University of California, Berkeley

Department of Psychology


Psychology 129 / Cognitive Science 102

Scientific Approaches to Consciousness

Spring 2011


Midterm Examination

Scoring Guide


What follows is the scoring guide prepared for the GSIs to help them grade the exam.  Other answers may also be appropriate, which is why it's called a scoring guide.  Based on the item analysis, I provide the percentage of the class who attempted each item, and the mean score achieved by those students.

Thanks to the GSIs for their grading of the midterm exams -- especially given the demands of Spring Break.  

The mean score on the exam, before editing, was 34.74 (SD = 7.11).  That's an average performance of 69.5%, which is within the range of 65-70% correct, which I like to see as a minimum average score on my exams.  

The mean number of questions answered was 24.02 (SD =2.01), with the majority of the class (66.4%) attempting all 25 questions, which suggests that the 25-question exam may have been a little too long.  Still, I needed 25 questions to cover the material.  

Then we looked for "bad" items, employing the criteria described in the Exam Information page.

On average, 95% of students attempted any particular question (SD = 4%).  Applying the "2 SDs" criterion for an outlier, two item, #s 12 and 19, made the cut -- although two additional items (#s 8 and 25) met a relaxed criterion of lying more than 1 SD away from the mean.
The average score for those attempting each item was 1.43 (SD = 0.41).  Applying the "2 SDs" criterion, #25 was identified as a clear outlier; #16  met a relaxed criterion of lying more than 1 SD away from the mean, and #21 came awfully close..
No item met the criterion for being a "double outlier", even though #25 met the relaxed "1 SD" criterion.  

Better to err on the side of caution, so I dropped these 6 items by giving everyone 2 points for each of them.  

Still, it was clear that the test was too long.  The most appropriate way to compensate for test length is to give everyone additional points representing the difference between the nominal test length (25 items) and the average number of test items actually answered by students (24.02) items.  Accordingly, I added 2 more points to everyone's score, and then truncated any revised scores above 50 to the maximum of 50 points. 

MT_S11.jpg (29617 bytes)This resulted in a revised average score of 42.95 (SD = 5.20), or 86%, amounting to a solid B on average -- pretty good performance.  The median score was even better, 44 -- a solid B+.



I've had to adjust for length in past years, so why did I keep the test at 25 items this year?   First, there are two kinds of tests, speed tests, which have time limits, and power tests, which don't.  Midterm exams are almost by definition speed tests, because they have to be completed in 50 minutes, and it's fairly common for some students not to finish them in the allotted time.  Final exams are closer to power tests, because the standard is to give students three hours to complete a two-hour exam.  If we had had 80 minutes (which we didn't), I'm sure everybody would have finished early.  

In the past, I have been unhappy with shorter test formats, because they didn't allow for comprehensiveness of coverage.  With the present format, there was at least one question drawn explicitly from each lecture, and at least one question drawn explicitly from each reading, so what the psychometricians call "content validity" was enhanced.  I  was prepared to learn that even 25 questions were too many, but in the end this is an empirical question, and there was no way to find out except to try.  

Still, fewer than 25 questions begins to compromise content validity, so I intend to stick with the 25-question format, allowing for limited choice, in the future.  More details on how that will work out in the future as we get closer to the final exam -- which, because of that extra hour, will give everyone plenty of time to complete both the cumulative and noncumulative portions of the exam.

In the item-analysis below, I provide the percentage of students attempting each question, and the average score they obtained.  

Exams will be returned in discussion section March 31-April 1.  Students will have until 5:00 PM on Monday, April 4 to request a regrading.  The request for regrading must be in writing, and must provide a written defense of the answer at issue (just one or two sentences will do).   Please don't simply request regrading in the hope that you will get lucky: the GSIs have been strictly instructed to read each answer "fresh", and it is possible that they'll like your answer even less the second time around.  Exams not written in ink are not eligible for regrading.


Answer each the following 25 questions. Each question is worth 2 points, so that the exam totals 50 points.

Do not provide long-winded answers. You have less than 2 minutes, on average, for each question, and we are grading accordingly. Use only the space provided for your answer. Just one or two sentences will do -- three at the most. Get right to the point.

Write your name at the top of every page, so that the exams can be separated for grading.

Write your answers in ink. Answers written in pencil will not be eligible for regarding.

And write legibly, or we won't be able to appreciate how wonderful your answers are. Complete sentences are always nice, but they're not absolutely necessary.

Also, please indicate your Discussion Section # (or time, or GSI) here:_____________



Section 1: Introduction.

1. Why were the behaviorists dissatisfied with the definition of psychology as a study of consciousness?

 99% of the class attempted this item; mean score = 1.9 (a good start!).  Science is objective, while consciousness is necessarily subjective. Science depends on public observation, while consciousness is necessarily private. So, they redefined psychology as the analysis of behavior, which can be publicly, and objectively, observed. [Lecture 1]



Part 2: Introspection

2. What did James mean when he said that "human thought appears to deal with objects independent of itself"?

98% attempted; mean = 1.85.  James meant that we're not just consciousness -- we're conscious of things. Mental states don't exist in the abstract, they're always about something -- we're perceiving or remembering, thinking or imagining, feeling or desiring something. It's what Brentano meant when he said that "Intentionality is the mark of the mental": Mental states represent other things. [Lecture 2]


3. How did the Structuralists propose to analyze consciousness scientifically?

97%; 1.80.  The Structuralists viewed psychology as a sort of "mental chemistry" in which conscious experiences, analogous to molecules, could be analyzed into their constituent elements, analogous to atoms. By asking observers to introspect on their mental states, and varying the stimuli presented to them, they hoped to uncover the fundamental structure of conscious mental life. [Lecture 3]


4. How does reductive materialism differ from eliminative materialism?

98%; 1.65.  Eliminative materialism holds that consciousness doesn't actually exist: it's a sort of illusion, and a genuine psychology of mind would eliminate it in favor of a description of "objectively real" brain processes. Reductive materialism views consciousness really exists, but it is to brain-processes, so that subjective, private consciousness can be reduced to objectively, publicly observable brain-processes. [Revonsuo, Chapter 1]


5. How did the Structuralists' research program depend on the psychophysical principle?

95%, 1.63.  The psychophysical principle states that every psychological quality of a sensory experience is related to a physical property of the corresponding stimulus. Therefore, the Structuralists assumed that changes in sensory stimuli would produce corresponding changes in sensory experience, permitting them to map the physical dimensions of consciousness. [Revonsuo, Chapter 2]


6. Describe two problems with relying on introspective reports as data for a science of consciousness.

99%, 1.96.  Any two will do. (1) We can only report on an experience we've already had ("All introspection is retrospection"), so our report may be affected by forgetting. (2) Because memories are reconstructions, the reconstructed experience may differ from the experience as it originally occurred. (3) Introspective reports may be contaminated by confabulations, as the observer fills in the gaps of a poorly remembered experience. (4) Qualia are ineffable, so it may be very difficult for even the most sensitive and articulate observers to describe their experience accurately. (5) The very act of introspecting and reporting may create experiences that would not occur otherwise, so that the introspective report is a distortion of the person's actual subjective experience. (6) The "demand characteristics" of the experimental situation may lead observers to report what the experimenter wants to hear, rather than their actual experience. (7) Observers may censor their experiences, so as not to embarrass themselves. (8) Because conscious experiences are private, there is no way to assess the accuracy of introspections. [Revonsuo, Chapter 3]



Part 3: Mind and Body

7. What is "Descartes' impasse"?

99%, 1.73.  Descartes argued that mind and body were composed of two qualitatively different substances. But he also knew that mind and body interacted, and he couldn't figure out how an immaterial substance could have a causal effect on a material one. [Lecture 4]


8. Which version of identity theory is assumed by cognitive neuroscience?

90%, 1.31.  A "bad" item.  Under type identity theory, every instance of a particular mental state corresponds to the same brain state. Thus, it makes sense to try to identify the particular neural correlate of a particular conscious state. But under token identity theory, there is no specific neural correlate of any particular mental state, and no point in trying to identify one. [Lecture 5]


9. How do placebo effects provide evidence for psychosomatic interactions?

99%, 1.83.  In placebo effects the patient's belief about a treatment affects some aspect of physiological functioning. But because every mental state (such as a belief) corresponds with some brain state (its neural correlate), the most convincing psychosomatic effects occur outside the central nervous system, as in ulcers. [Lecture 6]


10. How are parapsychological phenomena such as telepathy and clairvoyance relevant to consciousness?

98%, 1.72.  These and similar phenomena are claimed to represent instances where information was transmitted from one mind to another in the absence of a physical medium. The existence of such phenomena thus implies that mind can exist in the absence of, or independently of, a body. [Lecture 7]


11. In searching for the neural correlates of consciousness, what is the advantage of techniques such as fMRI and PET over EEG? And what is the advantage of fMRI over PET?

96%, 1.82.  EEG has relatively low spatial resolution, so it can offer only a gross specification of the brain area activated during some mental activity. Compared to PET, fMRI has relatively high temporal resolution, meaning that it is sensitive to changes in brain activity that occur over a relatively short interval of time. [Revonsuo, Chapter 7]


12. What is binocular rivalry and what does it tell us about consciousness?

83%, 1.44.  A "bad" item.  When two incompatible visual stimuli are presented to each eye, the subject experiences first one visual percept, and then the other. Thus, we can record changes in brain activity that occur when one percept appears, and then disappears, from consciousness. [Revonsuo, Chapter 9]


13. What is Searle's primary objection to Dennett's "multiple drafts" theory of consciousness?

94%, 1.72.  Dennett, in Searle's view, essentially denies the reality of conscious mental states; they constitute some sort of illusion. All that happens is that the brain (or some other information-processing device) receives stimulus inputs and generates response outputs, with a number of "discriminative states" inbetween. All of this can be given an objective, third-person description, but the essence of consciousness is its first-person subjectivity. [Searle, Chapter5]


14. Why does Chalmers think that a thermostat and the Milky Way galaxy might be conscious, each in its own way?

97%, 1.30.  Chalmers asserts that subjective experience is a property of information, and so every physical system that embodies information must have subjective experience. In Nagel's terms, there's "something it's like" to be a thermostat. [Searle]



Part 4: Attention and Automaticity

15. Marcel showed that masked (unconscious) presentation of a word like doctor primed lexical decisions about semantically related words like nurse. What does this tell us about the location of the attentional filter (assuming that there is one)?

97%, 1.30.  Early-selection theories of attention assumed that attentional selection occurred early in information-processing, based on the physical features of the stimulus, and that semantic processing could occur only after attention was directed to the stimulus. But masked semantic priming obviously requires semantic analysis of a prime toward which attention can't be directed, because it's invisible. Therefore, it seems like the attentional filter must be located later in information-processing, after the stimulus has received at least some degree of semantic analysis. [Lecture 8]


16. How did the concept of automaticity attempt to resolve the debate between early and late selection?

97%, 0.94.  A "bad" item.  Automatic processes occur in the absence of conscious attention, because they consume little or no attentional resources. In principle, any task can be performed preattentively, so long as it has been automatized through extensive practice. Therefore, it is possible for semantic processing to occur preattentively, apparently contradicting early-selection theory. [Lecture 9]


17. What is the argument against "the automaticity of everyday life"?

93%, 1.14.  Some theorists assert that most social behavior is automatically triggered by the environment, and results from preattentive or preconscious processing. But all they have demonstrated is that automatic processing plays some role in social interaction. Studies using Jacoby's "method of opposition" suggest, to the contrary, that even complex social behavior reflects a mix of conscious, controlled and automatic, unconscious processing. [Lecture 10]


18. What is Chalmers' "hard problem" of consciousness, and what makes it seem so hard?

97%, 1.46.  The "hard problem" is explaining how a physical system could produce conscious mental states. It's relatively easy to figure out what the neural correlates of conscious mental states are, but it just not easy to state how those neural states produce mental states, or why they should do so [Revonsuo, Chapter 10]


19. What is the "40 hertz hypothesis" of Crick and Koch, and why is it not quite right?

86%, 1.38.  A "bad" item.  Crick and Koch proposed that people become conscious of an object when the ensemble of neurons representing features of that object fire synchronously at a frequency of 40 cycles per second. But it turns out that 40-hz oscillations are neither necessary nor sufficient for conscious awareness. They're apparently critical to binding the features of an object together, but something else is needed for conscious awareness. [Revonsuo, Chapter 11]



Part 5: The Explicit and the Implicit

20. When does priming provide evidence of unconscious memory?

98%, 1.30.  Priming provides evidence of implicit memory when the prime occurred sometime in the past (i.e., after the person's attention has been directed away from the prime. But priming alone is not sufficient to show that the memory is truly unconscious. Priming must occur in the absence of conscious recollection of the prime, as evidenced by performance on tests of free or cued recall, or recognition; or at least independent of conscious recollection. [Lecture 11]


21. What is the major difference between implicit perception and implicit memory?

98%, 1.07.  A "bad" item.  In implicit memory, the subject was consciously aware of the prime when it was presented. In implicit perception, the subject was not consciously aware of the prime when it was presented -- because it was subliminal, or masked, or otherwise unattended. [Lecture 12]


22. What are some of the limits on "subliminal" perception?

98%, 1.32.  Any two will do. (1) The prime cannot be too brief, in terms of the "stimulus-onset asynchrony between the prime and the mask (it is not necessary to use the precise term "SOA"; the concept will suffice). (2) The interval (SOA) between prime and target cannot be too long. (3) Although semantic processing of a subliminal prime is possible, there appear to be limits to how much meaning analysis can be performed preconsciously. (4) The limits mentioned in (3) may be loosened, if the semantic processing in question has been automatized through extensive practice. [Lecture 13]


23. In what sense can emotional responses be unconscious?

98%, 1.20.  Any of the following will do. (1) Emotional responses can be triggered automatically, even though that the person is unaware of the triggering stimulus, as a sort of priming effect. In this way, a conscious emotional response can be a manifestation of implicit memory or implicit perception. (2) Emotional responses can be triggered automatically by a consciously perceived (or consciously remembered) stimulus, but the person may have no conscious control over the emotion itself. In this case, the process leading to the emotion is inaccessible to conscious control. (3) A state of desynchrony among the various components of an emotional response can lead a person to respond to a stimulus with the physiological or motor component of emotion (or both), in the absence of the subjective feeling state. In this case, the emotional state itself would be unconscious, and the physiological or motor responses would be manifestations of implicit emotion, analogous to implicit memory. [Lecture 14]


24. Consider the neurological syndrome of neglect: what do neglect patients neglect, and where is the lesion typically found?

94%, 1.13.  Neglect patients appear to be unaware of that portion of space contralateral (i.e., on the opposite side of) to their lesion; and because the lesion is typically in the right posterior parietal lobe, they typically ignore the left portion of space, and any objects, or features of objects, that are located there. [Revonsuo, Chapter 4]


25. What is the "disconnection model" of explicit/implicit dissociations?

88%, 0.07.  A "pretty bad" item.  The disconnection model postulates that various features of an event are processed by specialized modules (such as one for face recognition), whose outputs are normally passed to a single module that mediates conscious awareness -- like Schacter's "conscious Awareness System" (CAS).  When the connection between a processing module and the consciousness module is broken, the person is unaware of the feature in question. But the processing module can still influence behavior indirectly, outside of conscious awareness, as in priming effects. [Revonsuo, Chapter 5]


The scoring guide used to grade the exam

will be posted to the course website as soon as possible after the exam.


Exam grades will be posted on the course website

as soon as possible after March 21.


Exams will be returned in sections after Spring Break.