HW 07

The template for this assignment may be found in the repository at the following link: https://classroom.github.com/a/PVhzenpb

Today’s data come from a survey on resident perceived quality of life in Durham, NC, conducted in 2020.

Below is a data dictionary for the dataset.

• ID: the ID of the participant taking the survey
• qual_xxx: The quality rating score given to the fire department, EMS service, public transit system, water utilities, park system, public school system, etc. This is an ordinal variable taking on values 1 = very dissatisfied through 5 = very satisfied (3 = neutral).
• overall_image: An ordinal variable corresponding to overall image of the city as a whole.
• qol_durham: The overall quality of life in Durham
• years_in_durham: The number of years lived in Durham
• gender: self-identified gender of the respondent
• rent_own: whether the respondent rents or owns their primary residence
• city_limits: whether the respondent lives within city limits

1. The primary research question is whether long-time residents of Durham have differential perceptions of overall quality of life (qol_durham) compared to more recent transplants. You are considering an ordinal model vs. a multinomial model. In the context of these two models, explain what the proportional odds assumption and irrelevance of independent alternatives assumptions are in the context of the data. Which of the two models do you think is more appropriate? Explain.
2. Create a new variable corresponding to the respondent's average perception of the quality of city services.This variable should be created by averaging their responses to quality of police services, fire department, EMS, public transit, water utilities, park system, and public schools (among non-missing values for these variables). You may treat both this variable as a continuous variable; discard any observations that do not have any ratings at all for each of the seven city services listed (i.e., if you cannot calculate an average city services rating). What is the average city services score among all survey respondents? Provide your answer to at least 3 decimal places.
3. Regardless of what was chosen for Exercise 1, fit an ordinal regression model that aims to predict the resident's perceived QOL in Durham based on whether the respondent rents or owns their residence, whether they live in city limits, and the continuous variable you created in Exercise 2. Interpret the coefficients corresponding to rent vs. own status and the continuous variable in terms of relevant odds ratios.
4. Regardless of what was chosen for Exercise 1, fit a multinomial regression model that aims to predict the resident's perceived QOL in Durham based on the same predictors as in Exercise 3. Compare and contrast conclusions you make from this model to conclusions from the model in Exercise 3 in terms of relationships between rent/own and continuous resident average perception score vs. perceited overall QOL.