How to calculate the significance of latent variables with lavaan ?

[Matt]

Hi,

I am trying to figure out a way to calculate the significance of the factors correlations using lavaan in R 3.5.3.

Could anybody help me with this topic ?

For example. I have a F1 =~ X1 + X2 + X3; and F2=~ X4+ X5 + X6; How can I corr(F1,F2) for significance ?

Best Regards

Matt

[Terrence Jorgensen]

I am trying to figure out a way to calculate the significance of the factors correlations using lavaan in R 3.5.3.

If the factor covariance is a model parameter, just look in the summary() output.  If your null hypothesis is 0, that implies the covariance == 0, so it does not need to be standardized to test that hypothesis.

If your factors are not exogenous (they have predictors), then the summary() only provides a test of the residual covariance.

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

http://www.uva.nl/profile/t.d.jorgensen

[Matheus Raszl]

Dear Terance,

Thanks for your reply.

I've tried your suggestion but I cannot see on the regression relation between the latent variables (PE,BI and SI) the pvalue of the significance of this relationship.

I am using summary(UNI_LV, standardized=TRUE), but we can see for example the correlation between PE ~ BI is 0.755 , but how can I disply the pvalue for it ?

Regressions:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
  PE ~                                                                  
    BI                0.917    0.043   21.532    0.000    0.755    0.755
  SI ~                                                                  
    BI                0.794    0.039   20.113    0.000    0.675    0.675

Best Regards

Matt

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To post to this group, send email to lav...@googlegroups.com.
Visit this group at https://groups.google.com/group/lavaan.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/929a6545-71c3-4188-8cfb-8857490fb72f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Terrence Jorgensen]

I cannot see ... the pvalue

It is the column "P(>|z|)"  (the probability of a more extreme z statistic)

for example the correlation between PE ~ BI is 0.755

only in the case of a simple regression slope is the standardized slope equal to the bivariate correlation.  

[Matheus Raszl]

I cannot see ... the pvalue

It is the column "P(>|z|)"  (the probability of a more extreme z statistic)

But how could I see the p value in a more classical way ? Or at least find this variable and calculate by myself ? For example, p < 0.05 * , p < 0.01 **, p < 0.001 ***

for example the correlation between PE ~ BI is 0.755

only in the case of a simple regression slope is the standardized slope equal to the bivariate correlation. 

So, what lavaan is showing me when I force the path between two latent variables? Covariance or Correlation  ? I am sure this is not a loading.

Thanks in advance.

Matt

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To post to this group, send email to lav...@googlegroups.com.
Visit this group at https://groups.google.com/group/lavaan.

To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/0e877a61-58fb-46ff-8eb5-e60490375540%40googlegroups.com.

[Yves Rosseel]

On 7/6/19 2:26 PM, Matheus Raszl wrote:
> But how could I see the p value in a more classical way ? Or at least
> find this variable and calculate by myself ? For example, p < 0.05 * , p
> < 0.01 **, p < 0.001 ***

If the z-statistic is, say 21.532, then the p-value is computed as

z <- 21.532
pvalue <- 2 * (1 - pnorm(abs(z)))
pvalue

which will be (as good as) zero.

> So, what lavaan is showing me when I force the path between two latent
> variables?

The "~" operator defines a regression. The value in the 'Estimate'
column is a regression coefficient.

The "~~" operator defines a covariance.

To see the value of a correlation, you need the "~~" operator, in
combination with summary(, standardized = TRUE).

To see the p-value for this correlation, you need the
standardizedSolution() function.

Yves.