Different respondent numbers in same model?
Posted: Thu Apr 20, 2023 2:58 pm
Hello,
I have a question on using constructs and indicators with different respondent samples.
I would like to run a model where some indicators and constructs were only followed up with people who had answered 'yes' to a previous question. For example, one construct is knowledge about 'peanut butter', and then I have a different construct related to for example eating habits towards peanut butter- obviously only those who eat peanut butter would be exposed to this construct. This would mean that for some indicators and constructs I have say, 350 responses but for 1 or 2 constructs, I only have 200. Does the model treat these gaps as 'missing values'? Would this then only be possible if the difference or 'missing values' is only of a certain percentage? Can I even run a model with different sample sizes?
For any insights I would be ever so grateful!
Thanks very much!
I have a question on using constructs and indicators with different respondent samples.
I would like to run a model where some indicators and constructs were only followed up with people who had answered 'yes' to a previous question. For example, one construct is knowledge about 'peanut butter', and then I have a different construct related to for example eating habits towards peanut butter- obviously only those who eat peanut butter would be exposed to this construct. This would mean that for some indicators and constructs I have say, 350 responses but for 1 or 2 constructs, I only have 200. Does the model treat these gaps as 'missing values'? Would this then only be possible if the difference or 'missing values' is only of a certain percentage? Can I even run a model with different sample sizes?
For any insights I would be ever so grateful!
Thanks very much!