1 - Lab outline

  1. guideline for MplusAutomation workflow
  2. create an R project in a dedicated project folder (on the desktop or in a class folder)
  3. install & load packages
  4. read in data to R
  5. view data in R
  6. view metadata (from SPSS files)
  7. write .sav / .csv / .dat files
  8. fix character names to have less than 8 character
  9. filtering rows & selecting columns
  10. change variable classes in R
  11. visualize and explore data
  12. introduction to mplusObjects

2 - Preparing to work with MplusAutomation

  1. R PROJECTS: We will use R Projects for ALL labs & assignments.
  1. THE {here} package: To make filepaths unbreakable (reproducible)
  1. PROJECT SUB-FOLDERS: Thoughfully organize files in sub-folders.
  1. LOCATION OF PROJECT FOLDERS: on desktop or within a single enclosing folder.

e.g., if/your/filepath/has/many/nested/folders/it/will/be/longer/than/the/90character/limit/data.dat

2.1 Tools we will use in lab

Tool/Package Purpose/Utility Advantages
{MplusAutomation} package Current capabilities supporting full SEM modeling Flexibility (approaching infinite)
R Project Unbreakable file paths & neatness Reproducibility (kindness to your future self)
{tidyverse} package Intuitive/descriptive function names Accesability to new users
{here} package Unbreakable/consistent file paths across OS Reproducibility (for Science’s sake!)
{haven} package Viewable metadata in R from SPSS datafiles Getting to know your measures
{ggplot2} package Beautiful, customizable, reporoducible figures Publication quality data vizualizations
pipe operator (%>%) notation Ease of reading/writing scripts e.g., first() %>% and_then() %>% and_finally()

3 - Creating an R-Project

  1. create a project folder (that will enclose all files associated with a given lab or assignment)
  1. create a new project (upper right hand corner of the R-studio window)
  2. create two sub-folders in the project folder, one called “data”, and one called “basic_mplus”

4 - Installing & loading packages

  1. Copy the packages below into your R-script
  2. Save your R-script in your main project folder
  3. If a yellow banner appears at the top of your screen click “install packages”

4.1 install the “rhdf5” package to read gh5 files

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
 BiocManager::install("rhdf5")
# how to install packages?
install.packages("tidyverse")

4.2 load packages

library(tidyverse)
library(haven)
library(here)
library(MplusAutomation)
library(rhdf5)
library(reshape2)
library(corrplot)
library(GGally)
library(corrgram)

4.3 Keyboard shortcuts

  • ALT + DASH(-) = <-
  • SHIFT + CONTROL = %>%

5 - Read in data

# object_name <- function1(nested_function2("dataset_name.sav"))

exp_data <- read_spss(here("data", "explore_lab_data.sav"))

\(\color{white}{\text{.}}\)

6 - A couple ways to explore & view data in R

# 1. click on the data in your Global Environment (upper right pane) or use...
View(exp_data)

# 2. str() allows you to view the structure of dataframe including class of variables (e.g., factors, ordered,... )
str(exp_data)

# 3. summary() gives basic summary statistics & shows number of NA values
#              *great for checking that data has been read in correctly*
summary(exp_data)

# 4. names() provides a list of column names. Very useful if you don't have them memorized!
names(exp_data)

# 5. head() prints the top x rows of the dataframe
head(exp_data)

# 6. glimpse() and another way to look at the dataframe, just depends on your preference!
glimpse(exp_data)

7 - View dataframe with labels & response scale meta-data

7.1 Note: Use the “print” option to save a PDF as a codebook of metadata.

# the {haven} package keeps the meta-data from SPSS files

# package_name::function_within_package()

sjPlot::view_df(exp_data)

7.2 Types of data for different tasks

  • SAV (e.g., spss_data.sav): this data format is for SPSS files & contains variable labels (meta-data)
  • CSV (e.g., r_ready_data.csv): this is the preferable data format for reading into R (no labels)
  • DAT (e.g., mplus_data.dat): this is the data format used to read into Mplus (no column names or strings)

NOTE: Mplus also accepts TXT formatted data (e.g., mplus_data.txt)

8 - Writing, reading, and converting data between 3 formats

8.1 prepare datasets, remove SPSS labeling

# write a CSV datafile (preferable format for reading into R, without labels)
write_csv(exp_data, here("data", "exp_lab1_data.csv"))

# write a SPSS datafile (preferable format for reading into SPSS, labels are preserved)
write_sav(exp_data, here("data", "exp_lab1_data.sav"))
# read the unlabeled data back into R
nolabel_data <- read_csv(here("data", "exp_lab1_data.csv"))
# write a DAT datafile (this function removes header row & converts missing values to non-string)
prepareMplusData(nolabel_data, here("data", "exp_lab1_data.dat"))

9 - Preparing column-names to be MplusAutomation ready

9.1 Task: Make all variable names fit within the 8-character name limit (Mplus) while avoiding duplicates.

9.2 Renaming columns manually…

# use function: rename(new_name = old_name)
new_names <- nolabel_data %>% 
  rename( school_motiv1 = item1 ,  
          school_motiv2 = item2 ,
          school_motiv3 = item3 ,
          school_comp1  = item4 ,
          school_comp2  = item5 ,
          school_comp3  = item6 ,
          school_belif1 = item7 ,
          school_belif2 = item8 ,
          school_belif3 = item9 )

9.3 What do you do if you have a large dataset with many column names that are > 8 characters?

  • first, remove all characters greater than 8 using str_sub()
  • second, make sure you don’t now have duplicate variable names
  • third, locate and change all duplicate names
# remove characters from the variable names that are greater than 8 characters 
names(new_names) <- str_sub(names(new_names), 1, 8)

# check if culumn names are unique 
test.unique <- function(df) {  ## function to identify unique columns
  
  length1 <- length(colnames(df))
  length2 <- length(unique(colnames(df)))        
  if (length1 - length2 > 0 ) {
    print(paste("There are", length1 - length2, " duplicates", sep=" "))
  }     
}

test.unique(new_names)

# locate duplicates (this will find the column of the first duplicate)
anyDuplicated(colnames(new_names))

10 - Filtering rows (observations) & selecting columns (variables).

10.1 When to use?… create new dataframe subsets for particular analyses or purposes

# filtering observations & selecting variables

# new_data_frame <- existing_data_frame %>% 
#       function_filter_rows(variable == value) %>% 
#       function_select_columns(first_column:ninth_column)

females <- nolabel_data %>% 
  filter(female == 1) %>% 
  select(1:9) # column numbers you are selecting for new dataframe object

# an alternative way to select columns, write a list of variable names (case-sensitive)

males <- nolabel_data %>% 
  filter(female == 0) %>% 
  select(item1, item2, item4)

10.2 A note on coding style:

  • Naming conventions: Be consistent!
  • I use the style lower snake case (e.g., this_is_lower_snake_case)
  • Annotate code generously
  • Let your code breath: use return often to spread code chuncks out vertically (dense paragraphs of code are a headache to look at)

11 - Changing variable class types

e.g., change class numeric to factor

str(var_class)
var_class <- nolabel_data

# change variable "female" to a factor 
var_class <- var_class %>% 
  mutate(female = factor(female))

# change a set of variables to factors using "modify_at"
var_class %>%
  modify_at(c(1:9), as.factor) %>%
  str()

# change all factors back to numeric using "modify_if"
var_class %>%
  modify_if(is.factor, as.numeric) %>%
  str()

12 - visualizing data using ggplot()

12.1 boxplot

# using ggplot: 
# whenever you are using ggplot and you specify a variable name
# it needs to be within "aes()" which stands for aesthetic

# Making a box plot
nolabel_data %>%          # the data frame
  ggplot(aes(y=item1)) +  # ggplot with aes( y = variable_name)
    geom_boxplot()        # specify the type of plot

12.2 histogram

nolabel_data %>%               
  ggplot(aes(x = item2)) +    
    geom_histogram()

12.3 density plot

nolabel_data %>%
  ggplot( aes(x=item3)) +
    geom_density(fill="#69b3a2",  # change the aesthetics
                 color="#e9ecef", # add a color
                 alpha=0.8)       # make fill transparent

12.4 jitter plot

nolabel_data %>%
  ggplot( aes(x=item3, y=item4)) +
    geom_jitter(alpha = .5 )

12.5 qq-plot

# Quantile-quantile plot: 

nolabel_data %>%
ggplot(aes(sample = item3)) +
  geom_qq(size = .8, alpha = 0.5) +
  facet_wrap(~female) +
  stat_qq_line() +
  labs(title = "Quantile-quantile plots, check of normality")

13 - Visualizing & exploring sets of variables

In factor analysis you often want to look at a set of variables at once rather than one at a time…

# loops, use purrr::map() to make histograms for a series of items

 nolabel_data %>% 
   select(1:9) %>% 
   names() %>%
   map(~ggplot(nolabel_data, aes_string(x = .)) + geom_histogram())

      
# alternatively, use the facet wrap function with melt to make one graph

melt(nolabel_data[,1:9]) %>%                  # within brakets [rows, columns]
  ggplot(., aes(x=value, label=variable)) +   
  geom_histogram(bins = 15) + 
  facet_wrap(~variable, scales = "free")      # scales="free" allows the x-axes to vary

13.1 faceted scatterplot

# scatterplot 
melt(nolabel_data[,1:9]) %>% 
  ggplot(., aes(y=value, x=variable)) + 
  geom_jitter() + 
  facet_wrap(~variable, scales = "free")

13.2 violin plot

# violin plot
melt(nolabel_data[,1:9]) %>% 
  ggplot(., aes(y=value, x=variable)) + 
  geom_violin() + 
  facet_wrap(~variable, scales = "free")

14 - correlations - there are many ways to visualize bi-variate relations in R

14.1 package {corrplot}

f_cor <- cor(females, use = "pairwise.complete.obs")

corrplot(f_cor, 
         method="number",
         type = "upper")
corrplot(f_cor, 
         method = "circle",
         type = "upper", 
         tl.col="black", 
         tl.srt=45)

14.2 package {corrgram}

corrgram(nolabel_data[,1:9],
         order=TRUE, 
         lower.panel=panel.ellipse, 
         upper.panel=panel.pts, 
         text.panel=panel.txt, diag.panel=panel.minmax, 
         main="Explore Data")

14.3 package {ggpairs}

ggpairs(
  nolabel_data[,1:9],
  upper = list(continuous = "density", combo = "box_no_facet"),
  lower = list(continuous = "points", combo = "dot_no_facet"))

15 - Simple introduction to MplusAutomation (writing, running, & reading models)

15.1 What does the mplusObject() function do:

  1. It generates an Mplus input file (does not need full variable name list, its automated for you!)
  2. It generates a datafile specific to each model
  3. It runs or estimates the model (hopefully) producing the correct output. Always check!

16 PRACTICE: Using MplusObject() method (type = BASIC;)

m_basic  <- mplusObject(
  TITLE = "PRACTICE 01 - Explore TYPE = BASIC", 
  VARIABLE = 
 "usevar=
 item1 item2 item3 item4 item5
 item6 item7 item8 item9 female; 
 
 ! use exclamation symbol to make comments, reminders, or annotations in Mplus files",
  
  ANALYSIS = 
 "type = basic; ",
 
  usevariables = colnames(nolabel_data), 
  rdata = nolabel_data)

m_basic_fit <- mplusModeler(m_basic, 
                            dataout=here("basic_mplus", "basic_Lab1_DEMO.dat"),
                            modelout=here("basic_mplus", "basic_Lab1_DEMO.inp"),
                            check=TRUE, run = TRUE, hashfilename = FALSE)

## END: TYPE = BASIC MPLUS AUTOMATION PRACTICE

17 PRACTICE: Now explore descriptives for observations that reported as “female”

Add line of syntax: “useobs = female == 1;”

\(\color{white}{\text{.}}\)

fem_basic  <- mplusObject(
  TITLE = "PRACTICE 02 - Explore female observations only", 
  VARIABLE = 
 "usevar=
 item1 item2 item3 item4 item5
 item6 item7 item8 item9; 
 
 useobs = female == 1; !include observations that report female in analysis",
  
  ANALYSIS = 
 "type = basic;",
 
  usevariables = colnames(nolabel_data), 
  rdata = nolabel_data)

fem_basic_fit <- mplusModeler(fem_basic, 
                            dataout=here("basic_mplus", "fem_basic_Lab1_DEMO.dat"),
                            modelout=here("basic_mplus", "fem_basic_Lab1_DEMO.inp"),
                            check=TRUE, run = TRUE, hashfilename = FALSE)

18 PRACTICE: Exploratory Factor Analysis (EFA)

## EXPLORATORY FACTOR ANALYSIS LAB DEMONSTRATION

efa_demo  <- mplusObject(
  TITLE = "EXPLORATORY FACTOR ANALYSIS - LAB DEMO", 
  VARIABLE = 
 "usevar=
 item1 item2 item3 item4 item5
 item6 item7 item8 item9;" ,
  
  ANALYSIS = 
 "type = efa 1 5;
  estimator = MLR;
  parallel=50;",
  
  MODEL = "" ,
  
  PLOT = "type = plot3;",
 
  OUTPUT = "sampstat standardized residual modindices (3.84);",
 
  usevariables = colnames(nolabel_data), 
  rdata = nolabel_data)

efa_demo_fit <- mplusModeler(efa_demo, 
                            dataout=here("basic_mplus", "EFA_Lab_DEMO.dat"),
                            modelout=here("basic_mplus", "EFA_Lab_DEMO.inp"),
                            check=TRUE, run = TRUE, hashfilename = FALSE)

## END: EXPLORATORY FACTOR ANALYSIS LAB DEMONSTRATION

19 END OF LAB

20 References

Hallquist, M. N., & Wiley, J. F. (2018). MplusAutomation: An R Package for Facilitating Large-Scale Latent Variable Analyses in Mplus. Structural equation modeling: a multidisciplinary journal, 25(4), 621-638.

Horst, A. (2020). Course & Workshop Materials. GitHub Repositories, https://https://allisonhorst.github.io/

Muthén, L.K. and Muthén, B.O. (1998-2017). Mplus User’s Guide. Eighth Edition. Los Angeles, CA: Muthén & Muthén

R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/

Wickham et al., (2019). Welcome to the tidyverse. Journal of Open Source Software, 4(43), 1686, https://doi.org/10.21105/joss.01686