CASE: THE PSYCHOLOGY OF STUPIDITY

 

Case Author: Developed by Tanya Sharon, Summers 2003, 2004

 

Case Materials: Thomas Gilovich, (1991), How We Know What Isn’t So, New York, The Free Press.  (A moderately priced paperback available through some bookstores and Amazon.)

Various handouts are provided in the case notes.

 

Case Purpose: This is an extended case exploring, as the book title indicates, how people form and hold on to erroneous beliefs.  The book is somewhat similar to Michael Shermer’s Why People Believe Weird Things, but written by a psychologist and based on psychological research. It is used here as a tool to develop students’ critical thinking skills and to expose students to the methods psychologists use to answer questions about human thought and behavior.  The book is divided into 3 main parts with 3 chapters each. The first two parts explore cognitive and social determinants of questionable beliefs, respectively, and the third provides in-depth coverage of 3 examples of questionable beliefs- alternative health practices, ESP, and ‘questionable interpersonal strategies’ (this chapter is much weaker than the other chapters and one I skip).

 

Learning Objectives

1)      introduce some extremely common errors in thinking

2)      increase students’ awareness of and ability to recognize fallacies in their own and others’ thinking

3)      reinforce key strengths of science, including appropriate skepticism, clear operational definitions, a priori predictions, and looking for disproof (avoiding confirmation bias)

 

Relation to Course Objectives

This case gives students experience in evaluating decisions from a critical thinking perspective (objective #15), as well as in-depth exposure to how the scientific process can be used to expose irrational conclusions and flaws in thinking (objective #14).

 

Case Use

I teach this with 75 minute classes twice a week.  For 50 minute classes the reading assignments will need revision. A suggested schedule with 75 minute periods (note that if you used all the activities below you would most certainly exceed this schedule by quite a bit):

Class 1 –Part 1 (Cognitive Determinants), Ch 2- Random Data, & Ch 3- Incomplete Data

            Class 2- Part 1 (Cognitive Determinants), Ch 4- Ambiguous Data

            Class 3- Part 2 (Social Determinants), Ch 5-Motivations, & Ch 7- Imagined Agreement of Others

            Class 4- Part 2 (Social Determinants), Ch 6- 2^nd-hand Info

            Class 5- Part 3 (Examples), Ch 8- Alternative Health Practices

            Class 6- Part 3 (Examples), Ch 10 - ESP

 

The chapters do reinforce each other, but are not mutually dependent. You could skip some chapters or use them selectively without introducing confusion, although they do make a particularly effective package taken together.

 

Reference List: links to articles relevant to specific demonstrations are provided with the demonstrations, but no additional readings aside from the text are needed

An outline of the entire text is here.

 

Ch 1: Introduction

 

The introduction is only 6 pages long but I don’t assign it, lest the students feel horribly misused at being assigned 3 whole chapters at one time! It does make one critical point, however, which I simply present in the first class: that (the case title notwithstanding), Gilovich does NOT hold that people in general are stupid.  Rather, his central contention is that people’s faith in highly questionable beliefs “derives primarily from the misapplication or overutilization of generally valid and effective strategies for knowing” (p. 2).  Other authors have referred to these as ‘mind bugs’, analogous to computer programming errors that cause problems only in certain circumstances. This may be a hard concept for students to grasp.  One way to convey it is via demonstration of a few perceptual illusions. One example Gilovich uses in Ch 2 is St. Louis’ Gateway Arch; another simple example would be the Muller Lyer illusion. The point to emphasize with these perceptual illusions is that the illusion is so strong that they are not eliminated simply by knowing the correct answer.  Similarly, our cognitive illusions can be very compelling; in contrast, however, the entire point of Gilovich’s book is that with knowledge of these cognitive illusions we can override their impact and improve our thinking and decision-making.

 

 

Cognitive Determinants, Ch 2- Something Out of Nothing:

The Misperception of Random Data

 

Reading summary: Gilovich starts with the supposition that people are predisposed to dislike randomness and to like and seek order and patterns. This leads to phenomena such as the ‘clustering illusion’, the belief in a ‘hot hand’ in basketball, and the regression fallacy (all defined below).  The representativness heuristic is a major contributor to these errors. Having misinterpreted random data, we then cement our interpretations with ad-hoc explanations (illustrated by research on split brain patients).

 

I start with the big picture: whether people are rational or not. There is a long tradition in psychology, and before that in philosophy, of holding humankind to a rational ideal.  This viewpoint and its antithesis are both represented in the research Gilovich discusses, so it is worth bringing it out at the beginning.

    Some possible initial Discussion Questions:

Overall, are people rational or irrational?

What are some examples you can think of, of clearly irrational beliefs? (Gilovich lists some p 5-6 that are far enough out that students will have no difficulty recognizing them as such, eg killing seals to use their penises as aphrodisiacs. You can also press them to identify some irrational beliefs a little closer to home.)

Why do people hold irrational beliefs?

Does it matter if people hold irrational beliefs? Why or why not?

            (finally, to bring in Gilovich’s point in the introduction)- Does Gilovich think people are irrational and stupid?  (see above for discussion and demo)

 

At this point you are ready to get into the substance of the chapter. Here are some possible demonstrations of people’s difficulty recognizing chance/randomness:

• Coin-Flipping Exercise (5-10 mins). Tell students you will leave the room for 5 minutes. While you are gone, they are each to simulate as best they can the outcome of a series of coin flips (ie, write down a series of H’s and T’s on a piece of paper). They should also elect one person to flip a real coin 20 times and record the outcome.  When you come back, you will pick out the 1 person who really flipped the coin- it will be the person with the longest run of H’s or T’s.  This might be most effective if you do it at the end of the class preceding reading this chapter. Alternatively, you can make it a homework assignment, telling each student to first make up a 20-flip sequence, then actually do it, and compare the two. (Thanks to Thomas Gilovich for the idea.)

 

• Find Your Twin exercise (~20 mins).  Preface this exercise with a discussion of research on twins separated at birth and later reunited. For example, many of them will have heard of Jim Springer and Jim Lewis, who when reunited in their late 30s discovered an amazing number of similarities- both had married and divorced women named Linda, married second wives named Betty, drove the same model blue Chevrolet, and owned dogs named Toy (see here). Then give students a list of 20 questions to answer about themselves, e.g. their favorite kind of food, etc. Tell them to pair up with one other person and see how many questions they gave matching answers for.  Poll the entire class to find out the range of matches.  You should find at least one pair with a high number of matches (one of my classes had a 12-match pair, but in another the maximum number of matches was 2- including more characteristics in the list reduces the chances of this).  Use the exercise to launch a discussion of how, while a ‘high’ number of matches may seem remarkable and non-accidental, any 2 things can be similar if enough aspects are compared. This exercise can also be used to prompt discussion of the principle of parsimony, and/or demonstrate the problem of multiple endpoints in Ch 4. (The Birthday Problem discussed on p. 175-7 is somewhat similar in concept to this exercise.)

 

Next, having demonstrated the phenomenon, press students for an explanation for why people would have difficulty recognizing randomness. Lead into a discussion of Gilovich’s answer: in short, it is built into how our brains work. In more detail, you can discuss all or some of the following-

 

The clustering illusion (p. 16-21) is a term used by statisticians to capture people’s erroneous intuitions about how random events should look- the intuition that random events shouldn’t have any clusters at all but should be perfectly evenly distributed. Again, what bears emphasizing is that this illusion is so strong that it is not eliminated simply by knowing the correct answer.  In this way it is very similar to perceptual illusions, which retain their effect even when the illusion is revealed.  If you haven’t already covered this point (see Ch_1), you should here. The example Gilovich uses is St. Louis’ Gateway Arch; another simple example would be the Muller Lyer illusion. The ‘hot hand’ phenomenon in basketball is an example of the clustering illusion, and refers to the erroneous the belief that success/failure in making basketball free throws breeds more success/failure. Students easily generalize the hot-hand phenomenon to related phenomena but have more difficulty recognizing the broader illusion.

 

Gilovich’s explanation for why chance looks too ordered to us is partly based on the famous representativeness heuristic proposed by Kahneman and Tversky (see here). An enormous body of research exists demonstrating the impact of this rule of thought on people’s judgments and decisions. Basically, it refers to people’s reflexive tendency to assess things on the basis of salient or superficial features. For example, we expect effects to look like their causes (heartburn from spicy food not bland, etc), and categories to be like their members. In the coin-flipping example, we expect even short sequences to show the key feature of a 50-50 split between heads and tails. Another potent example is Steve the librarian (see below).

 

A demonstration of the representativeness heuristic:

·        Steve the librarian- Read students the following description: “Steve is very shy and withdrawn, invariably helpful, but with little interest in people or in the world of reality.  He has a need for order and a passion for detail.  Is Steve more likely to be a salesperson or a librarian?” Students of course say librarian, since the description matches our stereotype (is far more representative of) a librarian more than a salesperson. But judging that Steve is a librarian ignores the fact that salespeople are massively more common than librarians (by a 75:1 ratio, according to some estimates).  According to the classic psychological literature on judgment and decision-making (called the ‘heuristics and biases’ school), ignoring the base rate is a fundamental error in logic (see also Ch 6), and the correct answer is that Steve is more likely to be a salesperson.  In essence, there are going to be more salespeople who happen to be shy than librarians of any type. (If 95% of librarians are shy, but only 5% of salespeople are, you would still have more shy salespeople [7500*.05 = 375] than shy librarians [100*.95 = 95]). Explain this to students and get their reaction. Many people’s intuition shrieks ‘that makes no sense!’  In fact, there is another school of thought (the Brunswikian tradition) that explicitly rejects the premises and conclusions of the heuristics and biases school. Rather than measuring people against a pure ideal of rationality and finding them sorely lacking, the Brunswikian tradition stresses that people are designed for, and excel at, making sense of cues in a real environment. According to this second perspective, when you read the description of Steve, you were essentially providing cues to students, and it would be foolish for students not to make use of this information. Further discussion could address how science involves and adjudicates between different theoretical viewpoints, and/or revisit the question of what exactly it means to be rational or irrational. The important point is that, regardless of which theoretical camp is correct, the strength of the basic effect (students’ rejection of the answer ‘salesperson’) shows how strongly we are affected by representativeness. You can point out that this format is paradigmatic of classic research in cognitive psychology: presenting subjects with hypothetical scenarios and comparing their responses to a rational ideal.

 

Judgment by representativeness also contributes to the regression fallacy - people’s failure to recognize regression to the mean. Examples in the text include the Sports Illustrated jinx (p. 26); it also arises in Ch 8 regarding dubious health care practices. It is relatively straightforward to present. I draw a simple graphic (a straight line through a wavy line) and map a few examples to the graphic.  Do explicitly point out that where the line (the mean) lies can vary from person to person, eg Tiger Woods and John Doe both have good and bad days but Tiger’s baseline is in a very different place. You should also point out that regression to the mean only applies when there are enough datapoints to have a meaningful mean. Thus, it is more accurately applied to groups or populations than to individuals. 

            Gilovich includes a very interesting discussion (p. 27) of how regression to the mean can mask the effectiveness of praise and make punishment seem to be a more effective tactic than is the case. If you wanted to spend more time on the regression fallacy, this should be a good hook- students usually get quite engaged when discussion turns to views of discipline.

 

 

Cognitive Determinants, Ch 3- Too Much from Too Little:

The Misinterpretation of Incomplete Data

 

Reading summary: There are 2 main points to this chapter. The first concerns the excessive impact of confirmatory information on our beliefs. People commonly misinterpret incomplete data because we fail to appreciate the distinction between necessary and sufficient evidence.  We look for and give heavy weight to instances consistent with a belief, and tend to ignore other information, including instances inconsistent with that belief- in short, we show a confirmation bias. The second point is that we often lack the information that would allow us to disprove a belief, leading to an illusion of validity.  Self-fulfilling prophecies are a special case of hidden/absent data, in which beliefs lead to behaviors that change the world we observe, thus making it impossible to collect observations contrary to the initial belief.  But Gilovich again argues that most of the time our initial beliefs are not entirely irrational but have some basis in the real world- thus his title ‘too much from too little’ rather than from nothing.

 

Confirmation bias - refers to the tendency to focus largely or exclusively on information that can confirm a belief, rather than looking for disproof.  This is very helpfully conveyed by a 2 by 2 table (p 31). One example in the text is the erroneous belief that infertile couples who adopt are more likely to conceive naturally than infertile couples who don’t adopt. The former situation is much more memorable than the latter, thus leading people to overestimate its frequency. Students can get practice with avoiding confirmation bias from the following worksheet, which can be assigned as homework or done in class in groups.

 

Here are some possible demonstrations of confirmation bias:

• Deduce the Rule (~5 mins)- Give students the number sequence 2, 4, 6.  Their job is to determine the rule that created that sequence by creating sequences of their own; you will tell them if their sequence follows or violates the rule.  They can continue creating test sequences until they are confident they have figured out the rule. Most people only generate sequences that can confirm but not disconfirm their idea. In this case, the obvious first guess is the rule ‘increase by 2’, but the real rule was ‘any increase’. Someone proposing sequences according to the ‘increase by 2’ rule will erroneously believe they have the correct rule and thus never discover the real one.

 

• Design a Personality Test(~15-20 mins)- Have students form groups or pairs and tell them that their task is to design a 3-5 item questionnaire that can determine whether an individual is (pick one) agreeable, responsible, or open to new experiences (the example in the book is extraversion, p 35).  When students are finished, have them present their questionnaires and as a class analyze whether the proposed questionnaire would in fact be useful in identifying the specified personality trait or whether it would be likely to yield false positives (ie, shows a confirmation bias), and thus fail to discriminate accurately between people with and without that trait. Note that people do not always fall prey to the confirmation bias in this type of task.  If you want a guaranteed demonstration of confirmation bias, consider the next task.

 

• The Wason Card Task (pp 33-34) is a classic problem in psychological research on human reasoning (~ 10 mins for basic version, perhaps 45 for the longer?).  This task has been used in dozens of studies, virtually all of which show that people are very poor at logical reasoning. To demonstrate this with your students, present them with 4 cards in the layout below (I’ve changed the specifics from that in the text to make it more interesting in case any students actually read and remembered the example in the text).

 

E

R

4

9

 

Explain that each card has a vowel on one side and a number on the other.  Their task is to determine whether ‘all cards with a vowel on one side have an even # on the other side.’ Specifically, which card or cards do they need to turn over to ascertain whether or not the rule is true? Ask students to work out their own individual answers then discuss as a group. The most common response is E or E and 4, thus demonstrating the confirmation bias: The 4 card can only confirm the rule, not show whether it was broken. More generally, for any rule of the form ‘if p then q’, proving the rule requires turning over the card representing p and the card representing not-q; in this example, E and 9. Write the rule on the board and label each card with the appropriate p’s and q’s (now you know where that phrase came from!). Go over it until students have it. The point again is that most people are biased to look for confirmation rather than disconfirmation. In many studies, only 10-15% of American college students answer correctly.  Interestingly, research has shown that asking the question in the form of a search for disproof/violations leads to only a minor improvement in performance at best. You can take this opportunity to discuss how disconfirmatory evidence can be more decisive than confirmatory evidence.  Given the rule, ‘all geese are white’, you can have multitudinous instances of white geese, but a single example of a black/purple/whatever goose is all it takes to disprove the rule.

In a footnote, Gilovich discusses research that shows some exceptions to the confirmation bias.  Some researchers (e.g., Cheng and Holyoak, 1985) have shown that if the scenario is presented in a context involving permission, people overcome their confirmation bias.  For example, given a context involving checking id’s at a bar, most people effortlessly choose the logically correct cards. Present students with a new row of cards:

 

Beer

Coke

12 yrs

45 yrs

  

They all know the rule that applies here, ‘if drinking alcohol, then 21 or older.’ People easily choose alcohol (p) and 12 yrs (not-q) (85% correct and up).

 

This second demonstration can be used in two ways.  First, you can use it to illustrate the idea that our evolved mental heuristics have evolved to be very effective in the real world—it is when the rules are abstract and meaningless (ie, the letters and numbers) that we appear most irrational. For a more elaborate activity, you could give students both demonstrations and ask them to develop, in teams, a model to explain the different patterns of result. (I will be trying this in Fall ’04 and will update the case accordingly.)

 

           

The illusion of validity refers to a belief that enjoys a mistaken impression of support because the evidence to disconfirm it is missing or unavailable. One salient example in the text is the belief in the power of SAT scores for predicting college success. 

 

A simple example of a self-fulfilling prophecy is when a rumor of bank insolvency leads to a run on the bank and therefore causes the bank to become insolvent. A more elaborate example is the classic Prisoner’s Dilemma game (p. 45-46), which is used to show how a competitive/negative mindset can be self-fulfilling- people who interact competitively with others force them into being competitive in self-defense. There are many sites for playing this on the web; one is here.   You could assign students to play the game as homework then report or write up what strategy they used and how it worked. This assignment would also work best if students completed it before the reading assignment.

 

 

Cognitive Determinants, Ch 4- Seeing What We Expect to See:

Biased Evaluation of Ambiguous Data

 

Reading summary: Gilovich begins again with the thesis that people’s evaluation of information is sometimes biased, but now raises an important caveat, the issue of appropriate versus inappropriate bias. Some ‘bias’ is appropriate- if a belief has a lifetime of support, it is perfectly reasonable to be highly skeptical of contradictory info. (If using Carey, this can be connected to Carey’s tenet that ‘extraordinary claims require extraordinary evidence’.) Gilovoch also phrases this as a distinction between skepticism and close-mindedness; the case of cold fusion is briefly mentioned and could be developed further (see here or here).  Also raised is the issue of bias in the evaluation of scientific findings and (too) brief mention of some of the safeguards against bias (replicability, statistical analysis, control groups, blind observers, and clear a priori predictions). This is contrasted to the problem of multiple endpoints in e.g., horoscopes, where vaguely worded predictions can be ‘confirmed’ by just about anything. People’s tendency to accept vaguely worded descriptions as accurate of themselves is called the Barnum effect, after the circus promoter P.T. Barnum who excelled at providing something for everyone.

Note: I assign only through p 61; the remainder of the chapter gets fairly technical.

 

A possible demonstration of bias in evaluating data:

• A very simple example from perception is given in the text p. 52. The ‘bias’ arises from our pre-existing knowledge, which leads us to interpret the ambiguous stimuli in a certain way, based on years of experience with reading.  We would not show the same effect for entirely unfamiliar stimuli such as Chinese characters, unless of course we could read Chinese.

 

A possible demonstration of the problem of multiple endpoints via the Barnum effect:

• Horoscope Matching Copy the weekly horoscopes from some source students are unlikely to frequent, making sure that the month information is deleted.  Give each student all 12 horoscopes and ask them to identify (by #) which is theirs.  Record their selections, then give the ‘correct’ answer.  Calculate the number of students who chose correctly.  It should be less than 10% (chance = 1/12= 8.3%). If they complain that the horoscopes are too vaguely worded to choose one, all to the better!

 

Group Activity/Quiz (~40mins total):

Give groups 10-15 minutes to define each of the key concepts in the chapters to date (open-book or closed) then generate 1 good example of their own of that concept. I ask for the following: representativeness heuristic, regression fallacy, confirmation bias, illusion of validity, self-fulfilling prophecy, problem of multiple endpoints. When they are done, check their definitions and discuss examples as a class- this can be a very efficient way of identifying misconceptions, as well as getting great examples of the concepts to use in the future.

 

Homework:

Give students the handout on Appropriate Skepticism and Inappropriate Bias. It asks them to rank a series of statements in terms of the amount of evidence or support they would require to accept the proposition as true (similar to Carey’s exercise #16 in Ch 2). More importantly, they are to describe the evidence they would require to accept the proposition.  The point of the exercise is threefold: 1) to become more aware of when bias/closemindedness is present, 2) to become more aware of when bias/closemindedness is appropriate, and 3) to improve one’s ability to identify support for a belief (whether logical/rational, empirical, or faith-based). You will need to go over Ways of Knowing (LINK see the case for a useful handout) if that hasn’t yet been covered. They are rather awful at identifying relevant evidence- there’s a tendency to instead offer ‘arguments’ like ‘that’s just absurd’, so you will definitely need to analyze some example statements for them.

 

 

Social Determinants, Ch 5- Seeing What We Want to See:

Motivational Determinants of Belief

 

Reading summary: This chapter explores people’s tendency, “within limits”, to hold beliefs that reflect favorably upon ourselves (self-serving beliefs). In particular, people’s assessments of their own abilities are notoriously inaccurate. This is captured in Garrison Keiler’s famous Lake Wobegone effect, in which “all the men are handsome, all the women are strong, and all the children are above average”. A subtype of self-serving bias is the fundamental attribution error (p. 78), in which people tend to attribute their successes to internal factors (ability) and failures to external factors (unfair teachers, etc.). Three ways in which motivations influence beliefs is via 1) the framing of questions asked of evidence, 2) selective solicitation of opinions, and 3) stopping data collection once one’s opinion has been verified.

 

 A possible demonstration of motivations influencing beliefs:

• An elaborate demonstration could be set up with a Clue-like game.  For a simpler version illustrating mechanisms #1 and 3, one could use the Deduce the Rule exercise from Ch 2. The distorting motivation in this case is the desire to be right (rather than the desire to find out if one is right). This would also give the opportunity to review and underscore confirmation bias.

 

A possible demonstration of self-serving beliefs:

• Student Belief Survey- Before you assign this chapter, have students fill out and return a brief, anonymous questionnaire. Before class, calculate the proportion of students rating themselves as ‘above average’ on the various traits.  Based on previous research, you can expect to find that the majority of students rate themselves as above average (often significantly above average) on most traits.  You may also find somewhat less bias on traits that are relatively more objective, for example athletic ability compared to the more amorphous ‘character’ or ‘ability to get along with others’. If you get this pattern, you can use it to underscore Gilovich’s point that our self-serving beliefs are at least partly constrained by reality. You can point out again how simple pen-and-paper tests are used by psychologists to gain a wealth of knowledge about human psychology.

 

A possible demonstration of the fundamental attribution error:

• Trivia Game- Watson (1987) developed a classroom demonstration. Have students pair up with someone they do not know well.  Randomly assign half to be ‘quiz masters’ and half to be ‘quiz takers’. Have the quiz masters create 5 reasonably difficult general-knowledge questions (e.g., the capital of a distant state, the name of the current Secretary of State, etc.)  Then have each quiz taker answer the questions, and immediately be scored as correct or incorrect. Finally, have all students anonymously answer one question about how intelligent they perceived their partner to be (see handout). Collect their papers and make a histogram of the results (more practice with graphs!) by the role played. You should find an obvious difference, with quiz-makers rating their partners as less knowledgeable than vice versa. Students thus succumbed to the FAE, ignoring the fact that the exercise was set up to favor making quiz-makers look smarter than quiz-takers.  If the roles were reversed, the original quiz-takers could make their partners look ignorant. (For more details, see Watson, D.L., The fundamental attribution error. In Makosky, Whittemore, & Rogers, eds., Activities handbook for the teaching of psychology: Vol 2, pp. 135-137.)

 

 

Social Determinants, Ch 7- The Imagined Agreement of Others

 

Reading summary: We begin with Gilovich’s usual formula: that our beliefs are heavily influenced by a social factor, in this case our perceptions of others’ beliefs, that ‘within limits’ this is perfectly reasonable, but the limits are exceeded. Specifically, the problem here is that there is a ‘systematic defect’ in our ability to estimate others’ beliefs: we consistently overestimate agreement with our own beliefs, aka the phenomenon of false consensus. A number of factors contribute, including the fact that the vast majority of people are exposed to people and information that supports our pre-existing beliefs, unrecognized differences in how questions and situations are construed (e.g., American cuisine = MacDonald’s leads to a different evaluation than American cuisine = fusion food), and a lack of corrective feedback from others. The evidence for this last point is the weakest in the book (e.g., the existence and popularity of etiquette manuals, of all things), but students are able to see the point themselves. 

 

A demonstration of false consensus (~10-15 mins):

• Hand out the False consensus questionnaire at the end of the class prior to this one.  Have students either complete it in class or as homework (but before they read the chapter). It asks them to first estimate the proportion of Mercer students who hold the specific beliefs or opinions; these %s should be recorded in column 1.  In this class, you will fill in columns 2 and 3. Ask students to record in column 2 their own view of each belief or opinion (True or False).   Finally, fill in column 3 (a calculator is handy)- determine the proportion of students in class that day who hold the belief via a simple show of hands. Point out that the present students constitute a sample of all Mercer students. You could also discuss the issue of representative samples, although it’s a bit tangential to this demonstration (the %s are important for the bias they reveal in individuals, not for their actual values). Finally, ask students to compare columns 1 and 3.  For each statement, if column 2 is marked ‘true’ (i.e., if the student endorsed that belief) then by the false consensus effect you would expect the % in column 1 (the individual students’ estimate of general student support for that belief) to be greater than the % in column 3 (the actual % in your sample) (ie,  3 < 1).  Similarly, for each statement marked ‘false’ by the student you would expect 3 > 1. Ask students to circle each statement for which they showed the false consensus effect and count the total. Get a show of hands for the number of students who showed the effect on at least 1 statement, 2 statements, etc. You can then move into discussion. Do emphasize that this is a normal, expected effect and is not a judgment on anyone’s moral worth.

 

Group Activity/Quiz (~20mins total?):

Give groups 10 minutes to define the main concepts from Part 2 to date (open-book or closed) then generate 1 good example of their own of that concept. I would ask for the following: fundamental attribution error, self-serving belief, false consensus. When they are done, check their definitions and discuss examples as a class- this can be a very efficient way of identifying misconceptions, as well as getting great examples of the concepts to use in the future.

 

 

Social Determinants, Ch 6- Believing What We Are Told:

The Biasing Effects of Secondhand Information

 

Reading summary: This chapter gives a compelling overview of how information collects inaccuracies, distortions, and biases through the processes of transmittal (whether verbal or through mass media). Gilovich introduces a number of principles, derived from communication theory, that explain how and how and why information is passed on: justification (entertainment vs informativeness), immediacy, sharpening and leveling, ironic plausibility (this one causes difficulty- many students don’t seem to understand irony), and distortions from self-interest.  In discussing possible ways to deal with these distortions, he points out the difficulty inherent in one of the standard solutions offered by experts (give more weight to the base rate of events in a population), given that accurate information regarding the base rate is often very hard to gather from the mass media.  The issue of base rate neglect could be developed much further; a few additional resources can be found here.

 

A very effective demonstration of distortions in second-hand information:

• Telephone (~10 mins) - Announce to students that you will play one round of telephone.  Their task is to relay information to the next person as accurately as possible. Take one person out of the room and read him/her a passage approximately 5-8 sentences long. The actual number of sentences is less important than the number of ideas (distinct propositions)- that needs to be at least 7, but much more than that and students will be overwhelmed. It should not be something with which students are likely to be familiar- they could then just fill in from their pre-existing knowledge. I read the beginning of a traditional African tale, which reliably leads to incredible errors.  Each student then whispers it to the next student, all the way to the last person, who reports the final version aloud.  Then read the original version and point out and discuss the distortions. It can also be fun to have students at various points in the chain report the version they heard, so the class as a whole can see how the distortions accumulate.

 

Class activity (~45-60 min) and writing assignment:

One or 2 classes prior to this class, hand out the Urban Legend assignment. This assignment asks students to research an urban legend of their choice and become familiar enough with it that they can report it to the class. Start with the retelling of the legends; if you have the time and/or a small class you can have every student tell their story.  Write short names down on the board for each legend. When complete, ask students to vote for the most memorable legend- the one they will be most likely to go tell a friend.  Take the winner and discuss what makes it an urban legend- what makes it worthy of passing on.  Turn next to the principles outlined by Gilovich (see above) and ask students to generate the list and define/explain each (a quick team quiz is possible). Finally, apply the principles to the winning urban legend- essentially, model for them what they need to do for the written assignment. Do this for at least 4-5 legends, and do point out that not all the principles apply to every legend. The written assignment can be due 1-2 classes after this class. 

 

You may want to briefly review and discuss Gilovich’s suggestions for dealing with the fact that so much information is second-hand and therefore distorted (p. 109-110).  Alternatively, you could have groups brainstorm their own suggestions.

 

 

 

Examples, Ch 8- Belief in Ineffective Alternative Health Practices

 

Reading summary: This chapter is devoted to analyzing people’s belief in dubious alternative health care practices (in which he includes homeopathy), beliefs that carry an enormous cost in both money and lives lost. Factors he identifies as contributing to such beliefs include

1)      the will to believe

2)      inferring causality from sequentiality (post hoc ergo prompter hoc) and failing to recognize regression to mean

3)      murky outcomes (similar to multiple endpoints in Ch 4) and ambiguous criteria

4)      an aura of plausibility (e.g., handwriting analysis, projective tests in psychology- background information on the limitations of the latter can be found at ADD LINK) due to the representativeness heuristic

 

Group quiz/activity (~20-25 mins total):

Give groups ~10 minutes to brainstorm two lists: one of reasons that Gilovich cites for why people believe in ineffective health practices and the second their own list. Collect the lists then discuss as a class (one factor glaringly absent from Gilovich’s list is the lack of access to mainstream health care). Students seemed very engaged in this.

 

Class activity (~30-45 mins) and writing assignment:

Prior to this class, give students the assignment on Alternative Health Care Practices. They are asked to research an alternative health care practice of interest to them and be prepared to present a critical evaluation of the support they found for that practice.  Their goal is to be able to give informed advice to a friend who is considering using that practice. Students are sometimes unsure of what constitutes ‘alternative’ health practice. A working, somewhat conservative definition is anything that you would not receive in a doctor’s office. In class, the number of student presentations can be tailored to available time. With each presentation, press students to address and critically evaluate the kinds of evidence provided (Testimonials? Retrospective studies? Randomized studies? If a study is reported, how credible do they find it? Are there any obvious weaknesses?).  In short, model to them the kinds of question they should be prepared to answer in their writing assignment. This assignment requires a basic familiarity with the different types of causal studies; it would also be helpful to have already discussed the different ways of knowing (see LINK TO CASE).  Alternatively, this activity could be done as a group project with science-fair style posters as the product.

 

Examples, Ch 10 – Belief in ESP

 

Reading summary: This chapter discusses American’s high levels of belief in paranormal phenomena. He starts his argument by pointing out that a large number of studies on a phenomenon reveals nothing if most were of poor quality (e.g., studies of the portacaval shunt in the 1940s). Similarly, 130 years of research on ESP have not been able to provide convincing evidence. Why then the belief?

1)      the over-abundance of pseudo-evidence

2)      the sensationalistic bias in media coverage

3)      the will to believe

4)      personal experience that fails to recognize confirmation bias, the problem of multiple endpoints, etc.

 

Possible discussion questions:

-         Why do people believe in ESP? What common errors in thinking could contribute? What other factors?

-         What do you think of Gilovich’s treatment of ESP? Is it fair? Is it biased? [students have reported they find his tone in this chapter patronizing]

-         Are there any paranormal powers that you feel could be real? What is the basis for your belief?

-         Have you ever had an experience that did not seem explainable by conventional science? (or, to make less threatening, ask if they know someone else who ever has had such an experience).  Are other, non-paranormal explanations possible as well?

-         Can you design a test of the phenomenon that could decisively prove that the phenomenon is real? (This would also make a good group activity.)

 

Group activity: (~45 mins)

Have groups brainstorm their 5 top candidates for ‘signs of pseudoscience’ – how a layperson can distinguish between science and the pseudo-scientific.  Discuss as a class, writing down any viable ones (I incorporate the points from this exercise into the final exam.) If you used the What is Science? Majors Exercise (LINK) this provides a great opportunity to go back and review the earlier class discussion and decisions.