Page 1 of 7 ID# Exam 3 PS 217, Spring 2010 (You must use your real student ID#) This exam should contain no surprises (except, possibly, its brevity). Read each question carefully and answer it completely. Show all your work so that I might be able to give you partial credit. Finally, I want to thank you for a wonderful semester. I really enjoyed myself. I wish you all a peaceful summer, and I look forward to seeing many of you throughout the coming academic year. Take good care of yourself, and be gentle with those around you. 1. Bäuml, K.-H. T. and Samenieh, A. (2010) had an article published in Psychological Science titled The Two Faces of Memory Retrieval. They write: Does retrieval of a specific memory affect recollection of related memories? For instance, does selective remembering of part of a traumatic experience, or part of an incidentally observed crime, affect memory for other details of the event? Casual subjective experience suggests that memory retrieval can improve recollection of related memories. When one is talking with a friend about a common, long-forgotten vacation, remembering a first piece of the event often initiates a chain of retrieval processes, along which more and more of the seemingly forgotten memory is recollected. However, this subjective experience contrasts with scientific experiments that have demonstrated that selective remembering typically impairs recollection of related material. In their experiment, subjects studied a list of 16 words, one at a time, for 5 sec each. Of the 16 words, four were targets (each beginning with a unique letter, e.g., apple, carrot, mouse, table) and the remaining 12 words were nontargets (each beginning with different unique initial letters, e.g., banana, sandwich). After studying the list, half of the 80 total participants were told to remember the items in the list for a later test and half were told to forget the items in the list, because they weren’t needed for a subsequent test. The other factor was the number of nontarget items in an initial cued recall task. In each group of 40 participants (Remember or Forget), ten were first asked to recall 0 of the nontarget words before recalling the target words, ten were first asked to recall 4 of the nontarget words before recalling the target words, ten were first asked to recall 8 of the nontarget words before recalling the target words, and ten were asked to first recall all 12 of the nontarget words before recalling the target words. For the nontarget words, the participants were given a prompt of a couple of initial letters of the word (e.g., ba---- for banana and sa------ for sandwich). For the target words, the participants were given a prompt of the initial letter of the word (e.g., a---- for apple and m---- for mouse). The dependent variable was the percentage of the target words correctly recalled. Complete the SPSS table below and then interpret the results of the study as completely as you can. [15 pts]Page 2 of 7 Source Type III Sum of Squares df Mean Square F Sig. Partial Eta Squared Observed Powerb Task 577.8 .000 .438 1.000 Items 628.4 .000 .459 1.000 Task * Items 5360.2 .000 .879 1.000 Error 739.9 Corrected Total 7306.3Page 3 of 7 2. If the above ANOVA had been computed as a one-way ANOVA on Number of Nontarget Items Recalled (Items), how would the source table change (illustrate below)? [5 pts] Source SS df MS F Item Total 3. In the two-way ANOVA in Problem 1 and the one-way ANOVA in Problem 2, you should be able to estimate the population variance (σ2). What would be your best numerical estimate of the population variance in Problem 1? What would be your best numerical estimate of the population variance in Problem 2? Given that the data are identical in the two problems, why might your estimate differ? [5 pts] 4. Under which conditions might it make more sense to compute Spearman’s rho than Pearson’s r? [2 pts] 5. The full name is Pearson product-moment correlation coefficient. You can compute the statistic (r) as the product of two “moments.” Well, actually, the average (mean) of the product of two moments. What two values might you multiply, then take an average of the products in order to compute r? (Of course, you would never actually compute r this way.) [2 pts]Page 4 of 7 6. Helene Intraub (University of Delaware) has conducted lots of research on a phenomenon she calls boundary extension. In study after study, she and her students have shown that people will report having seen more of a picture than had actually been presented, as seen below: What people actually see at acquisition What people later report having seen This year, Ben Glicksberg conducted his senior thesis research with me. He examined possible limitations on the boundary extension effect. Essentially, Ben had three conditions: Condition First phase Second phase Test phase Control briefly look at each picture (as in typical Intraub) distractor task draw each picture Simultaneous Drawing while looking at each picture, construct a rough drawing distractor task draw each picture Post Viewing Drawing after looking at each picture, construct a rough drawing distractor task draw each picture As you might intuit, Ben reasoned that if people drew the picture either while it was being presented (Simultaneous) or immediately after seeing the picture (Post), then they might later exhibit less boundary extension than Intraub typically finds. And, of course, the Control condition should exhibit the typical boundary extension effect. Ben looked at his data in a number of different ways, but here’s one type of data he analyzed. The dependent variable was the percentage of the original picture represented in the final drawing. Thus, if the original object took up 68% of the original frame and a participant drew the object so that it took up 68% of the frame at test, then that person had a score of 1.0 and exhibited no boundary extension. However, if the participant drew the object so that it took up 34% of the frame at test, then that person had a score of .5 and exhibited boundary extension. (That is, as seen in the examples above, the object takes up a smaller proportion of the total picture than was true in the original picture.) Below is an SPSS output from Ben’s data. Complete the source table and then interpret his results as completely as you can. Do Ben’s results confirm