decent growth model fit but negative variance error

[Brie M]

Hi all,

I am working on a relatively simple growth model:

mod1 <- 

'

  # intercept and slope with fixed coefficients

 i =~ 1*baz12 + 1*baz5y + 1*baz10y 

 s =~ 0*baz12 + 5*baz5y + 10*baz10y 

log_hs_crp ~ s + i + A_age + sex + iron_supp

s ~ sex 

# residual variances

baz12 ~~ baz12

baz5y ~~ baz5y

baz10y ~~ baz10y

'

fit1 <- growth(mod1, data = data_lavaan, missing="FIML")

summary(fit1, standardize = T, fit.measures = T)

And I get an acceptable fit but also this error message: 

lavaan WARNING: some estimated ov variances are negative

With a weird negative variance for BMI z-score at 10y:

Variances:

                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all

   .baz12             0.567    0.052   10.873    0.000    0.567    0.642

   .baz5y             0.680    0.050   13.501    0.000    0.680    0.505

   .baz10y           -0.343    0.079   -4.337    0.000   -0.343   -0.250

   .log_hs_crp        2.559    0.148   17.247    0.000    2.559    0.932

   .i                 0.316    0.044    7.105    0.000    1.000    1.000

   .s                 0.014    0.001   10.949    0.000    0.995    0.995

It's literally growth data on a cohort of kids, and it should be pretty straightforward (though there was an obesity crisis happening in the middle of it...). Is there a way to fit a model here without negative variance?  I've attached data and also screenshots of individual growth curves (n=600) in case that is helpful/mildly interesting.

Thanks for your help!

[Chao Xu]

This warning is usually due to model misspecification (e.g., a covariance is fixed to 1 when it should not be). I would suggest you to write down the set of equations. It is easier to find the source of errors this way based on my experiences.