{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/132516"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/132516","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Essays on complexity and causality in political science research","abstract":"This is a dissertation consisting of three independent essays on the use of computational and causal inference methods in political science research. The first essay examines the validity of the synthetic control method under complex analytical settings. The synthetic control method is increasingly being used to estimate the effect of an intervention when units of interest are at the aggregate level and there does not exist suitable control units to serve as the counterfactual for the treated unit. However, the validity of the method has not been examined under complex settings where there could either be interaction effects among the covariates used to estimate the optimal synthetic control unit or where there could be a network cascading process happening within the treated unit at the same time as the intervention of interest. This paper shows that a number of commonly used synthetic control methods provide biased estimates of the causal effect of interest when there is only a small amount of pairwise interactions among the covariates and their coverage rates significantly decrease when there is an interaction between within treatment unit network cascading and the intervention of interest. As a result, this paper provides new insights about the validity of synthetic control methods under more complex analytical settings. The second essay examines the performance of bootstrap methods for quantifying the uncertainty of causal effect estimates from matching. In particular, it investigates the performance of three existing approaches as well as a newly developed one. The proposed non-parametric bootstrap method quantifies the uncertainty of average treatment effect estimate by first quantifying the uncertainty associated with the sample treatment group by bootstrapping the treatment group and then finding the counterpart control group by pair matching on estimated propensity score without replacement. Through Monte Carlo simulation and analysis of the Current Population Studies data set, the results show that all bootstrap standard error estimates can deviate from the true sample standard error if the data becomes in-permissible for matching. And most bootstrap methods provide satisfactory coverage rate of 95 percent or above but only when percent treated in the sample data is small enough such that sufficient number of control units are available to be matched to the treatment group units. The third essay makes use of a computational model to examine the interaction between partisan polarization and partisan segregation in the US in recent decades. Existing research that looks at mass partisan polarization and geography in conjunction focuses exclusively on the extent of partisan geographical sorting but neglects the substantial effect segregated geographical context can have on the formation of mass partisan polarization. This paper shows that while partisan geographical sorting can happen to a certain extent, it only happens in spaces where there is a certain level of partisan segregation to start with. In essence, geographical context's effect on partisan polarization is more substantial than partisanship' effect on geographical sorting, a surprising result given current literature's focus on the latter rather than the former.","abstract_html":"This is a dissertation consisting of three independent essays on the use of computational and causal inference methods in political science research. The first essay examines the validity of the synthetic control method under complex analytical settings. The synthetic control method is increasingly being used to estimate the effect of an intervention when units of interest are at the aggregate level and there does not exist suitable control units to serve as the counterfactual for the treated unit. However, the validity of the method has not been examined under complex settings where there could either be interaction effects among the covariates used to estimate the optimal synthetic control unit or where there could be a network cascading process happening within the treated unit at the same time as the intervention of interest. This paper shows that a number of commonly used synthetic control methods provide biased estimates of the causal effect of interest when there is only a small amount of pairwise interactions among the covariates and their coverage rates significantly decrease when there is an interaction between within treatment unit network cascading and the intervention of interest. As a result, this paper provides new insights about the validity of synthetic control methods under more complex analytical settings. The second essay examines the performance of bootstrap methods for quantifying the uncertainty of causal effect estimates from matching. In particular, it investigates the performance of three existing approaches as well as a newly developed one. The proposed non-parametric bootstrap method quantifies the uncertainty of average treatment effect estimate by first quantifying the uncertainty associated with the sample treatment group by bootstrapping the treatment group and then finding the counterpart control group by pair matching on estimated propensity score without replacement. Through Monte Carlo simulation and analysis of the Current Population Studies data set, the results show that all bootstrap standard error estimates can deviate from the true sample standard error if the data becomes in-permissible for matching. And most bootstrap methods provide satisfactory coverage rate of 95 percent or above but only when percent treated in the sample data is small enough such that sufficient number of control units are available to be matched to the treatment group units. The third essay makes use of a computational model to examine the interaction between partisan polarization and partisan segregation in the US in recent decades. Existing research that looks at mass partisan polarization and geography in conjunction focuses exclusively on the extent of partisan geographical sorting but neglects the substantial effect segregated geographical context can have on the formation of mass partisan polarization. This paper shows that while partisan geographical sorting can happen to a certain extent, it only happens in spaces where there is a certain level of partisan segregation to start with. In essence, geographical context's effect on partisan polarization is more substantial than partisanship' effect on geographical sorting, a surprising result given current literature's focus on the latter rather than the former.","abstract_has_math":false,"creators":["Li, Jing"],"institution":"University of Illinois Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Political Science","degree_department":null,"school":null,"contributors":["Tam Cho, Wendy K","Gaines, Brian","Winters, Matthew","Offer-westort, Molly"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-12","date_published":"2025-12","updated_at":"2026-07-22T22:25:07Z","subjects":["synthetic control method","network cascading","covariate interaction","non-parametric bootstrap","matching","average treatment effect for the treated","computational model","partisan ideological polarization","partisan geographical segregation","intergroup contact"],"languages":["en"],"rights":["Copyright 2025 Jing Li"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2142/132516","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Tam Cho, Wendy K","Gaines, Brian","Winters, Matthew","Offer-westort, Molly"]},{"key":"dc:creator","label":"Author","values":["Li, Jing"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025-12","2025-11-21"]},{"key":"dc:type","label":"Dc Type","values":["text","Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Political Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["synthetic control method","network cascading","covariate interaction","non-parametric bootstrap","matching","average treatment effect for the treated","computational model","partisan ideological polarization","partisan geographical segregation","intergroup contact"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2025 Jing Li"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/2142/132516"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["This is a dissertation consisting of three independent essays on the use of computational and causal inference methods in political science research. The first essay examines the validity of the synthetic control method under complex analytical settings. The synthetic control method is increasingly being used to estimate the effect of an intervention when units of interest are at the aggregate level and there does not exist suitable control units to serve as the counterfactual for the treated unit. However, the validity of the method has not been examined under complex settings where there could either be interaction effects among the covariates used to estimate the optimal synthetic control unit or where there could be a network cascading process happening within the treated unit at the same time as the intervention of interest. This paper shows that a number of commonly used synthetic control methods provide biased estimates of the causal effect of interest when there is only a small amount of pairwise interactions among the covariates and their coverage rates significantly decrease when there is an interaction between within treatment unit network cascading and the intervention of interest. As a result, this paper provides new insights about the validity of synthetic control methods under more complex analytical settings. The second essay examines the performance of bootstrap methods for quantifying the uncertainty of causal effect estimates from matching. In particular, it investigates the performance of three existing approaches as well as a newly developed one. The proposed non-parametric bootstrap method quantifies the uncertainty of average treatment effect estimate by first quantifying the uncertainty associated with the sample treatment group by bootstrapping the treatment group and then finding the counterpart control group by pair matching on estimated propensity score without replacement. Through Monte Carlo simulation and analysis of the Current Population Studies data set, the results show that all bootstrap standard error estimates can deviate from the true sample standard error if the data becomes in-permissible for matching. And most bootstrap methods provide satisfactory coverage rate of 95 percent or above but only when percent treated in the sample data is small enough such that sufficient number of control units are available to be matched to the treatment group units. The third essay makes use of a computational model to examine the interaction between partisan polarization and partisan segregation in the US in recent decades. Existing research that looks at mass partisan polarization and geography in conjunction focuses exclusively on the extent of partisan geographical sorting but neglects the substantial effect segregated geographical context can have on the formation of mass partisan polarization. This paper shows that while partisan geographical sorting can happen to a certain extent, it only happens in spaces where there is a certain level of partisan segregation to start with. In essence, geographical context's effect on partisan polarization is more substantial than partisanship' effect on geographical sorting, a surprising result given current literature's focus on the latter rather than the former.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2026-02-19 without embargo terms","The student, Jing Li, accepted the attached license on 2025-11-20 at 16:52.","The student, Jing Li, submitted this Dissertation for approval on 2025-11-20 at 17:22.","This Dissertation was approved for publication on 2025-11-21 at 18:20.","DSpace SAF Submission Ingestion Package generated from Vireo submission #22909 on 2026-02-19 at 18:25:09"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Essays on complexity and causality in political science research"]}]}],"canonical_facts":{"dc:contributor":["Tam Cho, Wendy K","Gaines, Brian","Winters, Matthew","Offer-westort, Molly"],"dc:creator":["Li, Jing"],"dc:date":["2025-12","2025-11-21"],"dc:description":["This is a dissertation consisting of three independent essays on the use of computational and causal inference methods in political science research. The first essay examines the validity of the synthetic control method under complex analytical settings. The synthetic control method is increasingly being used to estimate the effect of an intervention when units of interest are at the aggregate level and there does not exist suitable control units to serve as the counterfactual for the treated unit. However, the validity of the method has not been examined under complex settings where there could either be interaction effects among the covariates used to estimate the optimal synthetic control unit or where there could be a network cascading process happening within the treated unit at the same time as the intervention of interest. This paper shows that a number of commonly used synthetic control methods provide biased estimates of the causal effect of interest when there is only a small amount of pairwise interactions among the covariates and their coverage rates significantly decrease when there is an interaction between within treatment unit network cascading and the intervention of interest. As a result, this paper provides new insights about the validity of synthetic control methods under more complex analytical settings. The second essay examines the performance of bootstrap methods for quantifying the uncertainty of causal effect estimates from matching. In particular, it investigates the performance of three existing approaches as well as a newly developed one. The proposed non-parametric bootstrap method quantifies the uncertainty of average treatment effect estimate by first quantifying the uncertainty associated with the sample treatment group by bootstrapping the treatment group and then finding the counterpart control group by pair matching on estimated propensity score without replacement. Through Monte Carlo simulation and analysis of the Current Population Studies data set, the results show that all bootstrap standard error estimates can deviate from the true sample standard error if the data becomes in-permissible for matching. And most bootstrap methods provide satisfactory coverage rate of 95 percent or above but only when percent treated in the sample data is small enough such that sufficient number of control units are available to be matched to the treatment group units. The third essay makes use of a computational model to examine the interaction between partisan polarization and partisan segregation in the US in recent decades. Existing research that looks at mass partisan polarization and geography in conjunction focuses exclusively on the extent of partisan geographical sorting but neglects the substantial effect segregated geographical context can have on the formation of mass partisan polarization. This paper shows that while partisan geographical sorting can happen to a certain extent, it only happens in spaces where there is a certain level of partisan segregation to start with. In essence, geographical context's effect on partisan polarization is more substantial than partisanship' effect on geographical sorting, a surprising result given current literature's focus on the latter rather than the former.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2026-02-19 without embargo terms","The student, Jing Li, accepted the attached license on 2025-11-20 at 16:52.","The student, Jing Li, submitted this Dissertation for approval on 2025-11-20 at 17:22.","This Dissertation was approved for publication on 2025-11-21 at 18:20.","DSpace SAF Submission Ingestion Package generated from Vireo submission #22909 on 2026-02-19 at 18:25:09"],"dc:format":["application/pdf"],"dc:identifier":["https://hdl.handle.net/2142/132516"],"dc:language":["en"],"dc:rights":["Copyright 2025 Jing Li"],"dc:subject":["synthetic control method","network cascading","covariate interaction","non-parametric bootstrap","matching","average treatment effect for the treated","computational model","partisan ideological polarization","partisan geographical segregation","intergroup contact"],"dc:title":["Essays on complexity and causality in political science research"],"dc:type":["text","Thesis"],"thesis:degree_discipline":["Political Science"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:07Z"}