In the third part of my research program, I speak to ongoing debates using innovative methodologies and help make sense of data irregularities with novel use of statistical procedures. In particular, legions of past studies have examined and compared effects of political values (i.e., ideology) and identity (i.e., party identity) on support for political candidates. Yet, scholars continue to debate about whether party identity or ideology is more fundamental, powerful, meaningful, or relevant. Part of the reason that the debate is yet to be resolved is that party identity and ideology are theoretically and empirically related, and they can dynamically shape each other. As a result, disentangling and comparing the effect of party identity and ideology is challenging. 
            However, another problem that contributes to this challenge, but is often overlooked by researchers is an omitted variable problem--past research often fails to consider the party identity and ideology of citizens as well as the party affiliation and ideology of candidates  simultaneously. To counter this, I introduce the “multiple matching perspective,” which simultaneously considers all possible interactive effects of partisan and ideological variance in perceivers and targets (5). In the case of ideology and party identity, four types of matches need to be considered and compared together under this rubric: (a) citizen ideology-candidate ideology match, (b) citizen party-candidate ideology cross-match, (c) citizen ideology-candidate party cross-match, and (d) citizen party-candidate party match. Using this method, I have uncovered nuanced and previously overlooked patterns of effects of ideology and party identity (5). To the field in general, the multiple matching perspective will allow researchers to uncover previously overlooked patterns and phenomena by disentangling effects of related predictors that reside in perceivers as well as targets, helping arbitrate the predictors’ contributions with attenuated omitted variable bias.
            Finally, as a data scientist, I often face challenges with respect to data quality and unexpected response patterns. Noticing an unusual pattern of low data quality from Mturk in 2018, I investigated and revealed a Mturk data contamination issue with novel usage of conventional statistical tests (i.e., using Cronbach’s alpha to gauge data quality of a given sample) (13). Working with other scholars, I also developed a R-based program to help researchers to identify responses from questionable sources (14). Thus, besides being a researcher with substantive interest in social and political psychology, I am also a methodologist who will continue to identify new methods and tools to help researchers answer their research questions.


​
References
​
5. Bai, H. The Multiple Matching Perspective on Value versus Identity: Investigating How Political 
Ideology and Party Identity Contribute to Citizens’ Support for Political Candidates 
(under review) Pre-print: 
https://psyarxiv.com/t7jh9/

13. Bai, H, Evidence that a large amount of low quality responses on MTurk can be detected with 
repeated GPS coordinates. https://bit.ly/35z56gh
14. Prims, J.P., Sisso, I, Bai, H (2018). Suspicious IP online flagging tool. 
https://itaysisso. shinyapps. io/Bots