COMPARE DATASETS is a very useful procedure when one is concerned about quality assurance on data entry, inter coder/judge comparison, time period comparison, etc.
I suggest that SPSS add 2 features to COMPARE datasets.
(yes I have workarounds).
(1) add a column with a count of number of variables with mismatches. This would be the 4th column after "ID variable" "active" and "compare".
(2) optionally add a fifth column with further ID information.
For example, when double entering data the first column tells what pair of records goes together.
This fifth column would be the source variable for who did the coding/entry. This would be useful when several sources would enter/code the data or there were two time periods etc.
------
The workarounds are to use LAG and other syntax to get the count of mismatches and add this as a variable to the second record for each pair. The first record of a pair would have a fixed value say -1.
(a) compare the datasets using whatever variables are of interest. Save the dataset with mismatches.
(3) in the syntax window concatenate the count variable and source variable to the front of the list of variables used in (a).
(c) apply the syntax used in (a) to the dataset with mismatches. Because the count and source will have different values for the members of the pair they become part of an integrated document.
Art Kendall
Social Research Consultants
Locked
