Accuracy and reliability

Overall accuracy

As with all surveys of random samples, a degree of statistical error must be expected. This error may be stochastic or systematic, and is the cause of the process of drawing the random sample and the structure of non-response. Non-response occurs when a respondent does not answer the survey.

Non-response increases the systematic error of the survey, as the likelihood of non-response varies across the persons comprising the sample. I.e. certain groups, e.g. young people and non-ethnic Danes, have a higher rate of non-response, which negatively impacts the representativeness of the survey. This is to some degree remedied by the weighting of the sample through the known population values of demographic variables. Even though this weighting handles much of the systematic nature of the non-response, it cannot be ruled out that systematic non-response still affects the final estimates.

Non-response increases the stochastic error of the survey in two ways. Firstly because the stochastic error is in a direct way a product of the amount of responses. Secondly and less directly, because the systematic error arising from systematic non-response is corrected through weighting, which in turn can increase the stochastic error. The weighting uses less information from groups with many respondents and more information from groups with few respondents.

Other sources of errors, eg. memory lapses are also relevant for the Cultural Habits Survey. The respondents are asked about their cultural habits over the past year. Over such a long frame of reference it is possible that some respondents are unable to remember their objective behavior. This may result in errors, or in responding not from a direct recollection of their behavior, but instead from a perception of their typical behavior. Such responses based on self image may be more subjective.

Sampling error

Stochastic error is quantified as standard errors, which express the uncertainty of the estimated means. The stochastic error is affected by the size of the sample and by the weighting process. The error is roughly halved by quadrupling the sample size. The more completed interviews the survey's results are based on, the less stochastically uncertain the results are. The greater the non-response, and the greater variation in weight sizes arising from it, the greater the stochastic uncertainty. This fact arises because the weighting uses less information from groups with many respondents and more information from groups with few respondents.

The standard errors for the proportions who have participated in each cultural activity within the past 12 months, as measured in the shared part of the questionnaire, is below one percentage point in the first two quarters of 2024. The standard error of the indicator 'have read or listened to books within the past 12 months' is 0.71 percentage points. This means that the proportion of those who have read or listened to books within the past 12 months with 95 pct. certainty lies within +-1.39 percentage points of the estimated proportion of 68 pct. The stochastic error does however increase when one considers groups grouped by demographic variables, since the results of each group in that case will be based on fewer responses.

Non-sampling error

To correspond as close as possible to the target population, the random sample is drawn from population data no older than three months. The risk of measurement error is reduced by the pilot testing of the questionnaire. Measurement error may however still arise as a consequence of i.a. memory lapses. Stochastic error is worsened by high non-response. The response rate is approximately a third of the sample and the non-response is systematic. The weighting of respondents partially compensates for the systematic error caused by this systematic non-response, but increases stochastic error. The weighting is based on the assumption that the utilised demographic variables explain all relevant variation in cultural habits.

Quality management

Statistics Denmark follows the recommendations on organisation and management of quality given in the Code of Practice for European Statistics (CoP) and the implementation guidelines given in the Quality Assurance Framework of the European Statistical System (QAF). A Working Group on Quality and a central quality assurance function have been established to continuously carry through control of products and processes.

Quality assurance

Statistics Denmark follows the principles in the Code of Practice for European Statistics (CoP) and uses the Quality Assurance Framework of the European Statistical System (QAF) for the implementation of the principles. This involves continuous decentralized and central control of products and processes based on documentation following international standards. The central quality assurance function reports to the Working Group on Quality. Reports include suggestions for improvement that are assessed, decided and subsequently implemented.

Quality assessment

The survey is relevant because it meets identified user needs and it is developed in cooperation with the users.

Accuracy is ensured through openness about metadata such as response rates, non-response, calculation of statistical error, as well as any revisions and through detailed information on data processing.

The survey is timely it is published two months after the end of the surveyed period and is released frequently, with both quarterly and yearly releases.

Certain parts of the statistic is comparable with the previous versions of the Cultural Habits Survey from 1964 to 2012. Comparability with the Cultural Habits Survey 2018-2023 is challenged by a change in temporal frame of reference, which from 2018-2023 was three months. However, the answer options for frequency in the current survey contains an answer option, which under certain assumptions makes it possible to calculate the proportion who have participated in a given cultural activity within the past three months. Additionally the questionnaire has been updated to better capture changes in cultural habits.

Data revision - policy

Statistics Denmark revises published figures in accordance with the Revision Policy for Statistics Denmark. The common procedures and principles of the Revision Policy are for some statistics supplemented by a specific revision practice.

Data revision practice

Only final figures are calculated.