if a questionnaire with 5 questions is randomly missing 10% of the data, then on average about 41% of the sample will have at least one question missing.Īlso, it is often the case that the missing data is not randomly distributed. This problem is bigger than might first be evident. In this case, additional sample data elements may need to be collected.
This is particularly relevant when the reduced sample size is too small to obtain significant results in the analysis.
One problem with this approach is that the sample size will be reduced. In Identifying Outliers and Missing Data we show how to identify missing data using a data analysis tool provided in the Real Statistics Resource Pack.Ī simple approach for dealing with missing data is to throw out all the data for any sample missing one or more data elements. For example, in conducting a survey with ten questions, perhaps some of the people who take the survey don’t answer all ten questions. One problem faced when collecting data is that some of the data may be missing.
0 kommentar(er)
