University of California, Berkeley
Department of Psychology

Psychology 1
Fall 2007

Midterm Examination 2

Final Scoring Key

In the scoring key that follows, correct answers are marked with an asterisk (*).

The statistical analysis identified only a single bad item: #10. This item was rescored correct for all responses. Students who got any this item wrong, in terms of the preliminary scoring key, should give themselves one (1) additional point (do not give yourself the additional point if you got a bad item "right").

The average score on the rescored exam was 38.84, standard deviation = 6.31), or approximately 78% correct. The median score was 39. This was excellent performance by the historical standards of Psych 1, in which my usual mean is somewhere between 65-70% correct.

In what follows, I provide the percentage of the class that got each item correct, as well as the item-to-total correlation for each item, followed by commentary on why the right answer is right, and the others wrong.

Choose the best answer to each of the following 50 questions. Questions are drawn from the text and lectures in roughly equal proportions, with the understanding that there is considerable overlap between the two sources. Usually, only one question is drawn from each major section of each chapter of the required readings; again, sometimes this question also draws on material discussed in class. Read the entire exam through before answering any questions: sometimes one question will help you answer another one.

Most questions can be correctly answered in one of two ways: (1) by fact-retrieval, meaning that you remember the answer from your reading of the text or listening to the lecture; or (2) inference, meaning that you can infer the answer from some general principle discussed in the text or lecture. If you cannot determine the correct answer by either of these methods, try to eliminate at least one option as clearly wrong: this maximizes the likelihood that you will get the correct answer by chance. Also, go with your intuitions: if you have actually done the assigned readings and attended the lectures, your "informed guesses" will likely be right more often than they are wrong.

Be sure you are using a red Scantron sheet.

Fill in the appropriate circles with a #2 pencil only.

Be sure you put your name on the front of the red Scantron sheet.

Be sure you put your Student ID# on both sides of the red Scantron sheet.

Indicate Exam 002 (use all three digits) on the reverse side of the red Scantron sheet.

Retain this exam, along with a record of your answers.

1. Working memory differs from long-term memory in terms of the

a. length of time information can be stored before it is lost or forgotten.

b. degree to which information has been processed before being stored.

c. capacity of the memory store.

d. all of the above.*

92% of the class got this item correct; item-to-total rpb = .25. For our purposes, working memory is essentially the same as short-term memory. Short-term memory is, almost by definition, time-limited, and its capacity is limited to "the magical number seven, plus or minus two". But working memory is also the "workbench" where memories are processed. Therefore, items in working memory have not been processed as much as their counterparts in long-term memory. At least, not yet.

2. A researcher presents participants with a list of words. She asks the participants to count the letters in the words in group 1, to come up with rhymes for the words in group 2, and to produce synonyms for the words in group 3. Later, she tests the participants' memory for all of the words. Going from best to worst, which pattern correctly indicates how well words in each group will be remembered?

a. 1, 2, 3

b. 3, 2, 1*

c. 3, 1, 2

d. There is no basis for predicting differences among the groups.

87% correct, rpb = .30. This is a question about depth of processing, or the elaboration principle. Group 1 is processing only the visual attributes of the stimuli -- a very shallow task. Group 2 is both reading and pronouncing the stimuli, which will produce somewhat deeper processing. But Group 3 is processing the meaning of the stimuli, as well as both the physical (visual and auditory) aspects of the items. This is very elaborate processing, and should produce the best memory.

3. Paula received a severe blow to the head in a car accident. What is the most reasonable explanation for her amnesia of what led up to the accident?

a. The information was not yet consolidated in long-term memory.*

b. Working memory interfered with the long-term memory storage.

c. Semantic memory has been disrupted.

d. She has lost the use of her hippocampus; that is, she is like H. M., but not as severely impaired.

70%, .43. Head injuries generally produce a retrograde amnesia, which affects episodic but not semantic memory; and retrograde amnesias are generally attributed to consolidation failure. What mediates the amnesia is a concussive blow to the head, and thus a loss of consciousness, rather than any damage to the hippocampus.

4. A student is presented with a list of word pairs, including the pair "stone�chair." According to the encoding-specificity principle, which word would be the best cue for the recall of "chair" on a subsequent task?

a. table

b. furniture

c. charity

d. stone*

84%, .45. According to the encoding specificity principle, memories are best retrieved with cues that were processed at the time of encoding. Therefore, even though the associations between table and chair, and between furniture and chair, must be stronger in some absolute sense, stone will be the better retrieval cue in this instance, because it was stone, and not chair or furniture, that was processed at the time of encoding.

5. How does use of a schema improve memory?

a. A schema improves memory for details.

b. A schema provides a framework to use in interpreting a situation.*

c. A schema helps avoid making errors in remembering the details of a situation.

d. All of the above answers are correct.

73%, .33. A schema is defined as the cognitive framework of knowledge, expectations, and beliefs that provide a framework for perception, memory, and thought. Schemas improve memory in two ways: by providing extra cues for the retrieval of schema-congruent information; and by fostering elaborative activity during the encoding of schema-incongruent information.

6. A recent news report tells that a patient with serious and permanent anterograde amnesia has, through special training, been able to remember some things well enough to use them at a job. From what you know about memory, you guess that the kinds of things this patient has acquired entail __________ memory.

a. verbal

b. procedural*

c. nonverbal declarative

d. episodic

95%, .29. This probably refers to new cognitive and motor skills, and skills of all sorts are part of procedural memory. There's no mention of whether the knowledge in question is verbal or nonverbal.

7. In contrast to the elaboration principle, the organization principle refers to

a. rote repetition of items in short-term memory.

b. relating individual list items to each other.*

c. the mechanism by which groups of memories are lost from storage.

d. items in memory interfere with each other unless they are consolidated in memory.

90%, .25. Elaboration refers to item-specific processing: linking individual pieces of new knowledge with knowledge already stored in memory. Organization refers to inter-item processing, linking individual pieces of new knowledge with each other. Chunking to expand the effective capacity of short-term memory is an example of organization, but chunking goes beyond rote repetition.

8. The time-dependency effect in long-term memory reflects mostly

a. the fact that encoding a memory takes time.

b. loss from memory storage.

c. interference among available memories*

d. schema-incongruent rather than schema-congruent memories.

56%, .29. Time-dependency refers to the fact that, in general, forgetting increases with time. Several mechanisms have been postulated for time-dependency, including passive decay, displacement, consolidation failure, and interference. But decay and displacement don't occur in long-term memory, and consolidation (and thus consolidation failure) is the process by which items are stored in long-term memory. So the only option left is interference among items that are already encoded, consolidated, and stored, available in long-term memory.

9. The post-event misinformation effect shows that

a. memories, once encoded, remain permanently accessible in storage.

b. knowledge acquired after an event can modify a memory trace of that event.*

c. emotional events are more subject to distortion than nonemotional events.

d. memory, in contrast to perception, is not subject to illusions.

95%, .26. In the post-event misinformation effect, information acquired after an event has been encoded in memory (like the statement that the Datsun was stopped at the stop sign) affects memory for the event itself, leading to the false (illusory) recognition of the picture of the car with a stop sign as opposed to the yield sign. Memories, once encoded, are permanently available in storage, not permanently accessible -- there are lots of processes, like that failure to supply adequate or appropriate retrieval cues, that will render available memories inaccessible to retrieval. But what the post-event misinformation effect seems to show is that memory traces can be modified retrospectively, after the fact. Which is why the stop sign is falsely recognized, even though it was a yield sign that was originally encoded.

10. According to the normative model for human judgment and decision-making:

a. people attempt to minimize the ratio of gains to losses.

b. people multiply the optimality of an outcome by its utility.

c. people evaluate uncertain outcomes in terms of their probabilities.*

d. people take into consideration the how they achieved their current state when evaluating possible future outcomes.

40%, .16. A bad item. The normative model says that people reason logically, and make "rational" choices that maximize gains and minimize losses in the most efficient manner possible, paying attention only to their current state and the future states represented by their choices. In this model, the values of uncertain outcomes are calculated by multiplying the value of the outcome by its probability. Thus, a 50-50 chance of winning $10 is worth $5, as is a 1/20 chance of winning $100. I actually thought this was an OK item, but here's another instance where the statistical analysis trumps subjective impressions: the item-to-total correlation is unacceptably low, so out it goes.

11. The fact that category members differ in terms of "typicality" violates which assumption of the classical view of categories as proper sets:

a. category instances all share a set of defining features in common.

b. category members can be arranged into vertical hierarchies.

c. there are sharp boundaries between adjacent categories.

d. all members of a category are equally representative of that category.*

74%, .45. According to the classical view of categories as proper sets, all members of a category are equivalent because they all share the same set of singly necessary and jointly sufficient defining features. But even in cases where there is no question that the category is a proper set, like odd and even numbers, geometrical figures, and females, there are variations in typicality. This indicates that categorization is not governed by defining features, but rather by something else -- like prototypes.

12. The simulation heuristic differs from the availability heuristic in that:

a. availability is used to judge similarity, while simulation is used to make estimates of frequency.

b. in availability, choices are made out of habit, while in simulation, choices are made on the basis of authoritative pronouncements.

c. the availability heuristic is based on accessibility in memory, while the simulation heuristic is based on ease of imagination.*

d. availability bases judgments on objective probabilities, while simulation bases judgments on subjective probabilities.

90%, .38. In availability, judgments of frequency and probability are based on the ease with which instances can be retrieved from memory. In simulation, the same judgments are based on the ease with which examples can be generated in imagination.

13. The Roman numeral X is an example of a(n) __________ representation.

a. symbolic*

b. analogical

c. hypothetical

d. psychological

86%, .28. In the Roman numeral system, X is a symbol that stands for the number 10. It doesn't look anything like 10 anything. Resemblance is the distinguishing feature of an analogical representation. Thus, the Roman numerals I, II, and III are arguably analogical in nature, because they involve, respectively, one, two, and three things. But there aren't 10 things in the Roman Numeral X, and there aren't 100 things in the Roman numeral C (and, for that matter, there aren't to Xs in the Roman numeral C either!). X and C are symbols that represent 10 and 100, respectively.

14. A subject in an image-scanning task is asked to mentally "travel" on four different journeys from a single given location. Which journey would take the most time to make?

a. a journey of one mile

b. a journey of five miles*

c. a journey of a half-mile

d. a journey of three miles

83%, .42. Don't confuse image-scanning with the images produced by brain-scanning devices like PET and MRI (though if you did, that wouldn't hurt you on this question). The fact is, that when people inspect a physical map, it takes them longer to move their eye fixations between distant points, compared to nearby points. And the same holds true when people inspect a "mental" map. This kind of finding is generally considered proof of the equivalence of perception and imagery.

15. According to the network model of generic memory, to which of the following questions would you probably respond most slowly?

a. "Is a cat an animal?"

b. "Is aluminum a metal?"

c. "Is a pine a tree?"

d. "Is a penguin a bird?"*

96%, .25. In network models, concepts are represented by nodes, and the strength of the association between concepts is represented by the length of the link between them. Because cats are typical animals, aluminum a familiar metal, and pine a familiar tree, the associative links between these concepts are relatively short, so it takes less time for activation to flow down them, resulting in short reaction times. But penguins are very poor, uncommon examples of birds, so the associative link weaker, thus longer, thus it will take longer to verify that a penguin is, in fact, a bird. The inverse relationship between typicality and response latency -- high typicality, short latency -- is one of the problems with the classical view of categories as proper sets.

16. In a problem-solving situation, what is considered in a means-end analysis?

a. random associations

b. only the eventual goal

c. only the current state of affairs

d. both the current state of affairs and the eventual goal*

95%, .22. A means-end analysis calculates the difference between the current state and the goal state, and then takes steps intended to reduce that difference. Nothing else matters. it's an algorithm for problem-solving. And it always works.

17. When compared to a chess player who is competent but who is below the master level, at which type of memory is a chess master superior?

a. chunking memory*

b. visual memory

c. working memory

d. both a and b

35%, .35. Expert chess players have better memories for the board than do novices, or even less-expert players, because they don't have to memorize the positions of each individual piece. Rather, they can organize the positions into familiar categories, like "Queen's Gambit Declined", which take up less memory capacity, and thus allow them to retain more information, and retrieve it more efficiently. A hard item, perhaps, but the item-to-total correlation shows it was fair.

18. What does research on deductive reasoning indicate?

a. Research participants are generally quite good at judging the validity of syllogisms.

b. Research participants frequently assess a syllogism's conclusion based on its plausibility.*

c. Research participants usually evaluate a syllogism by determining whether the conclusion logically follows from the premises.

d. Both a and c are correct.

34%, .33. Another hard-but-fair item. In deductive reasoning, people draw conclusions about specific instances from general principles represented by the premises. There are logical rules for solving syllogisms, but people don't always follow them. Instead, they are likely to base their judgments on the inherent plausibility of the conclusions, not whether it logically follows from the premises. This is one of the departures of human reasoning from the normative model.

19. Last week, Mike heard about five separate airplane crashes on the news. Even though, overall, motorcycle accidents account for more accidents than plane crashes do, Mike decides to ride his motorcycle from Washington to Atlanta instead of flying. What is Mike's reasoning error based on?

a. the availability heuristic*

b. mental set

c. a syllogistic error

d. a problem of framing

89%, .49. This is a pretty standard example of the availability heuristic. Airplane crashes are fresher in memory, memory strength affects availability, and availability affects judgment.

20. Say you meet a woman who opposes the death penalty, then decide that, generally, women are more likely to oppose the death penalty than are men. That is, you assume that the individual case is representative of its category. Here you have fallen prey to

a. restructuring.

b. the representativeness heuristic.*

c. chunking.

d. confirmation bias.

92%, .47. And this is a pretty standard example of the representativeness heuristic. In representativeness, judgments are based on resemblance - -that is, that instances resemble their parent population. So, having met one woman who opposes the death penalty, you figure that most women do as well.

21. Unconscious processing tends to be fast and efficient. This raises the question, What is consciousness good for? One answer is as follows:

a. It forces different brain areas to make connections with one another.

b. It allows us to break away from routine and habit.*

c. It tends to produce insights that are more creative.

d. It allows for incubation.

52%, .40. Unconscious, automatic processing is habitual and routine. Therefore, conscious processing is not these things.

A student wrote: I understand how consciousness allows us to break away from routine and habit, but by doing so, doesn't it also tend to produce insights that are more creative? I imagine that creativity cannot be readily accomplished while unconscious (since it is routine), so it would require that the conscious be used for creativity. Response: B really is the best answer. As the text puts it (p. 306): "Perhaps we need consciousness in order to break away from routine and habit". some of these cognitive breakouts may produce the occasional creative insight, but there's nothing in the text that would suggest that it does this most of the time, which is what's implied by the word "tends" in Option C.

22. Which of the following is an example of a single morpheme?

a. f

b. trip*

c. tl

d. talked

85%, .49. Just as a phoneme is the smallest unit of sound in a spoken language, a morpheme is the smallest linguistic unit that carries meaning. "F" is a phoneme, but it doesn't mean anything. "TL" is a combination of two phonemes, but they still don't mean anything. "Trip" is a single morpheme that means something like stumbling and falling. There are no smaller, component morphemes inside. But "Talked" is a word composed of two morphemes: "TALK-", meaning to speak, and "-ED", indicating that the speaking occurred in the past.

23. A concept has a family resemblance structure when

a. its meaning is different from its reference.

b. it is described by a set of defining features.

c. it is described by a set of defining features, no one of which is individually either necessary or sufficient.*

d. it cannot be represented by a mental image, but it shares features with other words.

59%, .40. According to the classical view of category structure, all category instances share the same set of defining features. But according to the revisionist view of categories as fuzzy sets, there are no features that all category members have in common. Some birds fly, like robins and sparrows, others don't, like penguins and ostriches. Some birds have long beaks, like sandpipers and eagles, but other birds don't, like finches and bluebirds. Members of a family resemble each other in much the same way, as when we say that a child has her mother's eyes and her father's nose. What unites the members of a category is not some set of defining features, but rather a pattern of family resemblance.

24. Which of the following statements is false?

a. Initially, Japanese babies can distinguish between the sounds "la" and "ra."

b. Most Japanese adults cannot perceive the difference between the sound "la" and "ra."

c. By two months of age, infants will listen to their native language and to a foreign language for similar periods of time without habituating.*

d. Children learn the phonemes of their languages by ignoring the sound distinctions that are not used in their language.

57%, .21. Newborn Japanese babies can distinguish between Ls and Rs just fine, but they lose this ability as they are exposed to the distinctive phonology of Japanese, which lacks these particular phonemes. This happens pretty quickly, so that even as early as two months of age, babies distinguish between the phonemes of their caregivers' language and those of an unfamiliar foreign language.

25. Genie, a fourteen-year-old child, was abused all her life and never exposed to language. As a result, she was never able to develop normal language skills. Isabelle, a six-year-old child, was given only minimal attention and never exposed to language, yet was able to learn normal language skills within a year. In what aspect of their lives did these two children differ that so affected the eventual outcome of their language development?

a. Genie was past puberty when she was first exposed to language; Isabelle was prepubescent.*

b. Genie's parents could neither speak nor hear; Isabelle's mother could do both.

c. Genie was born with a lower than normal IQ; Isabelle was born with a nearly normal IQ.

d. Genie suffered severe physical abuse; Isabelle did not.

96%, .33. The decisive factor in language development is age. There's a critical period for normal language development, which ends roughly at puberty. Genie wasn't exposed to language until after that time, while Isabelle was.

26. In terms of language, psychologists generally agree chimpanzees can learn

a. to refer to events in the past tense.

b. a substantial number of manual signals for words.*

c. rules to put words together in different sequences so as to express different meanings.

d. All of the above answers are correct.

90%, .41. Chimpanzees can learn manual signs and other symbolic representations for concepts -- that is words -- but they don't seem to have the ability to use grammatical rules to string "words" to form new sentences. In other words, they've got some capacity for semantics, but not for syntax. And neither, apparently, do any other nonhuman species.

27. Which of the following research findings is inconsistent with Whorf's hypothesis about the relationship between language and thought?

a. The Dani people of New Guinea have only two color names, but they tend to agree with American observers about which colors are particularly good examples of their categories.*

b. The Mayan people who speak Tzeltal have no language labels for "left of " or "right of," nor do they distinguish between mirror images.

c. The Australian people who speak Guugu Yimithrr only speak of objects as positioned north, south, east, or west of other objects, and they regard layouts of objects as similar only if they are aligned in these ways.

d. All of the above answers are correct.

75%, .49. Whorf (and Sapir) argued that language somehow constrained thought -- that the way we spoke about things affected the way we could think about them. In order to test Whorf's hypothesis, you'd have to take a group of people, like the Dani, who have some distinctive feature of language (like the fact that they have only two color names), and see if that affects how they perform on some cognitive task (like a test of color perception). It runs out that Dani behave just like Americans, which is not consistent with Whorf's hypothesis. The observations in options B and C seem, superficially, consistent with Whorf's hypothesis. But notice that they don't have a control group. We don't know, from the information given, how people whose language has different properties would behave when tested under the same circumstances.

28. According to the Doctrine of Traits, traits are

a. dispositions to behave in particular ways.*

b. dispositions to evaluate people and other objects positively or negatively.

c. dispositions toward particular positive or negative feeling states.

d. dispositions to approach or avoid certain objects.

87%, .36. There are lots of different internal, personal determinants of behavior (as opposed to external, environmental or situational determinants). Traits are behavioral dispositions, like friendliness or conscientiousness -- dispositions to behave in friendly or conscientious ways. Attitudes are dispositions to evaluate objects (like minorities or Democrats) positively or negatively. Moods are dispositions to experience positive or negative emotions. And motives are dispositions to approach or avoid certain objects (like cheeseburgers or spiders).

29. Bem's theory of self-perception holds that

a. our perceptual apparatus has evolved in such a way that we can automatically gain access to our true attitudes, without any need for "higher-order" processes of reasoning and inference.

b. we have direct knowledge of our own attitudes, but must infer the attitudes of other people.

c. we form attitudes that are consistent with our behavior.*

d. attitudes predict behavior only in situations that we perceive to be highly self-relevant.

69%, .29. Self-perception theory is an illustration of the reciprocal causal relation between Behavior and the Person. What Bem argued, and what phenomena like the foot-in-the-door effect illustrate, is that while we usually think of attitudes causing attitude-consistent behavior, sometimes the behavior comes first, and the attitude later. That is, we perceive ourselves behaving in a particular way (like voting for a Republican), and then infer that we have pro-Republican attitudes. Note the emphasis on inference: according to Bem, we don't really know our attitudes: we just infer what they are from observations of our own behavior -- just as we infer the attitudes of other people from observing their behavior.

30. The presence of other people

a. facilitates or inhibits helping behavior, depending on what they are doing.*

b. tends to facilitate helping behavior.

c. tends to inhibit helping behavior.

d. none of the above.

61%, .08. The effect of other people on helping behavior depends on the details of what they are doing. In the Darley and Latane studies, the presence of other people who are just sitting around, doing nothing -- a state of pluralistic ignorance -- deters helping behavior (Option C). But in studies of modeling effects, the presence of others who are actually engaged in helping behavior tends to stimulate others to help as well (OptionB). So the best answer really is A.

31. Children are more likely to delay gratification if they

a. wait in the presence of the promised reward.

b. wait in the absence of the promised reward.*

c. they are given a choice of immediate or delayed reward..

d. think about what it will be like to consume the promised reward.

83%, .39. In the printed version of the exam, there were two Option Cs, but that shouldn't have mattered, because the correct answer is D. Anyway, there's no delay of gratification if children are given a choice (Option C), because gratification doesn't have to be delayed if reward can be delivered immediately. So delay of gratification only makes sense if the reward is actually delayed. Under these circumstances, the key is "out of sight, out of mind". Children can wait longer for a delayed reward if it's not physically present in the room. If it is physically present, they can wait longer if they engage in behaviors, such as closing their eyes, that put it out of sight. And if it remains in sight, they can wait longer if they think about it in nonconsummatory ways -- like thinking about promised marshmallows as clouds instead of as sweet treats.

32. Stereotyping and prejudice illustrate how people can affect the environment through

a. evocation.*

b. selection.

c. behavioral manipulation.

d. cognitive transformation.

33%, .25. Another hard-but-fair item. Stereotypes are generalizations, usually inaccurate, about whole groups of people -- people who are identified by some distinguishing characteristic such as skin color, gender, age, ethnicity, or religion. A member of a stereotyped group doesn't have to do anything in order to be treated differently. Rather, the distinguishing characteristic of the stereotyped group is enough to evoke prejudicial behavior from other people. Consider the "Baby X" studies -- simply being told that the infant was a boy or a girl led adults to treat it differently. Some of you went for D, which wasn't a bad answer, but when it comes to B, P, and E we're always looking at B from the point of view of P. And in stereotyping, the person isn't thinking anything. He's just different-colored skin, or has a funny accent, or he's a she. Other people may be thinking about him or her, but he or she him- or herself is just sitting there being stereotyped by others.

A student wrote: The question seemed somewhat ambiguous to me. I see how the people who
are stereotyped can evoke prejudice in their environment. But can't behavioral manipulation by the stereotyper also evoke prejudice in the environment. I feel that "people" could mean either of these two situations. Response: Behavioral manipulation by the stereotyper doesn't evoke prejudice. Rather, the stereotypers prejudiced behavior is the behavioral manipulation. But, again, we're looking at the situation from the point of view of the person being stereotyped, which is why I discussed stereotyping, as in the studies of gender stereotyping, in the context of evocation. And the point about evocation is that the stereotyped person doesn't have to do anything to be stereotyped, and subject to prejudice. S/he just has to be who s/he is.

And another student wrote: Looking over the answers to the exam, I'm having a hard time understanding why Evocation is a better answer than Cognitive Transformation for Question #32 on Stereotyping and Prejudice. Looking back at your Final Exam for 2000, I found this question:

76. People who believe that someone has been diagnosed with a mental illness often treat that person differently than they would if that person were "normal". This tendency reflects:

A. evocation
B. selection
*C. cognitive transformation
D. cognitive dissonance

32%, .18. A BAD ITEM. People affect the environment, which includes other people and what they do, in a number of ways. In evocation, the effect is mediated by the person's mere presence or appearance. In selection, by the person's choice of one environment over another. In behavioral manipulation, by the person's overt behavior, which changes the objective environment, for everyone. In cognitive transformation, by the person's thought, which changes the private, subjective, internal representation of the environment. In this case, a belief, or cognitive transformation, starts the process. If they did not believe that the person had been diagnosed with mental illness, or they had some other belief about mental illness, the process would have gone differently. Cognitive dissonance has to do with how we resolve conflicts between beliefs, or between our beliefs and our behavior.

This answer seems to support Cognitive Transformation as the answer to #32 because it is people's private and internal representations of other's characteristics that cause them to change their behavior towards them, in the same way a person hears of someone getting diagnosed and changes their behavior. Rather than a minority just being present to create prejudice against him, the majority must have a representation of that minority installed in their heads. If this past problem does not show any reason why Cognitive Transformation is correct, I think it at least shows that it should be considered a Bad Item. Response: Well, that's a different item, which refers more clearly to people's beliefs, which is a sure cue that a cognitive transformation is involved.

Reflection: I think that the difference between evocation and cognitive transformation is clear in the lectures -- the first is passive, based on a person's appearance or other characteristics, the second is active and mental. But I do think that I should probably make clearer, in the future, that evocation, selection, behavioral manipulation, and cognitive transformation are always viewed from the perspective of the Person -- whoever is the P in Lewin's pseudoequation. True, the environment contains other people, who react to the Person in various ways, and that is what may have made the question somewhat confusing. We can debate the wording of the question, but in the final analysis the decisive test of a bad item is to be found in the item analysis. This is an objective standard, and doesn't rest on anyone's -- yours or mine -- subjective interpretation or nuance. No doubt, this was a difficult item. But the fact remains that choosing Option A was positively correlated with performance on the rest of the test, and choosing any of the remaining options was, in fact, negatively correlated with overall test performance.

33. The fundamental attribution error predicts that in making attributions for the behavior of others, we tend to

a. overemphasize situational factors.

b. overemphasize dispositional factors.*

c. assume that their behavior has the same causes as our own behavior.

d. overemphasize the role of chance.

78%, .55. In theory, causal attributions are governed by Lewin's equation B = f(P, E). That is, people try to figure out whether a person's behavior is caused by some internal, personal disposition or by some external, environmental factor. In general, however, people seem to be trait theorists: that is, they tend to overemphasize dispositional factors like traits, and de-emphasize situational factors like external social incentives. At least, that is, when they explain the behavior of other people. When it comes to explaining their own behavior, people tend to be situationists -- to attribute their own behavior to factors in the situation in which the behavior takes place. This is known as the self-other (or actor-observer) difference in causal attribution.

34. Self-serving bias refers to the observation that we tend to

a. take credit for our successes and attribute our failures to external factors.*

b. take responsibility for both our successes and our failures.

c. take responsibility for our failures and attribute our successes to external factors.

d. attribute both our successes and our failures to external factors.

98%, .20. Here's another constraint on causal attribution: it turns out that, in explaining our own behaviors, the tendency to favor situational causes occurs mostly for failures and other negative outcomes -- they're just not our fault! But for successes and other positive outcomes, we actually make dispositional attributions -- things worked out so well, because that's the kind of people we are.

35. One of the factors that influences whether there is a relationship between attitudes and behavior is whether

a. the attitude is specific or general.

b. situational pressures favor or discourage behavior consistent with the attitude.

c. the attitude is strongly or weakly held.

d. All of the above.*

70%, .24. Attitudes are causally related to behavior, just as traits are, but -- as with traits (remember Mischel's personality coefficient of .30?) there are limitations on predictability. There's better prediction from specific attitudes than from general ones: we can pretty accurately predict how you'll vote on an anti-abortion referendum if we know your particular pro-life vs. pro-choice attitudinal stance; but knowing whether you're conservative or liberal, or Republican or Democrat, doesn't give us as much predictive leverage. And, of course, prediction is better when people's attitudes are strongly held. But there's also the nontrivial fact that dispositions like attitudes aren't the only determinants of behavior: the details of the evoking situation matter, too. Even the most committed pro-life or pro-choice proponent may vote in a contra-attitudinal manner, depending on the way the law is worded.

36. Dissonance theory would predict that those who work hardest in a class will find that class

a. less enjoyable.

b. more enjoyable.*

c. as enjoyable as someone who did not work hard.

d. boring but informative.

83%, .36. Dissonance theory asserts that people generally behave in accordance with their attitudes, and when they do not, the discrepancy between attitudes and behavior creates an unpleasant emotional state, which has to be reduced somehow.And one way to do that is to change the attitude to conform the behavior. And that's what happens in this case. People who make big sacrifices, of time or effort or whatever, justify these sacrifices by enhancing their favorable attitude toward the activity. Why would you work hard at something you don't enjoy? If the reward were great, as in admission to graduate school, you could attribute the hard work to the external incentive. But in this case, lacking information about external constraints and incentives, dissonance is reduced by enhancing one's attitude toward the activity.

37. One of Cannon's major objections to the James-Lange theory of emotion was that

a. different emotions produce the same bodily changes, so it is therefore difficult to tell how a person knows what he or she feels.*

b. the sympathetic nervous system was not given a more prominent role.

c. emotions have different intensities.

d. emotions are subjective experiences; therefore, no theory can be adequate in explaining them.

72%, .44. James and Lange thought that emotions reflected people's perceptions of bodily responses to a stimulus. Thus, emotions like joy, fear, and sadness are associated with distinctive patterns of physiological response. But Cannon couldn't find any distinctive patterns (and psychologists still haven't, not really). That suggested to Cannon that, physiologically, all emotional states were based on the same pattern of physiological change, which he called undifferentiated arousal. Of course, that left Cannon with the problem of how it is that we do distinguish among emotional states of joy, fear, and sadness -- a problem that he never satisfactorily solved, and which remains unsolved to this day.

38. Imagine an Asch-style social pressure experiment in which the judgments are made more difficult (e.g., 6.5 vs. 6.25 inches) than in the standard task. Most likely, compared to the standard task, participants in the new version will

a. yield more and be more emotionally disturbed.

b. yield more and be less emotionally disturbed.*

c. yield less and be more emotionally disturbed.

d. yield less and be less emotionally disturbed.

57%, .48. We tend to look to others for cues when the situation is ambiguous. So, when the discrimination task is difficult, subjects will be uncertain, will look for cues from other people, and will be less bothered by the fact that they conformed -- because there's little or no reason to think that the others' judgments are, in fact correct. But subjects will be less likely to conform when the task is easy, because they'll believe their own eyes; and when they do conform, they're going to be more upset, because their behavior contradicts their perception. Put another way, they'll be in a state of cognitive dissonance.

39. Laboratory studies indicate that leaders are least likely to be effective if

a. they are more intelligent and dominant than others in the group.

b. the task to be performed is clear-cut rather than ambiguous.

c. they have relatively little authority within the group.*

d. the members of the group get along well with each other.

91%, .25. This is almost a tautology: effective leaders have authority. But the point is that when people are randomly assigned to the role of leader, those who already have some authority within the group prove to be more effective leaders than those who do not. Intelligence and dominance, of course, are positively correlated with leadership (though the correlations aren't as strong as you might think). If a task is clear-cut, groups don't need much leadership: it's already clear to everyone what they should do. And, of course, leaders are more effective if the group functions well as a group.

40. Pluralistic ignorance refers to the fact that

a. if nobody knows what to do in an emergency, no action will be taken.

b. when there is a large number of people present, there is a tendency for bystanders to pretend that nothing is happening.

c. when other bystanders do not take action, those present are likely to define the situation as a non-emergency.*

d. in an emergency, large groups of people are easily swayed by a single dominant individual.

85%, .33. This comes from the Darley & Latane studies of bystander intervention. When a situation is ambiguous, people look to others for cues as to what to do (remember the Asch conformity experiment described above). But when the others are also looking around, instead of taking action, their behavior is cueing that there really isn't an emergency after all. It's through pluralistic ignorance, then, that the presence of others deters helping behavior.

41. It is generally true that

a. people who live far apart tend to find one another more attractive.

b. opposites attract.

c. people with similar characteristics and interests are attracted to one another.*

d. people who differ on some critical dimension, such as skill in sports, are attracted to one another.

94%, .38. Proximity breeds liking (through exposure), so does similarity, and familiarity, far from breeding contempt, breeds interpersonal attraction.

42. Correlation coefficients have one important limitation. What is it?

a. They are not usually reliable.

b. They do not provide a measure of the degree to which two variables are related.

c. Negative correlations are more common than positive correlations.

d. They can't, by themselves, tell us whether or not the variables are causally related.*

88%, .38. Correlation does not necessarily mean causation. Sure, effects are correlated with their causes, but it's also the case that two variables that are correlated with each other might both be the effect of some third variable.

43. The predictive validity of most scholastic aptitude tests is in the neighborhood of 0.55. This suggests that

a. these tests do not really measure scholastic aptitude.

b. factors other than scholastic aptitude influence academic performance.*

c. such tests have little or no predictive validity.

d. such tests provide a near-perfect index of academic ability.

75%, .28. Validity is concerned with whether a test measures what it's supposed to measure. So, with predictive validity coefficients in the mid .50s, there's clearly a substantial, if not perfect, correlation between predictor (test score) and criterion (academic performance). The predictive validity of test scores isn't perfect, however, because other dispositional and situational variables also affect academic performance. For example, some people with high scholastic aptitude might be low on achievement motivation: these people are sometimes called underachievers (or, simply, slackers). Or, some people with low scholastic aptitude might work really hard at their coursework: these people are sometimes called overachievers. Achievement motivation, of course, is a dispositional variable. But there might also be situational variables that would affect academic performance: for example, a person with relatively low scholastic aptitude might choose an easy major, and get all As; or a person with relatively high scholastic aptitude might choose a really difficult major, and get all Cs. Note, in this case, how Persons can affect Environments through Selection: by choosing an easy major, a low-ability person puts himself in an environment where he is likely to succeed.

44. Fluid intelligence refers to

a. previously acquired skills and information.

b. the ability to deal with new problems.*

c. verbal ability.

d. mathematical ability.

92%, .40. Of the various kinds of intelligence, fluid intelligence refers to "raw" intelligence that, like flowing water, can be moved in any direction -- toward math, or writing, or music, or whatever. By contrast, crystallized intelligence refers to knowledge and skills that the person has already acquired. Crystallized intelligence increases with age (if for no other reason than that old people have had more time to acquire more knowledge), while there is some evidence that fluid intelligence declines (you can't teach an old dog new tricks).

45. One piece of evidence supporting Howard Gardner's notion of multiple intelligences is based on the fact that

a. some people have multiple personalities.

b. people seem to lose one type of intelligence earlier than others.

c. brain lesions may impair some abilities while leaving others unaffected.*

d. at different times in our lives, we may excel at one type of intelligence at the expense of others.

65%, .47. If intelligence were really a highly generalized intellectual ability, like Spearman's g, performance on all sorts of intelligence tests would be correlated (and, arguably, they'd all be based in the same area of the brain, like the prefrontal cortex). But Gardner has observed that there are cases of brain damage where some intellectual functions are grossly impaired, even though others are spared. This suggested to him that intelligence isn't One Big Thing; rather, that there are a number of qualitatively different "intelligences", each mediated by a different brain system.

46. Researchers administer IQ tests to a group of five- and 15-year-old adoptees, along with their biological and adoptive mothers. The researchers then calculate correlations between the children and both sets of mothers. Most likely,

a. the highest correlations will be between the 15-year-olds and their biological mothers.

b. the lowest correlations will be between the 15-year-olds and their adoptive mothers.

c. comparisons of correlations will provide evidence for a genetic influence on IQ.

d. All of the above.*

61%, .16. A lot of you went for C, which is true, but A and B are also true. There is a substantial genetic contribution to intelligence. We know that because identical twins are more alike in their IQ scores than fraternal twins. Similarly, there is a higher correlation between the IQs of biological relatives than there is between adoptive relatives. So, children's IQs are more highly correlated with that of their biological mothers (and fathers, too, for that matter), than they are with that of their adoptive mothers (and fathers).

A student wrote: I'm thinking this may be a better time to try for an absolution on the test than after the key comes up online: On question #46, the IQ tests were administered to 5- and 15-year-old adoptees... adopted at 5 and 15? Or adopted as infants, and tested at 5 and 15? I answered B, on the logic that a 15-year-old adoptee would have spent less time with their high-IQ parents, because they were adopted at a later age. The correlation would be lower than those who were adopted at 5, and so spent all those additional years with the high IQ adoptive parents. Afterward, when I sat down and was reviewing the exam with a friend (who is not taking the course) she pointed out that you may have intended for us to assume that both groups had been adopted as babies...which would
exactly flip the answer. Eek. Of course, if I'm the only one who failed to pick up that intention, I'll
have to suck it up and accept that I have an unusually broad interpretation on the adoption question. But I thought it was worth mentioning, anyway. I guess I'll find out soon. Thanks for "listening". Shortly thereafter, the same student wrote: So now that I've seen the key, turns out I read the possible answers wrong anyway, and missed the point as well. I do think the wording is a shade
hinky, but in all honesty if I'd actually remembered the point (genetics over upbringing reflected in IQ) the wording would not have mattered to my answer. Response: Well, the wording might be a shade hinkey, as you say, but(as you also say) the basic point is not that heredity is probably more important than environment in determining IQ: the point is simply that genetics does make a contribution to IQ; it's not necessarily the case that heredity is more important than environment.

Along the same lines, another student wrote: I do not understand how the highest correlation would be between the 15-year-olds and their biological mothers. It seems more accurate that the highest correlation would be between the 5-year-old and their biological mothers. The 15-year-olds have spent much time away from the biological mothers whereas the five-year-olds have only been gone for a short time. So therefore other factors would alter the 15-year-old's IQ more so than the 5-year-old. Similarly, the 5-year-old correlation with the adoptive mother would seem lower than the 15-year-old with the adoptive mother because the 5-year-olds have not spent as much time with their adoptive mothers as the 15-year-olds have. Therefore, the 15-year-old would be similar much more because it has been in the same environment longer. Response: In adoption studies, the general finding is that, no matter how you cut the data, the correlations of children's IQs with their biological mothers (or fathers, for that matter) are higher than the correlations with their adoptive parents. This is prima facie evidence for a genetic influence on IQ, so Option C is true. Moreover, as noted on p. 541, it's an amazing fact, but true, that the correlation between children and their biological parents increases as children get older. So, while we would expect the child-biological parent correlations to be significant for both the 5- and the 15-year olds, we could expect that correlation to be higher for the 15-year-olds. And because child-biological parent correlations are higher than child-adoptive parent correlations, the the child-biological parent correlations will be highest for the 15-year olds. So, that makes Option A true. And because Option C is also true, the right answer must be Option D. But what about Option B? It turns out that this is also true, which further reinforces Option D as the correct response. In the first place, it turns out that there is essentially no correlation between the IQs of adoptees and those of their adoptive parents. So, we're already starting from a pretty low point. But precisely because children become more similar to their biological parents as they age, they will also become even less similar to their adoptive parents, than they were when they were younger.

Again: In the final analysis the decisive test of a bad item is to be found in the item analysis. This is an objective standard, and doesn't rest on anyone's -- yours or mine -- subjective interpretation or nuance. No doubt, this was a difficult item. But the fact remains that choosing Option D was positively correlated with performance on the rest of the test, and choosing any of the remaining options -- even Option C, which would be the most intuitively reasonable of the incorrect options -- was, in fact, negatively correlated with overall test performance.

47. Jerome and Jenny are each given two tests of fearfulness, both rated on a 10-point scale, on which zero means no fear and 10 means maximum fear. In the first test, both Jerome and Jenny are confronted with a vicious dog. Here, Jerome's fear rating is 9 points and Jenny's is 9 points. In the second test, both are about to take a difficult examination. Now, Jerome's rating is 3 points and Jenny's rating is 3 points. These results illustrate

a. situational effects.*

b. differences in personal traits.

c. a person-by-situation interactions.

d. None of the above.

79%, .29. I pretty much always ask a question like this. And the way to solve the problem is to calculate the averages across situations and across persons. In this case, Jerome shows 6 points' worth of fear, across the two situations, and so does Jenny. So, there's no difference in fear depending on the person: both individuals are equally fearful. But both Jerome and Jenny show 9 points' worth of fear of the dog, and 3 points worth of fear of the test. So there's a difference in fear depending on the situation. Note that the difference across the two situations is the same for both people: so, there's no person-by-situation interaction.

48. In critical evaluation of Freudian theory,

a. evidence from clinical practice cannot be totally objective.

b. Freud's theory makes few specific predictions and is very difficult to refute.

c. a patient's statements can be interpreted to mean what they say or the opposite depending on the analyst's interpretation.

d. All of the above.*

51%, .22. Freud's famous, but that doesn't mean he was right. In fact, each of these is a serious problem with Freudian psychoanalytic theory, which is why it's taught more as literature than as science these days. We give you just enough Freud so that you can appreciate some of the literature, plays, and the like that Freudian psychoanalysis inspired.

49. Children who can delay gratification in early childhood generally

a. lose this ability as they grow older.

b. are more self-reliant and perform better under stress as adolescents and adults.*

c. feel as if they have been cheated out of many things and develop negative characteristics, such as impatience, as adults.

d. make friends more quickly in social settings.

95%, .16. There is some stability to delay of gratification, such that young children who can delay gratification grow into adults who are more patient, and more self-reliant, and who can cope better with the problems they confront. The correlations are far from perfect, but they're there. There is some degree of temporal stability to behavior, just as the Doctrine of Traits would suggest.

50. According to Abraham Maslow, a major prerequisite for becoming self-actualizing is having

a. all of one's lower-order needs fulfilled.*

b. a major altruistic streak.

c. a very selfless nature.

d. suffered in the past so you can truly appreciate the good aspects of life.

88%, .52. Maslow, like Freud, is much read: Even though the scientific foundations of his theory of self-actualization are relatively weak, those of you who go on to business school, particularly in personnel, will encounter his ideas. For Maslow, self-actualization is the highest human need. But Maslow also postulated a hierarchy of needs, and asserted that people progressed up the ladder: fulfillment of a lower-order need is a prerequisite for fulfilling the next-higher order need, and so on, all the way up the ladder. This means that one can't even work at self-actualization until all of one's lower-order needs (physiological, safety, esteem, and the like) have been met. It's worth it, though, in Maslow's view, because becoming self-actualized makes it possible to have "peak experiences" -- which, so we're told, are pretty good things.

Retain this exam, along with a record of your answers.

A provisional answer key will be posted to the course website by 3:00 PM today.

The exam will be provisionally scored to identify and eliminate bad items.

The exam will then be rescored with bad items keyed correct for all responses.

Grades will be posted to the course website.

A final, revised, answer key, and analyses of the exam items,

will be posted on the course website when grades are posted.

Requests for rescoring must be received within

one (1) week of the posting of grades