0
$\begingroup$

I have a conditional model such as:

y= b0 + b1*a + b2*previous1

Previous1 contains all the measurements of y(i,j-1). To present an example

id_number   a   y   previous1
0           12  0     .
0           13  0     0
1           9   0     .
1           8   1     1
1           4   1     1
2           5   0     .
3           6   1     .
3           4   1     1

So id number 0 has two measurements in y, both being 0. The previous1 variable has one measurement. As you can see this is a transitional conditional model based on previous measurements.

My question is, when I run this model, I get high insignificance:

            Estimate   Standard Error     pvalue
intercept   2.4         1.2                0.23
a           -0.011      0.024              0.31
previous1   40          11324              0.90

The Standard Error of previous1 caught my attention as it is way too high. My interpretation is that since the number of measurements per id is very small( less than 3) and some id's have only one measurement so the previous1 would be just a missing value, the standard error is very big. Therefore I attribute this high inaccuracy to a very small sample size and cluster size.

Is my interpretation correct?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.