0
$\begingroup$

I need help in interpreting the output for mixed linear model.The estimates of fixed effect shows the parameters of my level 1 predictor variables, but there are some zero values, and i'm not sure why or how to interpret it. Could it be due to lack of data points in ITEM16? Here's a snapshot of the output..

Parameter   Estimate    Std. Error  df  t   Sig.    95% Confidence Interval 
                        Lower Bound Upper Bound
Intercept   .184285 .016181 40.209  11.389  .000    .151588 .216982
[ITEM=1]    -.033855    .008585 29812.714   -3.944  .000    -.050681    -.017029
[ITEM=2]    -.020130    .008915 29812.824   -2.258  .024    -.037603    -.002657
[ITEM=3]    -.064286    .008792 29814.834   -7.312  .000    -.081519    -.047053
[ITEM=4]    -.077827    .009920 29817.388   -7.846  .000    -.097270    -.058384
[ITEM=5]    -.040264    .011120 29804.602   -3.621  .000    -.062061    -.018468
[ITEM=6]    -.031131    .008790 29815.299   -3.542  .000    -.048359    -.013903
[ITEM=7]    -.037671    .010160 29809.248   -3.708  .000    -.057585    -.017757
[ITEM=8]    -.023466    .010947 29806.310   -2.143  .032    -.044923    -.002008
[ITEMC=9]   -.025810    .011665 29807.427   -2.213  .027    -.048673    -.002947
[ITEM=10]   -.004509    .011803 29805.702   -.382   .702    -.027643    .018626
[ITEM=11]   -.073471    .009903 29816.904   -7.419  .000    -.092881    -.054061
[ITEM=12]   -.059247    .010095 29806.327   -5.869  .000    -.079033    -.039460
[ITEM=13]   -.038357    .014345 29810.276   -2.674  .008    -.066474    -.010240
[ITEM=14]   -.032134    .013288 29807.876   -2.418  .016    -.058178    -.006089
[ITEM=15]   -.046073    .010736 29808.929   -4.291  .000    -.067116    -.025029
[ITEM=16]   0   0   .   .   .   .   .
[...}.....
...
...

Thanks!

$\endgroup$

1 Answer 1

2
$\begingroup$

It looks like ITEM is a categorical factor variable, in which case the model is parameterized with a dummy or indicator variable for each distinct value of it, and if there is a fixed intercept in the model, the last one of these will be redundant or linearly dependent on the prior ones. That gets aliased to 0 in the formation of the generalized inverse formed during estimation. This is to be expected and does not indicate a problem.

If you have additional factor variables you'll see similar situations for them, and for interactions involving factors you'll generally see multiple parameters aliased to 0.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.