Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It only takes a minute to sign up.
Sign up to join this community
I did a binary logistic regression test on SPSS with a sample size of 24. The study is about the correlation of teenage pregnancy and depressive symptoms. Of the 24, 11 are pregnant teenagers. Of the 24 only 1 is positive for the presence of depressive symptoms. Both variables are recoded to nominal, so that 1 means yes and 0 means no.
What could the abnormally high standard error mean? And is there any way to solve it?
Hoping for kind responses. Thank you.
You have almost-separation, as Scortchi notes.
Your data - 11 pregnant teenagers, none of which are depressed, and 13 non-pregnant ones, one of which is depressed - is consistent with a model that essentially says that if you are pregnant, you have a zero chance of depression, whereas if you are not pregnant, you do have a small chance. Logistic regression does not play well with chances of zero or one, and one symptom of separation is large standard errors.
And of course, it's not as if there really were a zero chance of depression during pregnancy. You will simply need to collect more data.
With a point prevalence of about 3% for MDD, 24 participants are way too few, in any case, unless you screened explicitly for depression. This should have been taken into account during sample size calculation.
