Offizielle Kursbeschreibung
[b]Experimental vs. observational research in political science[/b]
Seminar at the Institute of Political Science, TU Darmstadt
Bernhard Clemm, Postdoc (University of Amsterdam), [url=mailto:b.f.d.clemm@uva.eu]b.f.d.clemm@uva.eu[/url]
In this seminar, students will learn about the logic of experimental and observational research They will get to know key contributions from both traditions to political science, with a focus on political behaviour, political psychology and public opinion. The seminar language is English, but I am happy to help out in German where necessary.
Even though often considered the "gold standard" to test scientific causal theories, experiments have long been a stranger to political science. In his 1909 presidential address to the American Political Science Association, A. Lawrence Lowell pointed to the "impossibility of experiment" and concluded: "Politics is an observational, not an experimental science..." (Lowell 1910, 7). Only towards the end of the 20th century, scholars began to acknowledge the benefits of random assignment to test causal political theories. Today, observational and experimental research are equally important for political science, each with their unique strengths and weaknesses and areas of application. A wide range of topics, in particular in the fields of political behaviour, public opinion and political psychology, have profited from both traditions, e.g. media effects and framing (Iyengar & Kinder 1987, Druckmann 2004), attitude change (Taber & Lodge 2006), turnout and voting (Gerber & Green 2000; Lodge et al. 1989) and political and social trust (Kugler et al. 2008).
The sessions covering methods will start with a quick history of observation and experiments in the social science in general. The concept of causality, and how it can be shown in both approaches, will be discussed. Basic statistical methods such as group comparisons and regression will illustrated using examples in R (no prior experience required). The issue of internal and external validity as well as analysis of heterogeneity will be discussed. The final sessions will revolve about the importance of ethics, benefits and drawbacks of incentivization and practical implementation of experiments. The use of online apps I have programmed illustrate key statistical ideas efficiently in class.
The topics sessions will be initiated by general topic introductions by the teacher, followed up by student presentations on more particular topics in the literature. For example, a presentation topic could be selective exposure to political news. Students would not only discuss the most important studies (e.g. Iyengar & Hahn 2009, Frimer et al. 2017), and illustrate in what way experiments are advantageous over (or coming short of) observational studies (e.g. Stroud 2008).
In addition, I will run several 5-min in-class studies at the beginning of sessions that will illustrate how state-of-the-art political science works in practice. Students can easily participate in these studies on their computer. I will analyze the these during sessions so that students learn the entire process of conducting until presenting research. Some of these studies can also be presented by students in later sessions instead of a literature presentation.
A full course outline, as well as signup for presentations will follow.
Online-Angebote
moodle
[b]Experimental vs. observational research in political science[/b]
Seminar at the Institute of Political Science, TU Darmstadt
Bernhard Clemm, Postdoc (University of Amsterdam), [url=mailto:b.f.d.clemm@uva.eu]b.f.d.clemm@uva.eu[/url]
In this seminar, students will learn about the logic of experimental and observational research They will get to know key contributions from both traditions to political science, with a focus on political behaviour, political psychology and public opinion. The seminar language is English, but I am happy to help out in German where necessary.
Even though often considered the "gold standard" to test scientific causal theories, experiments have long been a stranger to political science. In his 1909 presidential address to the American Political Science Association, A. Lawrence Lowell pointed to the "impossibility of experiment" and concluded: "Politics is an observational, not an experimental science..." (Lowell 1910, 7). Only towards the end of the 20th century, scholars began to acknowledge the benefits of random assignment to test causal political theories. Today, observational and experimental research are equally important for political science, each with their unique strengths and weaknesses and areas of application. A wide range of topics, in particular in the fields of political behaviour, public opinion and political psychology, have profited from both traditions, e.g. media effects and framing (Iyengar & Kinder 1987, Druckmann 2004), attitude change (Taber & Lodge 2006), turnout and voting (Gerber & Green 2000; Lodge et al. 1989) and political and social trust (Kugler et al. 2008).
The sessions covering methods will start with a quick history of observation and experiments in the social science in general. The concept of causality, and how it can be shown in both approaches, will be discussed. Basic statistical methods such as group comparisons and regression will illustrated using examples in R (no prior experience required). The issue of internal and external validity as well as analysis of heterogeneity will be discussed. The final sessions will revolve about the importance of ethics, benefits and drawbacks of incentivization and practical implementation of experiments. The use of online apps I have programmed illustrate key statistical ideas efficiently in class.
The topics sessions will be initiated by general topic introductions by the teacher, followed up by student presentations on more particular topics in the literature. For example, a presentation topic could be selective exposure to political news. Students would not only discuss the most important studies (e.g. Iyengar & Hahn 2009, Frimer et al. 2017), and illustrate in what way experiments are advantageous over (or coming short of) observational studies (e.g. Stroud 2008).
In addition, I will run several 5-min in-class studies at the beginning of sessions that will illustrate how state-of-the-art political science works in practice. Students can easily participate in these studies on their computer. I will analyze the these during sessions so that students learn the entire process of conducting until presenting research. Some of these studies can also be presented by students in later sessions instead of a literature presentation.
A full course outline, as well as signup for presentations will follow.
Online-Angebote
moodle
- Lehrende: Anonym
Semester: SoSe 2021