College of Behavioral & Social Sciences
Permanent URI for this communityhttp://hdl.handle.net/1903/8
The collections in this community comprise faculty research works, as well as graduate theses and dissertations..
Browse
6 results
Search Results
Item Information Uncertainty Influences Learning Strategy from Sequentially Delayed Rewards(2023) Maulhardt, Sean Richard; Charpentier, Caroline; Psychology; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The problem of temporal credit assignment has long been posed as a nontrivial obstacle to identifying signal from data. However, human solutions in complex environments, involving repeated and intervening decisions, as well as uncertainty in reward timing, remain elusive. To this end, our task manipulated uncertainty via the amount of information given in their feedback stage. Using computational modeling, two learning strategies were developed that differentiated participants’ updates of sequentially delayed rewards: eligibility trace whereby previously selected actions are updated as a function of the temporal sequence - and tabular update - whereby additional feedback information is used to only update systematically-related rather than randomly related past actions. In both models, values were discounted over time with an exponential decay. We hypothesized that higher uncertainty would be associated with (i) a switch from tabular to eligibility strategy and (ii) higher rates of discounting. Participants’ data (N = 142) confirmed our first hypothesis, additionally revealing an effect of the starting condition. However, our discounting hypothesis had only weak evidence of an effect and remains an open question for future studies. We explore potential explanations for these effects and possibilities of future directions, models, and designs.Item Equilibrium Models with Dynamic Demand and Dynamic Supply(2019) Hui, Shen; Sweeting, Andrew T; Economics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This dissertation comprises two studies of equilibrium models with both dynamic demand and dynamic supply sides. The first is an empirical study of the US video games industry, and the second is a theoretical study. Chapters 1 and 2 develop a model for quantifying the role of social learning in consumers’ dynamic demand and finding optimal intertemporal prices for profit maximizing firms in a market populated by forward-looking social learners. Optimal prices are a result of a Markov perfect equilibrium played between the firm and the consumers. Nested in the market equilibrium is a demand equilibrium played among consumers who make the “right” purchase/wait decisions given endogenously produced product information. The empirical exercises are conducted in two steps. The first step estimates demand parameters, including those associated with social learning. Endogeneity of prices is remedied with a pseudo pricing policy function of relevant state variables. In the second step, optimal prices are found by the Mathematical Programming with Equilibrium Constraints (MPEC) approach. The model is applied to the US video games industry with sales data of PlayStation 3 games. The results reveal that (1) compared to static social learning, forward- looking social learning reduces equilibrium profits of games in the sample by $5.2M (28.4%) on average; (2) an incorrect belief of consumers’ forward-looking behavior reduces a firm’s profits by a maximum of 29.92%. These results indicate great value for researches on consumers’ forward-looking social learning behavior. In chapter 3 we study the effect of adding strategic buyers to the computational model of dynamic price competition when sellers experience learning-by-doing and organizational forgetting developed by Besanko et al. (2010) (BDKS). The addition is motivated by the presence of repeat buyers in many industries where learning- by-doing has been documented, and the role that the assumption of a monopsony strategic buyer has played in the theoretical literature. We characterize the degree of strategic buyer behavior using a single parameter, and show that even quite limited strategic behavior changes the equilibrium correspondance by almost entirely eliminating the multiplicity of equilibria emphasized by BDKS, and ensuring that no seller is likely to dominate the industry in the long-run. We examine how the welfare of both buyers and sellers varies with the degree of strategic buyer behavior.Item The benefits of testing: Individual differences based on student factors(2017) Robey, Alison Marie; Dougherty, Michael R; Psychology; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The testing effect, the notion that retrieval practice compared to restudying information leads to greater and longer retention, is one of the most robust findings in cognitive science. However, not all learners experience a benefit from retrieval practice. Many manipulations that influence the benefits of the testing effect have been explored, however, there is still much to learn about potential individual differences in the benefits of retrieval practice over restudy. As the testing effect grows in popularity and increasing numbers of classrooms begin implementing retrieval practice, it is essential to understanding how students’ individual differences and cognitive abilities contribute to the effect. For my dissertation, I explore how students’ cognitive abilities, specifically, episodic memory, general fluid intelligence, and strategy use, relate to the benefit of retrieval practice. In Study 1, I developed a new measure to simultaneously capture two aspects of strategy use: variation in what strategies learners use and variation in how learners use strategies. In Study 2, I examine how these two types of strategy use, along with episodic memory and general fluid intelligence can be used to predict the magnitude of the testing effect. Converging evidence from multiple analyses suggests variation in how learners use strategies was the only individual difference to influence the benefit learners receive from retrieval practice. More specifically, learners who are less adaptive and flexible in their strategy use show a greater benefit than more skilled strategy users. These findings have implications both for improving existing theories of the mechanisms of the testing effect and for determining how to best incorporate retrieval practice into classroom settings.Item Social Emotional Memory and Negative Symptoms in Individuals with Schizophrenia(2015) Bradshaw, Kristen Renee; Blanchard, Jack J; Psychology; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The current study investigated social and nonsocial emotional memory in schizophrenia, the relation of impairments in the recall of positive stimuli to increased negative symptoms, and the mediating role of defeatist performance beliefs. Twenty-three individuals with schizophrenia and twenty-four healthy controls completed clinical symptom interviews, social and nonsocial laboratory emotional memory tasks, and a measure of dysfunctional attitudes. Results indicated that on a social affective learning task, in comparison to controls, the schizophrenia participants were impaired in their ability to exhibit minimal affective learning of positive pairings. Defeatist performance beliefs did not mediate the relation between recall of positive stimuli and experiential negative symptoms. These findings suggest that it is primarily in the social domain that we see deficits in emotional memory in schizophrenia, and that this impairment in the ability to learn positive social associations may be linked to decreased motivation to engage in social, vocational, and recreational activities.Item CONTRIBUTIONS OF AGENCY VS. NON-AGENCY TO SEQUENTIAL MEMORY IN 3-YEAR OLDS(2010) Shuck, Lauren Haumesser; Woodward, Amanda L.; Psychology; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Two studies explored the effect of agency on memory in 3-year-old children when learning a sequence in a picture-book format. Previous research has shown that with both adults and older children, the inclusion of agency in free verbal recall is a central theme. However, very young children are often thought to have poor memory for social events because of their verbal limitations. By using a form of deferred imitation, Study 1 explored social episodic memory in a non-verbal sequential reconstruction task. Children who saw an agent in the picture sequence reconstructed more steps than those that did not see an agent present in the picture-books. Study 2 expanded upon these results by investigating the extent to which agency is necessary in order to improve memory, and what properties of the Study 1 increased performance. In this study, participants who were presented with an agent in only the first and last picture of the sequence did not reconstruct more steps than those that did not see an agent present. Taken together, agency may increase memory for a sequence but only if ample amounts of agentive cues are present throughout.Item Heterogeneity and Input Reallocation(2006-08-07) Pinto, Eugenio; Haltiwanger, John; Shea, John; Economics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)In this dissertation, we analyze some patterns of aggregate job reallocation that are significantly determined by the coexistence of heterogeneous businesses in any industry. First, we argue that the interaction of non-strictly convex adjustment costs and learning about true efficiency can explain the significant growth of survivors in a cohort of entering firms. Using Portuguese data we find evidence that survivors are the main source of growth in the cohort's average size, and that their contribution varies across sectors. By simulation, we show that we need adjustment costs to match this evidence with a selection model of industry dynamics. In a calibration of the model, we find that proportional costs and the fixed exit cost are key parameters in matching the evidence, and that firms in manufacturing learn relatively less initially about their efficiency, and are subject to much larger adjustment costs than firms in services. Second, we analyze how does structural heterogeneity across classes of firms affects the cyclical behavior of aggregate job flows. We find that types of firms whose optimal employment is relatively more determined by aggregate shocks than by idiosyncratic shocks influence the dynamics of aggregate job flows by more than they affect average aggregate flows. In Portuguese data, we conclude that large and old firms tend to affect aggregate dynamics by more than their already large employment shares would suggest. This tends to make job reallocation less procyclical than otherwise, and affects aggregate behavior in some sectors. Finally, as a background for the empirical analysis that is used in this dissertation, we analyze basic facts about the business cycle and gross job flows in Portugal from 1986 to 2000. We conclude that gross job flows are large and react in predictable ways to the business cycle and that patterns of job reallocation vary widely across sectors and firm's age and size.