Missing value imputation: delta values
Posted: Mon Jan 10, 2011 1:53 pm
Dear all,
Last week, I tried to perform sensitivity analysis for missing values using the MLWin Macro. I run random intercept models on proportions as the dependent variable. I have a cross-national data set with 17 countries and 25000 respondents, 3500 of which have a missing value on income. Hence, I want to assess how sensitive my models are.
At first, all goes well according to the outlines of the pdf-document available on the website.
However, when I run the si macro, I can only explore values of delta between -0.3 and +0.2. Otherwise (with higher or lower delta's) I get the following error:
"error while obeying batch file sens_analysis/calc_weights.txt at line number 22: calc c803= expo(-c803)
Numeric error(s) in calculate command. Affected entries replaced with system missing.. See output window for details."
Could anyone tell me what I might be doing wrong? Are there ways to fix or circumvent this issue?
Last week, I tried to perform sensitivity analysis for missing values using the MLWin Macro. I run random intercept models on proportions as the dependent variable. I have a cross-national data set with 17 countries and 25000 respondents, 3500 of which have a missing value on income. Hence, I want to assess how sensitive my models are.
At first, all goes well according to the outlines of the pdf-document available on the website.
However, when I run the si macro, I can only explore values of delta between -0.3 and +0.2. Otherwise (with higher or lower delta's) I get the following error:
"error while obeying batch file sens_analysis/calc_weights.txt at line number 22: calc c803= expo(-c803)
Numeric error(s) in calculate command. Affected entries replaced with system missing.. See output window for details."
Could anyone tell me what I might be doing wrong? Are there ways to fix or circumvent this issue?