May I ask a further question? In linear regression, if we use listwise deletion, would the models by stepwise, forward selection or backward selection be different?
In listwise, you should not worry about analysis/outcomes, whether linear regression or hierarchical regression or logistic regression. The basic idea is to remove the entire row/respondent data that is affected by one or more missingness. Just an addition - you should be fine if the data is missing at random (MAR), ex, due to mistake omission by respondents. BUT You may face a problem of low statistical power which leads to invalid conclusion , if the respondent(s) made the omission intentionally (maybe due to the fact that you asked sensitive or ambigious question).
In listwise, you should not worry about analysis/outcomes, whether linear regression or hierarchical regression or logistic regression. The basic idea is to remove the entire row/respondent data that is affected by one or more missingness. Just an addition - you should be fine if the data is missing at random (MAR), ex, due to mistake omission by respondents. BUT You may face a problem of low statistical power which leads to invalid conclusion , if the respondent(s) made the omission intentionally (maybe due to the fact that you asked sensitive or ambigious question).
