Error with sem function at R : differences in factors

Question

I wanted to use the function sem (with the package lavaan) on my data in R :

Model1<- 'Transfer~Amotivation+Gender+Age
Amotivation~Gender+Age

transfer are 4 questions with a 5 point likert scale 
Amotivation: 4 questions with a 5 pint likert scale
Gender: 0 (=male) and 1 (=female)
Age: just the different ages 

And i got next error:

in getDataFull (data= data, group =  group, grow.label = group.label,:
lavaan WARNING: some observed variances are (at least) a factor 100 times larger than others; please rescale 

Is anybody familiar with this error? Does it influence my results? Do I have to change anything? I really don't know what this error means.

Answer 1

Your scales are not equivalent. Your gender variables are constrained to be either 0 or 1. Amotivation is constrained to be between 1 and 5, but age is even less constrained. I created some sample data for gender, age, and amotivation. You can see that the variance for the age variable is over 4,000 times higher than the variance for gender, and about 500 times higher than sample amotivation data.

gender  <- c(0,1,1,1,0,0,1,1,0,1,1,0,0,1,1,1)
age <- c(18,42,87,12,24,26,98,84,23,12,95,44,54,23,10,16)
set.seed(42)
amotivation <- rnorm(16, 3, 1.5)
var(gender)      #    0.25 variance
var(age)         # 1017.27 variance
var(amotivation) #    2.21 variance

I'm not sure how the unequal variances influence your results, or if you need to do anything at all. To make your age variable more closely match the amotivation scale, you could transform the data so that it's also on a 5 point scale.

newage <- age/max(age)*5
var(newage) # 2.65 variance

You could try running the analysis both ways (using your original data and the transformed data) and see if there are differences.