Quality Assurance for survey data
Background
In producing our Official Statistics for the Gambling Survey for Great Britain and the Young Persons Survey, the Commission uses research suppliers to help with survey collection and reporting. We have high standards, and our suppliers are required to adhere to thorough quality assurance checks, ahead of supplying data to us.
Data supply checks
In addition to the research supplier’s quality checks on the dataset, the Commission also conduct their own quality assurance process. These checks include:
1. High-level checks - These checks run across the entire dataset and include the following:
- download the datafile to a secure area and check file received is for the expected date/timeframe
- check the correct number of fields exist in the datafile
- check the number of cases is correct
- check all fields are labelled correctly and makes sense, and
- check the content in any additional cross tabulations that have been supplied.
2. Question checks - These checks look at the answers to specific questions in the survey and include:
- ensure the data contained within the fields is where we would expect it to be, for instance, check those questions applicable to all respondents, have been answered by all respondents, and
- check that questions that require a selection, only contain the correct range of answers and these are correctly labelled.
3. Sense checks - This section focuses on whether the data makes sense in the context of its use, and include the following:
- the weight variable is present and can be applied to the data
- demographics and/or selection criteria have been met as per the research specification
- basic analyses show sensible responses in line with previous waves and expectations, and
- any unlikely figures are queried with data supplier.
4. Final data checks - A final set of checks are undertaken during the analysis phase and ahead of publishing the Official Statistics. These include the following:
- check if any further issues arise during the analysis phase, that is, do the cross tabulations make sense or does anything not look quite right, and
- follow an additional quality assurance process for the reporting and outputs phase, which covers items such as formatting of publications, charts and tables clearly labelled, and so forth.
Status of official statistics Next page
Revisions and corrections policy
Last updated: 29 February 2024
Show updates to this content
No changes to show.