Exploiting the Welsh Agricultural Census

1
Exploiting the Welsh Agricultural Census

• Stuart Neil
• Agriculture and Rural Affairs Statistician
• Welsh Assembly Government

2
The Welsh Agricultural Census

• Run every June since 1866
• Statutory survey
• Broad coverage of agricultural activity
• Wide use of regional results

3
Changes

• Changes to the survey structure
• Changes to what users want to do with the data
• Code of Practice
• Data Protection Act
• Freedom of Information Act

4
Strengths of the Welsh Agricultural Census

• Broad coverage of farm holdings
• Broad coverage of agricultural activities
• Very long time series
• Consistent UK coverage
• Understood by farmers

5
Sources of error in the Agricultural Census

• Missing data
• Holdings not sampled
• Non-response
• No compulsory register of farms
• Not geographically specific

6
Missing data

• The Census is actually a sample survey
• Larger holdings are in the sample every year
• Smaller holdings sampled less frequently
• Response rates falling
• For most variables just under 50 of the total
  value comes from responses

7
Coverage of the statistical register of farms

• The register is not compulsory
• Problem areas
• Specialist horticulture holdings
• Specialist pig producers
• Specialist poultry producers
• The smallest holdings

8
Geo-referencing farm holdings

• A farm holding is not a compact geographical unit
• It is a disjointed set of fields
• Holdings may also use common land
• The centroid of a holding may be estimated to
  allow aggregation to flexible geographies
• This is an estimate

9
General conclusions

• Analyse strengths and weaknesses
• Be open about data usage
• Be prepared to negotiate for special cases
• Is releasing unfit data helpful?

10
And finally...

• May the gods grant you what you asked for