Rachel_CleanData&HandleMissing.Rmd

---
title: "Rachel's Replication: Clean Data & Handle Missing"
author: "Rachel Ganly"
date: "9/27/2021"
output: html_document
---

Clear the data space
1. Load required packages for analyses.  
2. Converting data from the GSS into a tibble.  
3. Cleaning and coding variables to be used in the analyses.  
4. Dropping all variables but the ones used in analyses.  
5. Labeling those variables. 

```{r}

#Follow Tyler's instructions from the Shell File

#1. Load required packages for analyses

# Clear dataspace
rm(list = ls())

library(tidyverse)
library(haven)
library(skimr)
library(naniar)
library(Hmisc)
library(sjlabelled)
library(gt)
library(gtsummary)
# Rachel's comments: Found these packages in the Readme list on GitHub


```

```{r}

# 2. Converting data from the GSS into a tibble.  

gss_rachel<-read.csv("gss_2017.csv") #Loaded file gss_2017 per the instructions
gss_rachel<-as_tibble(gss_rachel)


# 3. Cleaning and coding variables to be used in the analyses.  
# Rachel's comment: cannot find a list of variables to be used, except by reading the code

gss_rachel <- gss_rachel %>%
  select(c(SRH_110,                   # Self-Rated Health
           SRH_115,                   # Self-Rated Mental Health
           EHG3_01B,                  # Highest Educational Attainment
           AGEC,                      # Respondent's Age in Years
           SEX,                       # Respondent's Sex
           MARSTAT,                   # Marital Status
           FAMINCG2,                  # Family Income
           CHRINHDC,                  # Number of Children in Household
           VISMIN,                    # Visible Minority
           WGHT_PER))                 # Individual Survey Weight

# Create new values
  # Self-Rated Health (Ordinal)
      # Coded Excellent (1) - Poor (5)
  # Self-Rated Mental Health (Ordinal)
      # Coded Excellent (1) - Poor (5)
  # Educational Attainment (Ordinal)
      # Coded HS or Less (1) | Some College (2) | Bachelor's + (3)
 # Age in Years (Continuous; capped at 80)
  # Sex (Binary)
      # Coded Male (0) | Female (1)
  # Marital Status (Changing from Nominal to Dichotomous)
      # Married/Common Law = Married (1) | All Else = Not Married (0)
  # Family Income (Ordinal)
      # Values are: 1) Less than $25k; 2) $25k to $49.999k; 3) $50k to 74.999k;
      #             4) $75k to $99.999k; 5) $100k to $124.999k; 6) $125k or more
  # Number of Children in Household (Discrete)
 # Visible Minority (Binary) 
      # Coded Yes (1) | No(0)

gss_rachel<-gss_rachel%>%mutate(srh=SRH_110,
                  mental_srh=SRH_115,
                  education=case_when(EHG3_01B<=2~1,
                                      EHG3_01B>=3&EHG3_01B<=5~2,
                                      EHG3_01B>=6&EHG3_01B<=7~3),
                  age=AGEC,
                  sex=case_when(SEX==1~0,
                                SEX==2~1),
                  marital_status=case_when(MARSTAT == 1 | MARSTAT == 2 ~ 1,
                                          MARSTAT >= 3 & MARSTAT <= 6 ~ 0),  
           family_income=as.integer(FAMINCG2),                  # Family Income
           number_children=as.integer(CHRINHDC),                  # Number of Children in Household
           visible_minority=case_when(VISMIN == 1 ~ 1,
                                      VISMIN == 2 ~ 0),                 # Visible Minority
           individual_survey_weight=WGHT_PER)             # Individual Survey Weight

                  


# LABELLING VARIABLES AND VALUES


#4. Dropping all variables but the ones used in analyses. 
# Rachels Note: could only find the list of variables to be used by looking at the code and had already dropped most values earlier so I initially skipped this step but went back to add it later



#5. Label variables. 
# No notes on labelling of variables except by looking at the code

save(gss_rachel, file="GSS_Cleaned_Rachel.csv")

```


```{r}
#Handle Missing Data


#1. Recodes missing values of self-rated health and self-rated mental health (other measures were assigned missing values during the cleaning phase).  
# 2. Creates an index called "sampmiss" that is a count of how many variables that each respondent has missing values.  
#  3. Creates a dataset called "sample" that only contains cases with no missing values.

# Rachel's comment: I could not find out how to do this using the instructions or comments (for example how were the values recoded in step 1); so I had to read and copy the code

# CODING DATA VALUES AS MISSING
    # Note: Variables not included have no missing data or were already coded 
    # in the cleaning stage

gss_rachel <- gss_rachel %>% replace_with_na(replace = list(srh =       c(7, 8, 9),
                                              mental_srh =     c(7, 8, 9)))


# CREATING SAMPLE VARIABLE FOR ANALYSES
    # Will use this variable to filter out cases missing any data
gss_rachel <- gss_rachel %>% mutate(sampmiss = rowSums(is.na(.)))
gss_rachel %>% count(sampmiss)

# SAMPLE DATASET WHICH EXCLUDES MISSING DATA (LISTWISE DELETION)
sample <- gss_rachel %>% filter(sampmiss == 0)

save(gss_rachel, file="GSS_NoMissing_Rachel.csv")


```