1. The document discusses different types of randomized experiments used in political science research, including survey experiments, field experiments, and randomized control trials.
2. It provides examples of survey experiments that randomly assign participants to different experimental conditions involving country conflict scenarios to measure the effect on policy preferences.
3. A field experiment is described that randomly assigned state legislators to email requests from constituents with either stereotypically black or white male names to examine differences in responses and whether they were influenced by the legislator's own attributes like race.
1. The document discusses different types of randomized experiments used in political science research, including survey experiments, field experiments, and randomized control trials.
2. It provides examples of survey experiments that randomly assign participants to different experimental conditions involving country conflict scenarios to measure the effect on policy preferences.
3. A field experiment is described that randomly assigned state legislators to email requests from constituents with either stereotypically black or white male names to examine differences in responses and whether they were influenced by the legislator's own attributes like race.
1. The document discusses different types of randomized experiments used in political science research, including survey experiments, field experiments, and randomized control trials.
2. It provides examples of survey experiments that randomly assign participants to different experimental conditions involving country conflict scenarios to measure the effect on policy preferences.
3. A field experiment is described that randomly assigned state legislators to email requests from constituents with either stereotypically black or white male names to examine differences in responses and whether they were influenced by the legislator's own attributes like race.
International Politics and Economy 7th Randomized Experiment
2 last quiz For the following items regarding the 1st quiz, consider whether there are any pairs of possible endogeneity factors and multicollinear factors. Result: Daily Sales of Pre-University Ice Cream Shop Factor: Advertisement/promotion costs for the most recent week 1 holiday or weekday, 2 temperature, 3 Average price of all products, 4 Presence or absence of university events, 5 customer satisfaction, 6 kind of taste, 7 Whether competitor stores are open or not 3 last class Self-selection bias (endogeneity, confounding variable) Random allocation Divide into 2 groups with or without treatment Align features other than target explanatory variables average effect 4 random experiment Identify effects based on random allocation An experiment conducted artificially and randomly according to a predetermined procedure for assigning treatment groups and non-treatment groups High internal validity Accurately identify effects; cf. simple comparison Multiple types of external validity general discussion =Whether the argument can be applied to groups other than the target group depends on the breadth of the target of analysis.; ex. Targeting specific cities vs. Targeting all of Japan 5 survey experiment Mainly conducted by public opinion polls Divide the survey subjects into two groups (1. specific information (intervention) group, 2. other group) Compare information-related preferences in two groups 6 Examples of survey experiments Tomz 2007 audience cost Country A threatens Country B as opposed to Country A and Country B. ex. "If we do not withdraw from the disputed territory, we will invade by force." Country B refuses A's request Any choice of country A will incur costs Execution of threat: Damage caused by confrontation/conflict with country B Withdrawal of threats: voter dissatisfaction = audience costs Self-selection bias ``Retracting threats'' and ``lowering voter support'' If voter support is low, the government will not intimidate = There is a high possibility that there is bias in cases that demand withdrawal of threats 7 Fictional stories Background: "Country A is attacking a neighboring country." Response: (1) "The United States did not send troops to country A." (2) "The United States said it would send troops to Country A, but it did not." Question: Do you support the U.S. government's decision? Two-group classification Selected 1127 people based on the overall characteristics of the US population (Age, gender, state, etc.) Survey via Internet Sent scenarios (1) to 650 people and (2) to 477 people * Scenarios with different characteristics of country A and the influence of the United States are prepared, but all are randomized 8 See figure. 9 field experiment Give information randomly and observe behavior Mainly surveyed voters and politicians Analysis of actual behavior makes it easy to link to policy proposals 10-12 Butler and Broockman (2011) Do politicians treat voters equally? Emails sent to 4859 state legislators in 44 states asking for help with voter registration Decide on the format, but change the contents randomly Male/female, white/black, partisan support, etc. ex. Sender's name is Jake Mueller typical white male name, DeShawn Jackson typical black male name page 11: see figure 13 Butler and Broockman (2011). See figure. 2747 email replies Obvious differences in correspondence between names The response differs depending on whether the lawmakers are white or not. White lawmakers, regardless of party, tend to ignore black name emails Democrats, non-white lawmakers respond to black name emails 14 Attributes of legislators influence how they treat citizens Social minorities tend to be ignored, but minority legislators respond Citizen representation should correspond to the composition of society based on attributes (descriptive representation).