No description available.

Objective: The project aimed to compare marginal structural models, and propensity score adjusted models with Cox Proportional Hazards models to address confounding by indication due to time-dependent confounders. These methods were applied to data from approximately 120,000 women in the Women’s Health Initiative to evaluate the causal effect of antidepressant medication with respect to diabetes risk. Methods: Four approaches were compared. Three Cox Models were used. The first used baseline covariates. The second used time-varying antidepressant medication use, BMI and presence of elevated depressive symptoms and adjusted for other baseline covariates. The third used time-varying antidepressant medication use, BMI and presence of elevated depressive symptoms and adjusted for other baseline covariates and propensity to taking antidepressants at baseline. Our fourth method used a Marginal Structural Cox Model with Inverse Probability of Treatment Weighting that included time-varying antidepressant medication, BMI and presence of elevated depressive symptoms and adjusted for other baseline covariates. Results: All approaches showed an increase in diabetes risk for those taking antidepressants. Diabetes risk increased with adjustment for time-dependent confounding and results for these three approaches were similar. All models were statistically significant. Ninety-five percent confidence intervals overlapped for all approaches showing they were not different from one another. Conclusions: Our analyses did not find a difference between Cox Proportional Hazards Models and Marginal Structural Cox Models in the WHI cohorts. Estimates of the Inverse Probability of Treatment Weights were very close to 1 which explains why we observed similar results.

The purpose of this study was to examine the potential for using propensity score-based matching methods to estimate teacher contributions to student learning. Value-added models are increasingly used in teacher accountability systems in the United States in spite of ongoing qualms about the validity of teacher quality estimates resulting from those models. Using a large national dataset, teacher effects were estimated for 435 teachers using both value-added and propensity score-based approaches. The two approaches resulted in teacher effect estimates that were moderately correlated, with propensity score-based estimates more highly correlating with the value-added estimates as the matching ratio was increased. For many teachers' students, finding a set of matched control students was impossible unless the set of matching variables was reduced. Results suggest that many teachers have classroom compositions that are unusual, making evaluation of the teachers' impacts on student outcomes problematic. It was also found that, while value-added estimates were relatively insensitive to covariate inclusion choices or method of effect estimation, propensity score-based estimates were somewhat sensitive. Propensity score-based teacher effect estimates offer promise both for better accounting for classroom composition and student background variables and for indicating when a teacher's context is unique with respect to those variables, making the teacher's impact challenging to evaluate.

In this study, I examine the employment effects on average earnings and duration to work during a 45 month period, after receiving social assistance (SA) in October 2019, compared to receiving unemployment insurance (UI) the same month. A distinction is made between two treatment groups; receiving SA in addition to UI (treatment I) and receiving SA (treatment II). Using propensity score matching (PSM), I estimate the average treatment effects on the treated on earnings, as well as duration to work by using the Kaplan-Meier survival estimator with the matched observations. I use newly released Swedish administrative micro level data of individuals’ monthly labour market status (BAS) between 2020-2023, from Statistics Sweden. During this thesis process, where Statistics Sweden allowed me data access, I was allowed an additional year of data, for 2019. Results showed that the inflow of SA recipients in October 2019, on average had 25.5 percent lower earnings between November 2019-July 2023, compared to the inflow of UI recipients the same month. In addition, the inflow of SA recipients in October 2019, on average spent 4 months longer in unemployment, compared to those receiving UI the same month. However, results were insignificant when comparing effects between the inflow of those receiving SA in addition to UI in October 2019 with the inflow of UI recipients the same month. Results for this group were insignificant for both employment outcomes; average earnings and duration to work.

In this study, I examine the employment effects on average earnings and duration to work during a 45 month period, after receiving social assistance (SA) in October 2019, compared to receiving unemployment insurance (UI) the same month. A distinction is made between two treatment groups; receiving SA in addition to UI (treatment I) and receiving SA (treatment II). Using propensity score matching (PSM), I estimate the average treatment effects on the treated on earnings, as well as duration to work by using the Kaplan-Meier survival estimator with the matched observations. I use newly released Swedish administrative micro level data of individuals’ monthly labour market status (BAS) between 2020-2023, from Statistics Sweden. During this thesis process, where Statistics Sweden allowed me data access, I was allowed an additional year of data, for 2019. Results showed that the inflow of SA recipients in October 2019, on average had 25.5 percent lower earnings between November 2019-July 2023, compared to the inflow of UI recipients the same month. In addition, the inflow of SA recipients in October 2019, on average spent 4 months longer in unemployment, compared to those receiving UI the same month. However, results were insignificant when comparing effects between the inflow of those receiving SA in addition to UI in October 2019 with the inflow of UI recipients the same month. Results for this group were insignificant for both employment outcomes; average earnings and duration to work.

京都大学 / 新制・課程博士 / 博士(医学) / 甲第24288号 / 医博第4904号 / 新制||医||1061(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 山本 洋介, 教授 中本 裕士, 教授 森田 智視 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Heterogeneity in the school education quality plays an important role for those who want to pursue a bachelor's degree in Peru since access to higher education is highly correlated with socioeconomic status. In that sense, an intervention for disadvantaged students took place for the first time in 2022 and was constrained to the assessment of a scholarship called “Beca 18”, the biggest scholarship that the public institution called PRONABEC addresses every year since 2012. The intervention included additional tools for a group of applicants: (i) full-time online classes for 2 to 4 months; (ii) an electronic device with an internet connection; and (iii) the admission exam payment fully covered up to 2 times. The objective of this thesis is to evaluate the effectiveness of this intervention in increasing the likelihood of the treated to access higher education through the 2022 “Beca 18” scholarship process. As the treatment was not randomly assigned, a control group was estimated using the Propensity Score Matching methodology based on individual characteristics. Results showed that there is no statistically significant effect of the intervention in the treated applicants and invite to re-evaluate its design and implementation.

Bipartite experiments have recently emerged as a focal point in causal inference. In these experiments, treatment is administered to one set of units, while outcomes of interest are gauged on a distinct set of units. Such experiments are especially valuable in scenarios where pronounced interference effects transpire between units on a bipartite network. For instance, in market experiments, designating treatment at the seller level and assessing outcomes at the buyer level (or vice-versa) can lead to causal models that more accurately reflect the inherent interference between buyers and sellers. Although bipartite experiments can enhance the precision of causal effect estimations in specific contexts, it's imperative to conduct the analysis judiciously to avoid introducing undue bias through the network. Drawing from the generalized propensity score literature, we demonstrate that it's feasible to achieve unbiased estimates of causal effects for bipartite experiments, given a conventional set of assumptions. Furthermore, we delve into the formulation of confidence sets with accurate coverage probabilities. By employing a bipartite graph from a publicly accessible dataset previously explored in bipartite experiment studies, we illustrate, via simulations, a notable reduction in bias and augmented coverage. / Statistics

No description available.

No description available.