Class Notes (806,874)
Canada (492,493)
POL101Y1 (1,114)

Jan 22 2014.doc

4 Pages
Unlock Document

University of Toronto St. George
Political Science
Kenichi Ariga

Jan 22 2014 POL 322 H1 How Do We Test Our Theory? Research Design Theory Obstacles to Establishing Causality: 1. Reverse Causality & Simultaneous Relationship 2. Confounding Variables- Spurious, Misguided and Hidden Relationship Research Designs: Strategies that researchers can use to empirically test whether X causes Y 1. Experiments 2. Observational Studies - There are advantages and disadvantages for various research designs Research Design I: Experiments Logic of Experimentation What is an experiment?  2 components: 1. A researcher controls values of independent variable (x) 2. Researcher randomly assigns values of (x) Example from slide X=1 group received treatment (Treatment group) By controlling value of x, we can observe different rates and this will exclude possibility of reverse causality X=0 did not receive treatment (control group) Z= characteristics of participants (these are potential confounding variables for relationship of x and y) Z will balanced between two groups (X=0, X=1) -If we compare average outcome between control and treatment group we can see impact of x on y Treatment Assignment (=Control of X) - Must take great care when assigning value of x to different groups. - If we do not give anything to a control group, placebo effect may become a confounding variable (control group may believe that their condition has either improved or not when they were not administered the drug). Cannot distinguish between placebo and actual effect - Usually there is an ethical issue when concerning drug test, when the merit of the studies outweigh the risks of ethical issue then a clearance to conduct experiments will be given Different Types of Experiments 1. Laboratory Experiments 2. Field Experiments 3. Survey Experiments 1. Laboratory Experiments (Example: Turner 2007 in blackboard) - Perception of Bias (ex. CNN and Fox News) CNN is biased towards middle, Fox biased towards Conservative - Perceived Ideological Bias of News what determines this perception of people? - People could perceive perception bias because of “actual bias” OR News Media’s reputation - Difficult to distinguish between actual bias vs. news media’s reputation because they could be confounding variable to each other - Replicated News from the two different channels and changed their broadcast names or no attribution his theory is that there is a reputational effect in the media, any difference you can find can be controlled for Results: When there is no attribution of channel to the news, it is neutral. When it has fox news logo it was conservative. When it was CNN logo Neutral results. - The difference can be attributed to representational attribution. Causal impact on representation of media by perceived bias of individuals Validity of Experimental Studies 1. Internal Validity 2. External Validity 3. Laboratory Experiments 1. Any conclusion drawn from this experiment has high internal validity. The degree of confidence in the casual conclusion drawn from analysis. 2. About the theory of Confidence in generalizability of the casual conclusion drawn from the analysis. Cannot generalize for the entire pop. 3. Relatively high internal validity Different Types of Experiments Laboratory Experiments Field Experiment (conducted in a real world setting) Ex. • Clientelism votes Studying: relationship between campaign promise and election outcomes • Local Transfers Votes - Campaign promises based on clientelism, nepotism, and personal relationships - Some argue that in developing societies, clientelism is stronger, but no serious empirical studies that verifies these two different campaign promises ^ - We cannot distinguish the impact of both of them by just pure observation Ex. Wantchekon Example: 2001 Presidential election in Benin (a relatively stable democratic society) Candidates: Kerekou Incumbent president
More Less

Related notes for POL101Y1

Log In


Don't have an account?

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.