The foundation.
THE PRINCIPLE
Analysis is only as good as the records beneath it.
WHAT POOR DATA PRODUCES
Confident wrong answers.
Worse than no analysis, because people act on them.
WHAT AFFECTS QUALITY
- Completeness: are records missing?
- Accuracy: are values correct?
- Consistency: is the same thing recorded the same way?
- Timeliness: is it current?
- Uniqueness: are there duplicates?
WHAT TO CHECK REGULARLY
Record counts against expectations Totals against the source system Unique values in categorised columns Missing values by column
WHERE POOR QUALITY COMES FROM
Manual entry without rules Several people entering differently Systems not connected Nobody checking
WHAT FIXES IT
Rules, selection lists, and periodic checks.
WHAT TO DO BEFORE ANY IMPORTANT ANALYSIS
Verify the data first.
WHAT TO DO IF IT IS POOR
Say so, rather than presenting the result as reliable.
WHAT TO INVEST IN
Collection, before analysis.
