QuestionI need help writing and forming my introduction Instructions There… I need help writing and forming my introductionInstructionsThere is an outline provided for the introduction.You will need 3 references (you can use the article on this page provided as one of the references). (APA)Your job is to ?? an introduction for the experiment you will take in this lesson. ( I chose experiment 2) Therefore, start by explaining associative memory. Then use your three references to talk about research in associative memory.Finally, end with the task you completed and the hypothesis about this task. You will need to include references in the introduction and include references ?.My 3 Reference links ( I can only provide the abstract if I find full pdf; i will post in comments)https://pubmed.ncbi.nlm.nih.gov/16982044/https://pubmed.ncbi.nlm.nih.gov/25813899/https://pubmed.ncbi.nlm.nih.gov/31586278/OutlineTitle of paper here Statement of problem: Define associative memory Define judgments of associative memory Review of literature One reference Two reference Third reference (what did they do? What did they find?) Hypotheses Trying to see if people could discriminate between high relationships and low relationships. Can you judge different associative relationships? What are we going to test? Defined associative memory Free association task Judgments of associative memory task What did we find? People suck. (not very good at differentiating)Research on psycinfo: Maki, William S. Koriat (Koriat and Bjork, ~04) “judgments of associative memory”ExperimentsSeparating Bias and Sensitivity in Judgments of Associative MemoryWilliam S. MakiTexas Tech UniversityExperiment 2: Effects of Numerical Limits and Perspective on RatingsIn previous associative judgment experiments, participants had no guidelines regarding the numbers assigned to cue-response pairs, so the summed ratings over three levels of associative strengths could total more than the numerical maximum for any single rating. The hypothetical perfectly calibrated participant should assign ratings of 10, 40, and 70 to pairs with those same FSGs, so the expected total would be 10 + 40 + 70 = 120. However, it was possible to assign maximum ratings of 100 to each pair thus producing a total rating of 300. Perhaps putting a cap on the total rating would eliminate the overestimation bias and expose a sharper slope (as in the left panel of Figure 2). A second question motivating this experiment targets the instructions used in Experiment 1 (these instructions were also used by Garskof & Forrester, 1966; Maki, in press). The standard instructions in Experiment 1 asked for ratings of what other college students would do in the way of producing responses to the cue word. Perhaps the ratings observed to date are inaccurate because participants have difficulty judging what other people do in free association tasks. On this line of thinking, then, engaging the self by asking the participants to put their own ratings on the cue-response pairs might sharpen the discrimination among associative strengths.MethodParticipants. Forty undergraduate students from the Texas Tech University Department of Psychology human subject pool participated and were compensated by course credit. Apparatus and materials. The laboratory and computer stations were the same as those in Experiment 1. The word pairs were the same as those used in Maki’s (in press) Experiment 7. Forty-eight cue words were selected from Nelson et al.’s (2004) norms such that each cue word had four response words. The average FSGs for the four response words were 0.379, 0.226, 0.110, and 0.047. The stimulus properties of the cue and response words are given in Table 2. The focus during the stimulus selection process was on finding cue words that had at least four response words that were separated by at least 0.05 FSG. Other properties of the words and their pairings were not controlled. The order of the 48 cue words was randomized separately for each participant. Within each trial, the order of the FSG of the response words was randomized. Procedure. The participants were tested in groups of six or fewer. All participants were read the same general background information about free association tests as in Experiment 1. The participants were then told that they would see a series of cue words each accompanied by four response words that were known associates of the cue words. The participants were instructed to give a numerical rating to each response word such that each rating was in the range of 0 – 100, and the total of the ratings was exactly 100. The computer program was written to implement these constraints, and its operation was demonstrated to the participants before data were collected.Two versions of the experiment were conducted, each with a specific set of instructions that was given to the participants. In the first version (N = 29), the participants heard (and later read) the standard JAM instructions emphasizing how “other” college students would rate the pairs of words. The second version of the experiment (N = 11) was run after the first version (but within the same month). These participants were given a different set of instructions emphasizing how the participant (“self”) would rate the pairs. Following the instructions, participants viewed 48 displays, one for each cue word. The program advanced to the next display only when all response words had been rated, and the total of the four ratings was equal to 100.Results and DiscussionFor the “other” instructions, the intercept of the JAM function was 18.5, and the slope was 0.34; for the “self” instructions, the intercept was 20.8, and the slope was 0.22. Both JAM functions were strongly linear, R2 s >0.96. Both slopes fell within the range of the slopes observed in previous experiments (Maki, in press, and the present Experiment 1). Thus, numerically constraining the ratings appears not to have improved discrimination among associative strengths. A 2 x 4 mixed analysis of variance performed on the average ratings for each participant included Group (self vs. other) and FSG categories as factors. The FSG categories differed significantly, F(3, 114) = 58.92, MSE = 9.01, and the FSG effect depended on Group, F(3, 114) = 2.88, MSE = 9.01. The latter, interactive effect resulted from the slope of the JAM function for the “self” group being somewhat lower than the slope for the “other” group. These effects were due to strongly linear (and significant) trends relating judged and normed associative strengths. The answers to the two questions motivating this experiment seem clear. Numerically constraining the magnitude of the ratings did not increase the discriminability among associative strengths. The discriminability was also not much improved by changing the perspective of the ratings (from what others would do to what the participant him- or herself would do).General DiscussionThe questions motivating the present experiments concerned the causes of the low level of discrimination among associative strengths indexed by the shallow slope and the apparent bias toward high ratings indexed by the high intercept of the JAM function (see Figure 1). Training on many pairs with errorcorrection feedback (Experiment 1) had no effect on the slope of the JAM function. However, those groups that received errorfeedback training had lower intercepts than those groups that did not receive the training. This lowered intercept was attributed to instructions that caused a more cautious strategy in producing the ratings. Thus, the bias toward high ratings was reduced without any effect on the sensitivity (as measured by the slope). The control of bias was more direct in Experiment 2. Participants rated four associates of each cue word. The numerical ratings were constrained to sum to 100. Although this manipulation reduced the average numerical ratings, the slope of the JAM function remained shallow. Thus, with regard to the central question of concern here, the effects of bias and sensitivity are separable so that the model represented in the left panel of Figure 2 is rejected. A related issue concerns the causes of the invariance of the JAM slope. As noted in the introduction to this article, the slope has resisted several attempts to increase it (Koriat et al., 2006; Maki, in press). The present results are additional instances of the immutability of the JAM slope. No manipulation to date has significantly sharpened the slope of the JAM function, a fact that leads to the conclusion that there is a theoretical limit on the discrimination among associative strengths. Maki (in press) showed that a multiple-trace memory model (MINERVA 2; Hintzman, 1988) could be adapted to produce free association probabilities and JAM functions that fit the available data quite well, exhibiting both shallow slopes and high intercepts. If it turns out that other models are capable of producing the JAM function (e.g., Nelson, McEvoy, & Dennis, 2000), then there would be additional support for the contention that the JAM function reflects some fundamental characteristics of our memory system.Social SciencePsychology PSY 394
