Lecture 17: Longitudinal modeling in practice

Study on sleep and quality of life

Published

March 20, 2024

library(tidyverse)
library(knitr)
library(lme4)
library(broom.mixed)
library(skimr)
sleep <- read_csv("data/sleep-study.csv") |>
  mutate(sex = as_factor(sex), 
         educ_kat = as_factor(educ_kat), 
         estat = as_factor(estat), 
         hincome_6 = as_factor(hincome_6))

Introduction

The analysis for this AE is based on the paper Kudrnáčová and Kudrnáč (2023). We will use the following variables:

Response variables

  • lfsat_OT: life satisfaction
  • wellbe_OT: Wellbeing
  • happy_OT: Happiness

Sleep variables

  • slequal_OT: Sleep quality
  • SDweek_OT: Sleep duration
  • jetlag_OT: Social jetlag

Other covariates

  • sex: 1: male, 2: female
  • educ_kat: highest level of education attained
    • 1: basic and secondary vocational
    • 2: secondary with maturita
    • 3: tertiary education
  • hincome_6: household income
    • 1: up to CZK 22,999:
    • 2: CZK 23,000 to 29,999
    • 3: CZK 30,000 to 34,999
    • 4: CZK 35,000 to 39,999
    • 5: CZK 50,000 to 74,999
    • 6: more than CZK 75,000
  • age: age in years
  • estat: employment status
    • 0: employed
    • 1: self-employed
    • 2: unemployed
    • 3: students
    • 4: retired
    • 5: maternity leave
  • deti5_OT: number of children below age 5 in the household
  • wave:
    • 3: fourth wave (2018)
    • 4: fifth wave (2019)
    • 5: sixth wave (2020)

Exploratory data analysis

Exercise 1

Examine the distribution of your response variable. What do you observe?

# add code
Exercise 2

Examine the response variable over time. What do you observe?

# add code
Exercise 3

Examine the relationship between the response variable and the sleep variables. What do you observe?

# add code

Modeling

Exercise 4

Fill in your response variable in the code below to make the linear mixed model (similar to the models in Table 2 of the paper).

Remove #| eval: false once you have filled in the code.

sleep_model <- lmer(______ ~ SDweek_OT + slequal_OT + jetlag_OT + SDweek_OT:wave + slequal_OT:wave + jetlag_OT:wave + wave +  sex + educ_kat + age + 
             hincome_6 + estat + deti5_OT + (1|w1_hid) + (wave|w1_pid), 
           data = sleep, control = lmerControl(check.nobs.vs.nRE = "ignore"))

tidy(sleep_model) |> 
  kable(digits = 3)
Exercise 5

What initial conclusions do you draw from the model?

References

Kudrnáčová, Michaela, and Aleš Kudrnáč. 2023. “Better Sleep, Better Life? Testing the Role of Sleep on Quality of Life.” Edited by Fakir Md Yunus. PLOS ONE 18 (3): e0282085. https://doi.org/10.1371/journal.pone.0282085.