Thesis (M.A.)--Boston University

Previous studies across multiple domains (e.g. pain, negative film clips, and learning word lists) have established that the end of an experience is heavily weighted when making summary judgments. However, these studies have not typically involved the type of tasks that individuals complete in everyday life. Moreover, they generally focus on retrospective evaluations of an event rather than its immediate affective impact. We sought to leverage these findings and ask how the order in which people complete hard and easy tasks might have consequences for how they feel after they are finished. To test this, we first ran a pair of between-subjects studies where participants completed one hard and two easy tasks with minimal expectations about the nature and length of the experience. We systematically varied whether the hard task occurred first, second, or third in the sequence and measured affect before and after the set of tasks. Consistent with predictions generated from these prior studies, those who completed the most difficult task at the end of a sequence had a greater drop in affect than those who completed it earlier. Also, final task affect was significantly predicted by the difficulty and enjoyment of the final task in the sequences. Related to this, the affective experience of the tasks in isolation was very similar to sequences that end on those same tasks. Taken together, these findings suggest an end effect in our data. We next sought to replicate the observed order effects when participants had prior knowledge of how many tasks they would be completing. We saw a very similar pattern in this study as well, with participants who completed the most difficult task at the end of the sequences having the greatest drop in affect. We also replicated our end effects, and observed that knowledge of task number led to greater affect in all orders. Our final studies tried to answer the question of whether or not participants predict that completing the most difficult task at the end of a sequence will lead to worse affect than completing it earlier. Across two studies, we did not find that participants who read about the tasks predicted affective differences as a result of task order. We also did not see evidence of a clear end effect in these participants. However, when compared to those who completed the tasks, we did observe a general overestimation of negative affect across all orders, regardless of hard task position. Although it has not been shown for task sequences, this finding is consistent with literature on ‘affective forecasting,’ which suggests that people overestimate the magnitude of expected negative affect. Finally, we asked participants in all studies what order they would have preferred to complete the sequences in. The majority of all participants would have preferred to complete the hard task at the end of a sequence rather than earlier. This was despite the affective consequences that many of them experienced from recently completing it at the end of a sequence. However, those in the prediction groups who merely had the hardest task presented to them first showed a disproportionate preference to also complete it first. And those who only completed a single task would prefer to complete it first in a hypothetical sequence with two easy but unknown tasks. Thus, despite the affective consequences of task order, many people do not seem to select orders that may diminish negative affect following a sequence. However, these data also suggest that completing easy tasks at the end of a sequence can improve affect, and there may be scenarios where individuals make more adaptive choices.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

Problem. This study assessed the predictive validity of criterion-referenced tests in a military setting with cutoff scores set by the Angoff and conventional score-setting methods.;Procedure. Thirty-six instructors and thirty-six specialists assessed each test item for job relevance and the probability that a minimally competent person would answer each question correctly, resulting in a new test cutoff score. Intragroup variability and interrater reliability of judgments were calculated. Test predictive validity assessment compared classroom test scores, supervisory rating scores, and skill qualification test scores of 100 job performers based on the two score-setting methods. Sample sizes varied from 17 to 100. Behaviorally anchored rating scale was used to estimate soldier performance effectiveness. Hypotheses were tested using analysis of variance, a correlation procedure by Ebel, t-test, and Pearson Product-Moment correlation. Null was accepted or rejected at.05 level of significance.;Results. Findings follow: (1) intragroup variability and interrater reliability of judges' estimates were statistically significant; (2) strengths of correlation coefficients for classroom test scores (CTS) and supervisory rating scores (SRS) under Angoff method exceeded r values for scores under conventional method; (3) strength of correlation coefficient for CTS and skill qualification test (SQT) scores under conventional method exceeded r value for scores under Angoff method; (4) correlation coefficients for CTS and SRS were statistically significant for Angoff "accepts" but not for Angoff "rejects" in three of four job performance areas, and means of SRS of the two groups of job performers were significantly different; and (5) correlation coefficient for CTS and SQT scores was statistically significant for Angoff "accepts" but not for Angoff "rejects" and means of SQT scores of the two groups of job performers were significantly different.;Conclusions. The Angoff cutoff score-setting method provides an effective means for setting criterion-referenced test cutoff scores. The Angoff and present score setting methods yield significantly different test standards. The score derived by the empirical method is a better measure of minimum job requirements of an entry-level performer, thereby enhancing the predictive validity of the classroom test. Recommendations for future research are included.