The BYU effect

2_the_byu_effect.r

# load packages
library(tidyverse)
library(sjlabelled)
library(clipr)

data <- read.csv("./data/fire/fire.csv")
labels <- read.csv("./data/fire/labels.csv")

# checking labels
labels |> 
    filter(Variable == "school") # mormon == 3

# create variables for analysis
data$party <- factor(case_when(
    data$partyid %in% c(1,2,3) ~ "Dem",
    data$partyid %in% c(4) ~ "Ind",
    data$partyid %in% c(5,6,7) ~ "Rep",
    data$partyid %in% c(8) ~ "Other",
), levels = c("Rep", "Ind", "Dem", "Other"))

data$BYU <- as.integer(data$school == "Brigham Young University-Provo")
data$UU <- as.integer(data$school == "University of Utah")
data$USU <- as.integer(data$school == "Utah State University")
data$school_rc <- dplyr::case_when(
    data$school == "Brigham Young University-Provo" ~ "BYU",
    data$school == "University of Utah" ~ "UU",
    data$school == "Utah State University" ~ "USU",
    TRUE ~ "Other"
)


data$is_anxious <- as.integer(data$anxious %in% c(3,4,5))
data$is_lonely <- as.integer(data$lonely %in% c(3,4,5))
data$no_time_self <- as.integer(data$notime %in% c(3,4,5))
data$is_depressed <- as.integer(data$depressed %in% c(3,4,5))
data$is_stressed <- as.integer(data$stressed %in% c(3,4,5))

recodes <- c(
    '1' = 'business', #Accounting
    '2' = 'social sciences', #African/Afr-American Studies
    '3' = 'natural sciences', #Agricultural/Food Sciences
    '4' = 'natural sciences',#Animal Science
    '5' = 'social sciences',#Anthropology
    '6' = 'natural sciences',#Archaeology
    '7' = 'engineering',#Architecture
    '8' = 'humanities/arts',#Art History
    '9' = 'humanities/arts',#Asian/Mideast Lang & Lit
    '10' = 'humanities/arts',#Asian/Mideast Studies
    '11' = 'natural sciences',#Physics and Astronomy
    '12' = 'natural sciences',#Biochemistry and Molecular Biology
    '13' = 'natural sciences',#Biological Sciences
    '14' = 'business',#business
    '15' = 'natural sciences',#Chemistry
    '16' = 'humanities/arts',#classics
    '17' = 'social sciences',#Cognitive Science
    '18' = 'humanities/arts',#Communication
    '19' = 'engineering',#Computer Science
    '20' = 'humanities/arts',#Dance
    '21' = 'engineering',#Data Science
    '22' = 'social sciences',#Development Studies
    '23' = 'business',#Digital Media Design
    '24' = 'natural sciences',#Earth Sciences
    '25' = 'social sciences',#Economics
    '26' = 'teaching',#Education
    '27' = 'engineering',#Education
    '28' = 'humanities/arts',#English
    '29' = 'natural sciences',#Environmental Studies
    '30' = 'social sciences',#Ethnicity and Race Studies
    '31' = 'humanities/arts',#European Cultural Studies
    '32' = 'humanities/arts',#Fashion
    '33' = 'humanities/arts',#Film and Media Studies
    '34' = 'business',#Finance
    '35' = 'humanities/arts',#Fine Arts
    '36' = 'humanities/arts',#French and Italian
    '37' = 'natural sciences',#Geography
    '38' = 'humanities/arts',#German Studies
    '39' = 'humanities/arts',#History
    '40' = 'business',#Hotel Administration
    '41' = 'social sciences',#International Relations
    '42' = 'humanities/arts',#Jewish Studies
    '43' = 'humanities/arts',#Lat Am/Latino/Carib Studies
    '44' = 'social sciences',#Law/Criminology
    '45' = 'humanities/arts',#Linguistics
    '46' = 'business',#Marketing
    '47' = 'engineering',#Mathematics
    '48' = 'humanities/arts',#Music
    '50' = 'natural sciences',#Neuroscience
    '51' = 'natural sciences',#Nursing
    '52' = 'humanities/arts',#Philosophy
    '53' = 'teaching',#Physical Education
    '54' = 'social sciences',#Political Science
    '55' = 'humanities/arts',#Portuguese and Brazilian Studies
    '56' = 'natural sciences',#Pre-Medical
    '57' = 'natural sciences',#Psychological and Brain Sciences
    '58' = 'natural sciences',#Public Health
    '59' = 'social sciences',#Public Policy
    '61' = 'business',#Real Estate
    '62' = 'humanities/arts',#Religion
    '63' = 'engineering',#Robotics and Intelligent Systems
    '65' = 'social sciences',#Sociology
    '66' = 'humanities/arts',#Spanish and Portuguese
    '67' = 'natural sciences',#Statistics
    '68' = 'humanities/arts',#SStudio Art
    '69' = 'humanities/arts',#Theater
    '70' = 'other',#Undecided
    '71' = 'social sciences',#Urban Studies
    '72' = 'humanities/arts',#Visual Arts
    '73' = 'social sciences',#Women's and Gender Studies (and Sexuality)
    '74' = 'other'#other
)

data$major_type <- as.character(recodes[data$major1])

data$religattend_rc <- car::recode(data$religattend, '8:9=1;else=0')

data$gender_rc <- factor(car::recode(data$gender, '1="male";2="female";3:6="other";-99=NA'), 
    levels = c('male', 'female', 'other'))

data$grad_level <- factor(car::recode(
    data$gradyear, 
    "2024='Senior';2025='Junior';2026='Sophomore';2027='Freshman'"), 
        levels = c('Freshman', 'Sophomore', 'Junior', 'Senior'))

data$grad_level2 <- factor(car::recode(
    data$gradyear, 
    "2024='Junior/Senior';2025='Junior/Senior';2026='Freshman/Sophomore';2027='Freshman/Sophomore'"), 
        levels = c('Freshman/Sophomore', 'Junior/Senior'))

# its too sparse so need to collapse major type again
data$major_type2 <- as.character(recode_factor(data$major_type, 
    'business' = 'business/engineering/natural sciences',
    'social sciences' = 'social sciences/humanities/arts/teaching/other',
    'natural sciences' = 'business/engineering/natural sciences',
    'engineering' = 'business/engineering/natural sciences',
    'humanities/arts' = 'social sciences/humanities/arts/teaching/other',
    'teaching' = 'social sciences/humanities/arts/teaching/other',
    'other' = 'social sciences/humanities/arts/teaching/other'
))

# Summarizing data for post
## Church attendance over time
data |> 
    filter(religion == 3) |> 
    group_by(grad_level, BYU) |>
    count(religattend_rc) |> 
    drop_na() |>
    mutate(prop = n / sum(n)) |> 
    pivot_wider(id_cols = c(grad_level,BYU),names_from = religattend_rc, values_from = prop) |> 
    arrange(BYU, grad_level) |> 
    write_clip()

# using school_rc variable
data |> 
    filter(religion == 3) |> 
    group_by(grad_level2, school_rc) |>
    count(religattend_rc) |> 
    drop_na() |>
    mutate(prop = n / sum(n)) |> 
    pivot_wider(id_cols = c(grad_level2,school_rc),names_from = religattend_rc, values_from = prop) |> 
    arrange(school_rc, grad_level2) |> 
    write_clip()

# quick peak at majors
data |> 
    filter(religion == 3) |> 
    group_by(major_type) |>
    count(religattend_rc) |> 
    drop_na() |>
    mutate(prop = n / sum(n)) |> 
    pivot_wider(id_cols = c(major_type),names_from = religattend_rc, values_from = prop) |> 
    arrange(desc(`1`)) |> 
    select(major_type, `1`)

# quick peak at majors
data |> 
    filter(religion == 3) |> 
    group_by(gender_rc) |>
    count(religattend_rc) |> 
    drop_na() |>
    mutate(prop = n / sum(n)) |> 
    pivot_wider(id_cols = c(gender_rc),names_from = religattend_rc, values_from = prop) |> 
    arrange(desc(`1`)) |> 
    select(gender_rc, `1`)

data |> 
    filter(religion == 3) |> 
    group_by(gender_rc, major_type) |>
    count(religattend_rc) |> 
    drop_na() |>
    mutate(prop = n / sum(n)) |> 
    pivot_wider(id_cols = c(gender_rc,major_type),names_from = religattend_rc, values_from = prop) |> 
    arrange(major_type) |> 
    filter(gender_rc != "other" & major_type != "other") |> 
    select(gender_rc,major_type, `1`)
# gist: https://gist.github.com/acbass49/d419ab4f3e7551a382ed8eee85adb91f