Polling surveys have a number of limitations. The sample sizes (see Annex A for details) limit meaningful subgroup analysis and both surveys are based on non-probability research panels, which means representativeness is achieved using quotas. Furthermore, the nature of online research inherently excludes those who do not have internet access. These data sources nevertheless provide useful and timely information, as long as appropriate caveats are applied.
As each of the surveys has been conducted using a quota sample, rather than a random probability sample, statistical significance can only be used on an indicative basis. Differences over the survey waves are only highlighted if they are likely to be significant, however not all significant differences have been described.