immigration {cregg} | R Documentation |
Immigration Conjoint Experiment Dataset from Hainmueller et. al. (2014)
Description
A dataset containing the results of a conjoint survey of a representative sample of American adults who were asked to choose which hypothetical immigrants they think should be admitted into the United States. Each row corresponds to a single profile presented to the respondent. The dataset results from a mostly full factorial design with restrictions on two combinations of features. (1) Profile immigrants from ‘CountryOfOrigin’ “India”, “Germany”, “France”, “Mexico”, “Philippines”, and “Poland” could be paired only with ‘ReasonForApplication’ “Seek better job” or “Reunite with family”; profiles from the remaining countries could be paired with any ‘ReasonForApplication’. (2) Profile immigrants with ‘Job’ “Financial Analyst”, “Computer Programmer”, “Research Scientist”, or “Doctor” could not be paired with ‘Education’ levels “No Formal”, “4th Grade”, “8th Grade”, or “High School”. All other features were fully randomized against all other features.
Usage
data(immigration)
Format
A data frame (with additional “cj_df” class) with 13960 observations on the following 16 variables.
- ‘CaseID’
a numeric vector indicating the respondent to which the particular profile corresponds
- ‘contest_no’
a numeric vector indicating the number of the task to which the profile corresponds
- ‘Education’
a factor with levels “No formal”, “4th grade”, “8th grade”, “High school”, “Two-year college”, “college Degree”, “Graduate degree”
- ‘Gender’
a factor with levels “Female”, “Male”
- ‘CountryOfOrigin’
a factor with levels “India”, “Germany”, “France”, “Mexico”, “Philippines”, “Poland”, “China”, “Sudan”, “Somalia”, “Iraq”
- ‘ReasonForApplication’
a factor with levels “Reunite with family”, “Seek better job”, “Escape persecution”
- ‘Job’
a factor with levels “Janitor”, “Waiter”, “Child care provider”, “Gardener”, “Financial analyst”, “Construction worker”, “Teacher”, “Computer programmer”, “Nurse”, “Research scientist”, “Doctor”
- ‘JobExperience’
a factor with levels “None”, “1-2 years”, “3-5 years”, “5+ years”
- ‘JobPlans’
a factor with levels “Will look for work”, “Contract with employer”, “Interviews with employer”, “No plans to look for work”
- ‘PriorEntry’
a factor with levels “Never”, “Once as tourist”, “Many times as tourist”, “Six months with family”, “Once w/o authorization”
- ‘LanguageSkills’
a factor with levels “Fluent English”, “Broken English”, “Tried English but unable”, “Used interpreter”
- ‘ChosenImmigrant’
a numeric vector denoting whether the immigrant profile was selected
- ‘ethnocentrism’
a numeric vector
- ‘profile’
a numeric vector giving the profile number
- ‘LangPos’
a numeric vector
- ‘PriorPos’
a numeric vector
Note
This is a modified version of the ‘hainmueller’ dataset available from the cjoint package.
Source
Hainmueller, J., Hopkins, D., and Yamamoto T. 2014. “Causal Inference in Conjoint Analysis: Understanding Multi-Dimensional Choices via Stated Preference Experiments.” Political Analysis 22(1): 1-30. http://doi.org/10.1093/pan/mpt024
See Also
Examples
data("immigration")
# view constraints between features
subset(cj_props(immigration, ~ Job + Education, id = ~ CaseID), Proportion == 0)
subset(cj_props(immigration, ~ ReasonForApplication + CountryOfOrigin,
id = ~ CaseID), Proportion == 0)
# AMCEs with interactions for constraints
f1 <- ChosenImmigrant ~ Gender + Education * Job +
LanguageSkills + CountryOfOrigin * ReasonForApplication +
JobExperience + JobPlans + PriorEntry
cj(immigration, f1, id = ~ CaseID)