1 Environment preparation

# ### Data import ###
# install.packages("readr")     # read datasets
# install.packages("qualtRics") # read qualtrics datasets
# ### Data manipulation ###
# install.packages("dplyr")     # pipes and data manipulation
# ### Visualization ###
# install.packages("ggplot2")    # graphing capabilities
# ### Estimation ###
# install.packages("cjoint")    # base amce package
# install.packages("cregg")     # amce and mm 
# install.packages("factorEx")  # amce with non-uniform distribution

## Custom build functions 
# library(devtools)
# install_github("albertostefanelli/cjoint") # fixes some problem with cjoint

### Data import ###
library("readr")     
library("qualtRics") 
### Data manipulation ###
library("dplyr")     
### Visualization ###
library("ggplot2")    
### Estimation ###
library("cjoint")   
library("cregg")     
library("factorEx")  

2 Data

1. From the paper “The Ties that Double Bind” (Teele, Kalla, and Rosenbluth 2018)
2. We are going to use the Legislators and Voters sample
3. The authors want to test different type of gender biases
   • Among Legislators
   • Among Voters

• Variables in the dataset
  • CJ choice task= “winner”
    
  • Type of sample = “sample”
  • Respondent’s gender=“female_respondent”
    
  • Respondent’s PID= “democrat_respondent”, “republican_respondent”
    
  • Candidate’s gender= “orig_cand_female”
    
  • Candidate’s political Experience= “orig_0ys”, “orig_1ys”, “orig_3ys”, “orig_8ys”
  • Candidate’s number of Children= “orig_0ch”, “orig_1ch”, “orig_3ch”
    
  • Candidate’s political Post= “orig_may”, “orig_leg”
    
  • Candidate’s occupation= “orig_law”, “orig_teach”
    
  • Candidate’s spouse’s Occupation= “orig_UN_sp”, “orig_FM_sp”, “orig_MD_sp”
    
  • Candidate’s candidate’s Age= “orig_29”, “orig_45”, orig_65”

3 Task

1. Postulate Hypothesises/Expectation
2. Estimate the AMCE for the candidates’ traits
   • See if your Hypothesises/Expectation hold
   • Interpret the AMCE
3. Re-estimate the AMCE using a simple OLS
   • What happened?
4. Re-estimate the AMCE using a logit model
   • What happened to the coefficients? How can we interpret them?
5. [OPTIONAL] Estimate the AMCE in the voter sample
   • Are there any differences?
6. Plot the results
7. Test for respondent’s fatigue

# load the data 
teele_data <- readr::read_csv("https://github.com/albertostefanelli/conjoint_class/raw/master/data/teele.csv")