Trouble with Confirmatory Factor Analysis in R using lavaan package

CFA
1

I’m trying to run a Confirmatory Factor Analysis (CFA) in R using the lavaan package, but I’m hitting a snag. Here’s what’s happening:

model <- '
  # Factor definitions
  Quality =~ q1 + q2 + q3 + q4
  Service =~ s1 + s2 + s3 + s4
  Value =~ v1 + v2 + v3 + v4

  # Regressions
  Satisfaction + Loyalty ~ Quality + Service + Value
  Loyalty ~ Satisfaction

  # Covariances
  Satisfaction ~~ Loyalty
  Satisfaction ~~ Quality
'

result <- cfa(model, data=survey_data)

When I run this, I get an error saying some variables are missing from my dataset. But I’m sure they’re there! I’ve double-checked the column names and everything.

I’ve also tried using sem() instead of cfa(), but no luck. Am I missing something obvious? Any ideas on how to troubleshoot this?

2

Hey SurfingWave! :ocean:

Oof, CFA can be a real head-scratcher sometimes, right? I feel your pain!

Have you tried running names(survey_data) to double-check all your variable names? Sometimes sneaky spaces or capitalization differences can throw things off.

Also, I’m curious - are Satisfaction and Loyalty actually in your dataset? Your model is treating them as latent variables, but they’re not defined with the =~ operator like the others. Maybe try defining them explicitly?

Satisfaction =~ sat1 + sat2 + sat3
Loyalty =~ loy1 + loy2 + loy3

Just a thought! Let me know if that helps or if you want to bounce around more ideas. CFA can be tricky, but we’ll figure it out! :blush:

What kind of data are you working with, by the way? Sounds like an interesting project!

3

yo surfingWave, i’ve been there! check if ur variables are actually in the data with head(survey_data). sometimes R gets weird with case sensitivity, so make sure ur var names match exactly. also, try running the model step by step, adding one factor at a time. that way u can pinpoint where its freaking out. good luck dude!

4

I’ve encountered similar issues with lavaan before. One thing to check is whether all your variables are numeric. Sometimes, if a variable is stored as a factor or character, lavaan won’t recognize it properly.

Try running str(survey_data) to inspect the structure of your dataset. If any variables are non-numeric, you might need to convert them:

survey_data ← lapply(survey_data, as.numeric)

Also, ensure there are no missing values in your data. lavaan can be sensitive to NAs. You could use na.omit(survey_data) to remove rows with missing data, but be cautious as this might affect your sample size.

If these don’t solve the issue, you might want to try fitting a simpler model first (maybe just one factor) to isolate where the problem is occurring. Gradually build up to your full model, adding components one at a time.