Scotland’s Census Quality Survey 2011 Neil Jackson General Register Office for Scotland
Jan 12, 2016
Scotland’s Census Quality Survey 2011
Neil JacksonGeneral Register Office for Scotland
Agenda
BackgroundOverview of CQSUse of IT in 20112011 survey design2009 rehearsal experienceSummary
Why have a quality survey?
Self-completion census
Provide measure of reliability
Feed into future censuses
Format of CQS
Post enumeration survey
Samples households that have responded to census
Re-asks census questions and compares results
Asks reason for any discrepancies
You’ve just told me that you have a mortgage, but in the census you said you own the house outright. Which was correct on census day?
You’ve just told me that you have a mortgage, but in the census you said you own the house outright. Which was correct on census day?
I have a mortgage. I must have misunderstood the question on the census.
I have a mortgage. I must have misunderstood the question on the census.
Correct response: Owns with a mortgage
Reason for difference: Misunderstood question
Example 1
You’ve just told me that you are retired, but in the census you said you were working full time. Which was correct on census day?
You’ve just told me that you are retired, but in the census you said you were working full time. Which was correct on census day?
I was working on census day, but I have since retired.
I was working on census day, but I have since retired.
Correct response: Full-time employee
Reason for difference: No difference
Example 2
Output
Quality report giving accuracy rate for each question
Some examples from 2001:sex 99.9%hours worked 80.2%qualifications 60.6%number of rooms 77.5%
Use of laptops
Lightweight mini-laptop
Preloaded with relevant census data
All data encrypted
Microsoft Access application
Regular uploads to census HQ
Sample screenshot
Example 1: Same response
Example 2: Equivalent response
Example 3: Different response
Sample design
1000 householdsNot randomThree areas of different types:– Urban– Rural– Semi-rural/market town
Early census returns onlyResults still useful
Fieldwork design
3 teams of 7 interviewers (including manager)50 households eachAssistant interviewersInterview takes up to 1 hourHousehold and person questions
Lessons from rehearsal
Carry out as soon as possible after census day– Memory effect– Movers
Don’t ask about overnight visitorsUpload from home via broadbandSecurity features must be unobtrusive
Some results from the rehearsal
Self-contained accommodation 98.7%
Country of birth 99.5%
Number of rooms 66.0%
Nature of disability 75.6%
Summary
CQS gives measure of accuracy levels
Use of IT in field
1000 households in selected areas
Rehearsed in 2009