Of the 4 methods of replacing missing data, which method ensures that the replacement number does not introduce any bias in the data with respect to the average scores of the entire dataset? Replace missing value with a value from a similar respondent Replace missing value with an average of a subsample of the respondents with similar characteristics Replace missing value with an average of responses to similar questions Use average of the entire sample that answered the question as a replacement value