The challenge
An education research agency needed a national survey of parents to show how they choose schools. National parent survey weighting was central to making the results reliable. Online samples over-represented some regions and higher-income families, which risked skewing results on fees, distance and school reputation.
This case study shows how national parent survey weighting can correct those imbalances without hiding how the data was collected. The aim was simple: results that a reader could trust and the agency could defend.
Our approach to the national survey of parents
We recruited to soft quotas, then weighted the final sample to national targets and documented every step. This kind of clear record supports the transparency that the ICC/ESOMAR International Code expects of researchers.
- Recruited parents to soft quotas by region, child age and school type
- Checked the unweighted profile against national population targets
- Applied rim weighting with capped factors across region, income band and school type
- Delivered weighted tables with weighted and unweighted bases side by side
Why weighting survey data matters in a parent survey on school choice
Any online survey of parents will attract some groups more readily than others. Parents in certain regions may respond faster, and families with higher incomes are often easier to reach online. As a result, the raw sample can look quite different from the national population of parents with school-age children.
That matters because school choice depends on factors that vary by place and by household. Fees, travel distance and reputation do not carry the same weight for every family. If the sample leans towards one group, the headline findings lean with it.
Weighting survey data brings the sample back into line with known population targets. However, it works best as a second step. Good market research fieldwork aims for a balanced sample first, and weighting then corrects the remaining gaps.
Soft quotas and weighting working together
Soft quotas guide recruitment towards the right mix of parents. They do not close the survey to people who do not fit a cell. This keeps the sample flexible and avoids turning away useful respondents. For a parent survey on school choice, quotas by region, child age and school type give a sound start.
Weighting then deals with whatever imbalance is left. Because the two steps work together, the weights stay moderate. As a result, the findings rest on real responses from each group, not on a few heavily weighted cases.
How rim weighting works
Rim weighting, which some researchers call raking, adjusts the sample to match target totals on several characteristics at once. The process runs in rounds. Each round adjusts the weights to fit one characteristic. The rounds repeat until the sample matches all the targets as closely as possible.
This suits a national parent survey well. The targets sit on separate variables such as region, income band and school type. Rim weighting does not need you to know the full combination of every group in advance.
Capping the weights
Capping weight factors stops any one respondent from having too much influence. Without a cap, the weights could scale up a small, under-represented group too far. A handful of answers could then drive the result. A cap trades a little precision on the targets for more stable results, which is usually the better choice.
Reporting weighting efficiency
Weighting reduces the effective sample size, because some respondents count for more than others. Weighting efficiency summarises that loss. Reporting it lets the agency explain how much the weights affect precision. It also shows readers that the agency has hidden nothing.
Reading weighted survey tables
Clear weighted survey tables show both the weighted and unweighted bases. The weighted base shows the result as it applies to the population. The unweighted base shows how many parents actually answered. This tells the reader how much confidence to place in each figure.
Showing the two side by side lets analysts spot small groups quickly. For example, if a school type has a low unweighted base, the agency can treat differences there with care. It can say so in its report.
Good practice for national parent survey weighting
- Choose targets from a reliable and recent national source, and record which one you used
- Compare the unweighted sample with the targets before applying any weights
- Weight only on variables that relate to the questions being asked
- Cap the weight factors and report the effect on the effective sample size
- Keep weighted and unweighted bases visible in every table
- Document each step so another researcher could repeat the work
The results
- Results that represent parents across every region
- Weighting efficiency reported, so the agency could explain the numbers
- Weighted tables ready for the agency's report
The agency could present its findings on how parents choose schools. It also had a clear account of how we built and adjusted the sample. Readers could see what the data showed, how we weighted it and how much weight to give each finding.
Illustrative scenario based on typical engagements. Client details are anonymised.


