chapter 8 Talking different languages of value
Taking Part Survey and the data on culture
‘Museums and Happiness’ includes findings from quantitative research that used data from the Taking Part Survey1. The report explains why TPS data were chosen over other surveys, based on the different variables available, sample size and so on.((If you are interested in more information on the differences across BHPS, TPS and Understanding Society at the time, and why they mattered, or indeed, want to see another example of how research like this makes decisions, do look at the original report.)) It points out that while this technique had been used before, it had not yet been used on this dataset.
We have talked about why TPS was established in Chap. 6: that it was part of an instrumental project by DCMS to address a need for evidence. Here, we are going to think about the data itself and the context in which it is generated. TPS covers England only, rather than the whole of the UK. It employs interviewers to go to people’s homes, if they agree, of course, and interview them face-to-face with a questionnaire. The survey questionnaire asks about all different types of activities. A script asks the interviewer to request the interviewee to think in great detail, and to be specific, for example:
Firstly, I would like you to think about all the walking you have done. Please include any country walks, walking to and from work or the shops and any other walks you may have done.
In the last four weeks, that is since [TODAY’S DATE MINUS FOUR WEEKS] have you done at least one continuous walk lasting at least 30 minutes?((All the questions outlined, where specifcally worded, can be found in UK Data Service (2009). You can find questionnaires for each year here: https://www.gov.uk/government/publications/adult-questionnairetaking-part-survey-2009-to-2010. It is worth being aware that TPS has a longitudinal element, which is an adapted questionnaire, as it wants to accommodate change. Therefore, the adapted questionnaire wants to also know ‘why the change?’ For example, ‘you say you have participated more or less in this than last year. Why do you think that is?’))
Were you to participate, you would be asked this, and more questions and clarifications, and finally whether your walking was ‘for the purpose of health or recreation (not to get from place to place)’. The questionnaire then asks the same questions about cycling, for example; and so on. All in, you would expect to be speaking to the interviewer for about 40 minutes. In Box 8.1 you can see the museum questions from 2009–2010 survey.
Box 8.1 The Museum Questions from the Taking Part Survey 2009–2010
The museum questions((The formatting of the questions does change slightly over time, as adaptations and improvements are made. Again this is from the 2009–2010 schedule, available at: https://www.gov.uk/government/publications/adult-questionnaire-taking-part-survey-2009-to-2010.)) were phrased slightly differently from those on cycling and walking and listed below:
During the last 12 months, have you attended a museum or gallery at least once?
1. Yes 2. No -1. Don’t know
In the last 12 months, have you attended a museum or gallery…?
1. In your own-time 2. For paid work 3. For academic study 4. As part of voluntary work 5. For some other reason -1. Don’t know
How often in the last 12 months have you been to a museum or gallery [in your own-time] [or] [as part of voluntary work]? 1. At least once a week 2. Less often that once a week but at least once a month3. Less often than once a month but at least 3 or 4 times a year 4. Twice in the last 12 months 5. Once in the last
12 months -1. Don’t know
To make everyone’s answers analysable, all responses are combined into one dataset. Alongside questions on activities are questions about personal characteristics, for example, income, how many people live in your household, age, marriage status, whether you have children and so on. These variables allow researchers to understand how many different types of people ‘take part’ in different activities. These data also allow DCMS to see whether the percentages of different groups of people participating in different activities go up or down over time. These are reported on by DCMS in ‘Statistical Releases’, in which the results are synthesised and available for anyone to access.((The statistical release page is currently in DCMS (2013).))They include information on, for example, the numbers of people who have participated in the arts in the last twelve months, and the same for sport, and so on. DCMS2. For example, in 2015((Focus on reports from 2015 can be found in DCMS (2015b).)) there were ten of these reports, including one on well-being3 and one on art forms4; in 2016, there was one on diversity5.
Box 8.2 Variables: A Reminder
A variable takes different values in different situations. These values vary between cases or observations (which in this case are people but aren’t always). They also vary over time or space.
So, for example, height varies across people, because some people are taller than others, but also within people over time, because people get taller as they grow up.
It is a variable because it varies. It is this change or variability that is measured, whether over time, or to compare characteristics.
In a regression, you would analyse the relationship between an independent variable, or independent variables, and a dependent variable.
Because we look at how variations in independent variables can predict values of a dependent variable,
• independent variables are sometimes called predictor variables,
• dependent variables are sometimes called outcome variables.
So, if you want to see the relationship between age and museum attendance, presumably, you are not expecting museum attendance to make someone age, but you might want to understand if older people are more likely to attend museums. Therefore, age would be your independent variable.
To measure the culture–well-being relationship, we need an independent variable (for culture) and if we wish to measure culture’s relationship with well-being, then we need the chosen well-being
variable to be the dependent variable. For ease, let us say because we want to see whether people who participate in culture have higher well-being.
*We also want to add other independent variables to make sure that we’re not inadvertently measuring other relationships as well. For example, if married people report higher well-being on average, and are more likely to attend cultural events, we should include marital status as an additional independent variable. Without accounting for it in the analysis, marriage could be a confounding variable, meaning it could exaggerate the results. Therefore, here marriage would be controlled for, even though it is not of primary concern in the outcome.
