personally favor.) Then the students will be presented with opportunities to identify (out of a set of five) the best reason for accepting the belief and afterward its opposite. It is hypothesized that uncritical thinkers will have difficulty identifying the best reasons for both the beliefs they hold and for the opposing beliefs; that closed- minded critical thinkers will be adept at identifying the best reasons for the beliefs they hold but not for their opposites; and that fairminded critical thinkers will identify the strongest reasons for the beliefs they don't hold as well as the beliefs they do. It is not assumed that students will judge the best answers from their personal point of view but rather, to the best of their abili- ty, from the point of view of one who holds the basic belief being supported and who must defend it before an audience of reasonable persons. So that each question is of the form "Imagine yourself to be a per- son who holds this belief to be true. Which supporting reason would be easiest to de- fend before an open-minded audience of reasonable people?" The "fairness" of mind that is being tested here is that achieved by individuals who have developed some ability to distinguish strongly relevant from weakly relevant and irrelevant reasons and con- siderations within the context of empathical- ly entering into the points of view of those who hold beliefs they do not themselves hold. As part of this capacity, a fairminded person must be able to tell the difference between a reason that will seem best only to a narrow circle of believers (who often share a variety of questionable if not biased assumptions) from a wide circle of rational persons who will be moved most by reasons that are least dependent on assumptions as questionable as the belief itself. I suggest that each person be given dou- ble points for identifying the strongest reasons for accepting those beliefs opposite to their own. By weighing more heavily the skill involving reciprocity, the "fair- mindedness" component is accentuated. Little needs to be added to this careful explanation. If one wishes to test someone' s fairmindedness, it seems entirely reasonable to first identify which of two opposing beliefs they hold, and then to see how well they argue for both sides of the question. However, this turns out to be harder than one might expect. We look first at some problems with Paul's opposing pairs of beliefs. 3. The Fairmindedness Subtest: Part I a) The Subtest has two parts. In the first part the candidate is presented with nine pairs of statements (the members of each pair being "opposites") and he or she has to choose which of each pair" seems ... to be the most reasonable or correct". Some of the pairs present reasonably straightforward choices, for example: 8. a) The crime of murder should be punished by death. b) The crime of murder should not be punished by death. However, other pairs do not present straightforward choices. Here are two examples: 5. a) Being patriotic is a good thing. b) Being patriotic is a dangerous thing. 6. a) Poor people are often poor workers. b) Poor people are often good workers. Though the members of some of the pairs are contradictories (as with 8), in these two examples they are not, and candidates could regard both members of each pair as equally "reasonable or correct". Even with examples whose members are contradictory, they often seem to force a choice where it would be quite reasonable to say "I don't know" or "It depends" , for example: 7. a) Rich people usually deserve the money they have. b) Rich people usually don't deserve the money they have. Some candidates might believe that neither of these is more reasonable ("Some rich people deserve their wealth, some don't; I don't know which is more common"). Other examples force a choice where a candidate might want to say "It depends on what is meant" (e.g. 2. "Brief political adver- tisements should [not] be allowed on T.V. "). b) There appears to be a general problem here which will arise with this part of any test of fairmindedness. The concept which is being tested requires candidates to choose between opposing points of view. However, with several of Paul's pairs, it is possible to imagine circumstances where a choice would misrepresent the candidate's position- and the proposed scoring method would not therefore correctly measure his or her fair- mindedness. This problem is not a (mere- ly) theoretical one, since if many of those for whom the test is intended have difficulty in choosing which of the opposing pair repre- sents their own view, the items cannot func- tion as intended, as a test of fairrnindedness. On the other hand, this observation almost certainly points the way to the solu- tion to the problem. If most candidates see the pairs between which they have to choose as representing "opposing" points of view, and readily choose one member of each pair as expressing their own point of view (so that the pair genuinely puts the popUlation into opposing camps) then the basic require- ment for testing fairmindedness appears to be met. This suggests that if one wishes to test fairmindedness, one must first identify those "opposing views" in the intended population where fairmindedness is gen- uinely at issue, and that this must be done empirically. Testing Fairmindedness 33 4. The Fairmindedness Subtest: Part II The Directions for part II of the test are summed up by Paul as follows: I) For each item imagine yourself to be a person who really believes in the position stated, 2) Imagine your needing to select one ofthe answers as the basis of your present- ing a defense of this position before an au- dience of reasonable people, 3) Select one of the answers, to the best of your ability, on these considerations alone. As we shall see shortly, there are some general difficulties about these instructions, but there are also specific difficulties about particular items. Here is an example: 9) Being patriotic is a good thing because: I) everyone ought to be ready to fight for his country 2) everyone ought to try to make their country as good as it can be 3) we live in the best country in the world 4) if someone doesn't love his country that person ought to leave and find a better one 5) everyone ought to support the policies of their government The credited response is 2), but (writing from a British perspective at least) it is not easy to see what 2 has to do with 9. In general, it is not easy to say what would be a strong reason for 9 (partly because it is a large and categorical claim in a realm in which surely most people would want to qualify what they say-and certainly critical thinkers would); however, one standard way to defend such a claim would be to say that it has good consequences (which outweigh the bad). In the absence of a response of that kind it is hard to know which response to choose. Consider another example: 17) What high government officials say can usually he trusted to he true because: I. they are basically honest people 2. they realize that the truth will prob- ably come out in the long run 34 Alec Fisher 3. they have taken oaths of allegiance to uphold the constitution and the laws of the country 4. they have good sources of information and usually have no reason to lie 5. they realize that a government offi- cial can be sent to jail for lying to the people The credited response is 2. The problem with this example is that the candidate has to know whether responses 2, 3, 4 and 5 are true. Prevailing political institutions and practices determine what a believer in 17 would regard as the strongest reason for it. 2 might not matter much (the high officials would have got away with it by the time the truth came out!); 3 might be a very impor- tant and effective constraint-or it might be ineffective; 5 might be true and important-or it might be false. 5. The General Problem of Validation The problems with these two examples illustrate what is in fact a general problem with Paul's items in this part of the test, namely that no evidence is provided that what is being tested is fairmindedness- rather than something else. From the fact that a reviewer agrees with nearly all the credited responses (as I found that I did) it does not follow that what is being tested is fairmindedness. More importantly, even if the test statistics show that the test is reliable it does not follow that what is being tested is fairmindedness. What is needed is evidence of validity, evidence that the test measures what it is intended to measure, namely fairmindedness. This would be equally true of any other attempt (besides Paul's) to devise a test of fairmindedness of course. What is required is that a justification or rationale be provided for each item to show that it tests fairmindedness rather than something else. The best method of valida- tion in a case like this would appear to be Norris's (1989). In short, we need a clear account (from Paul or some other expert) explaining why it requires fairmindedness to arrive at the credited response in each item; we must then interview candidates who have answered the test to ensure that it was fairmindedness which caused them to choose the credited response (rather than shared background beliefs or whatever) and a lack of fairmindedness which caused them to choose other responses. Only then can we be sure that what is being measured is fairmindedness properly conceived. Let us attempt to explain by means of an example what is required to validate items in this domain and why it is difficult. Suppose that Mary is Californian, of Norwegian stock, and an atheist; suppose that Lech is from Chicago, of Polish stock, and a Roman Catholic. Suppose the state- ment at issue is S, "Miracles happen." It is one thing for Mary to decide what would seem to her to be the strongest reason in favour of the statement S, in which she does not believe; it is another thing for Mary to know what Lech, who believes in S, will regard as the strongest reason in favour of S. But if she is asked to say what Lech will present as the strongest reason for S when arguing his case before an audience of "un- biased and open-minded" people, is she gen- uinely being asked something different from both of the others? On Paul's conception she certainly is. But in that case, to validate the items it is necessary to show that generally speaking candidates cannot and do not arrive at the credited response by answering either of the other questions. There is another, and related, problem. It is that the instructions for the second part of the test require the candidate to do something quite complicated (to put oneself in the shoes of someone who believes Sand then to decide what they would choose as the strongest reason for S if they were presenting it before an audience of open- minded and reasonable people). The problem is that it is easy for a candidate to forget such complicated instructions. The present writer did. In fact I realised after doing the test that I had been simply answer- ing the question, "What is the best reason for S?" Given that I agreed with nearly all Paul's credited responses, this may be evidence that the test is not measuring fair- mindedness, but simply the ability to judge what is the strongest reason for each state- ment (from something similar to Paul's perspective) . 6. In Conclusion In generaL a test of fairmindedness is important; it is important to devise some kind oftest if those who try to teach strong critical thinking are to have their efforts prop- erly evaluated. In the absence of a satisfactory test, such teachers are in danger of prom- ising to deliver what people want without good evidence that their programme has any real impact-the very antithesis of what (strong) critical thinkers want. The lessons Testing Fairmindedness 35 of this critique of Paul's test are that two conditions need to be met if one is to devise an acceptable test of fairmindedness: (i) the opposing points of view between which can- didates have to choose must be seen by them as opposing, and must be such that most candidates readily accept one of the alter- natives as articulating their own point of view; such pairs must be identified em- piricaJly, and (ii) it is necessary to validate the items in the main part of the test, by showing that candidates choose the credited response if and only if they are fairmind- ed; this is not an easy task, but Norris (1989) shows how it can be done. Finally, it should be said that these criticisms of Paul's Fainnindedness Subtest were presented at his conference on "Critical Thinking and Educational Reform" in Sonoma in August 1989. Paul took part in the presentation and gave an impressive display of the very fairminded- ness he describes in our opening quotation. Hence this paper! References Norris, S.P. (1988). Controlling for Background Beliefs When Developing Multiple-Choice Critical Thinking Tests. Educational Measure- ment 7(3): 5-11. Paul, R., Glaser, E. & Bowen, B. Critical Thinking Appraisal: Fairmindedness Subtest. (Unpublished, but available from Richard Paul, Center for Critical Thinking Ilnd Moral Critique, Sonoma State University, Rohnert Park, CA 94928.) Paul, R., Binker, A.J.A. & Charbonneau, M. (1986). Critical Thinking Handbook K -3. Center for Critical Thinking and Moral Critique, Sonoma State University. ALEC FISHER DEPARTMENT OF PHILOSOPHY SCHOOL OF ECONOMIC AND SOCIAL STUDIES UNIVERSITY OF EAST ANGLIA NORWICH NR4 TTJ ENGLAND Editors' Note In connection with the topic discussed in Dr. Fisher's article. we thought readers might be interested in two insufficiently-known works which treat a concept closely related to fairmindedness. Hare, William. (1985). In Defence of Open- mindedness. Kingston and Montreal: McGiIl- Queen's University Press. (1979). Open-mindedness and Education. Kingston and Montreal: McGilI- Queen's University Press.