chapter 7 Evidencing Culture for Policy
Two reports on the relationship between being an artist or working in a creative occupation and well-being
The two reports we will turn to were published in subsequent years. Their titles and their named approaches suggest that they both contain findings from research using similar methods to answer a similar research question about the well-being of ‘creatives’. This enables us to see how ‘culture’ can be operationalised as being and working as an artist, and how this can relate to well-being. It also continues to allow us to familiarise ourselves with looking at others’ research as it appears in reports, and to think more about what might be happening under the bonnet.
Report 1, Artful Living: Examining the Relationship Between Artistic Practice and Subjective Wellbeing Across Three National Surveys was funded by the National Endowment for the Arts in the US . The research looked at different cohorts of arts practitioners and graduates in the US, using three different surveys. Contrary to the received wisdom that music and the performing arts are associated with the largest increases in well-being , Tepper et al. found that fine arts and crafts consistently related to higher well-being; music did so for some groups and not others; and participating in theatre ‘seemed unrelated to wellbeing’ in the data they had on arts practitioners and graduates . Overall, the authors say that there was ‘strong support’ that what they call ‘artistic practice’ is associated with higher life satisfaction and lower anxiety, as aspects of subjective well-being.
Report 2, Creative Occupations and Subjective Wellbeing is a working paper for NESTA, a UK Thinktank. This study used data from the UK’s Annual Population Survey (APS). This research concurs with Tepper et al.  that creative occupations are associated with higher than average life satisfaction, worthwhileness and happiness, ‘although most creative occupations also have higher than average levels of anxiety’ . This is contrary to Tepper et al.’s findings on anxiety from their data, but is corroborated by a number of studies, including the recent book, Can Music Make You Sick? .
We are going to break down the ways that these studies may seem similar, yet differ. Both Tepper et al. and Fujiwara et al. use multiple regression of cross-sectional ‘national survey’ data that ask subjective well-being questions from people with an artistic practice in the case of the US or a creative occupation in the case of the UK. This means that these data include variables based on questions asked by the organisations who administer the survey; the named researchers (or authors) don’t ask these questions of the participants themselves. Some of the datasets used include creative practitioners and people who are not creative practitioners. This is fairly common, and the researchers simply distinguish which cases (people in the data) meet the criteria of their research question, meaning they analyse the people who have a creative occupation/artistic practice and remove those who do not form from the model.
Box 7.5 Multiple Regression and Cross-Sectional Data
What is multiple regression of cross-sectional data?
Let’s look at these separately.
Regression analysis is common in statistical analyses. It involves estimating the relationship between a dependent variable and one or more independent variables.
In an analysis (e.g. a regression) you distinguish between
(1) Independent variables: that can take different values. You use an independent variable to predict the dependent variable. That is why it is sometimes called a predictor variable.
(2) Dependent variables: that can take different values. When you are measuring your relationship, you are interested in how the dependent variable is affected by the independent variable. It is, therefore, sometimes called the outcome variable to reflect this.
Say you are interested in private music tuition in childhood and creative occupations. You are not expecting an adult professional occupation to retrospectively generate experience of music lessons, but you might want to understand if the opportunity of private tuition affected a later career. So, occupation would be your dependent variable, and music tuition in childhood would be your independent variable.
So, we are still interested in private music tuition in childhood and creative occupations. We have established we want to understand how the first affects the second (and not the other way around). You might decide on other things that you think predict being a creative, such as gender, which previous research may suggest affects the likelihood of entering a creative occupation. Therefore, you would bear this in mind as another possible independent variable.
This is what makes it ‘multiple’ because we have now got more than one independent variable to predict our dependent variable.
A regression to explain how many people work in creative occupations could be conducted with either cross-sectional or longitudinal data.
Cross-sectional data are collected from a survey from a specific point in time, or time period. The same survey questions can be repeated, but these questions will have been asked from different people.
Longitudinal data hold information on the same people over
time. This means you can ask the same questions, year on year, to see change over time. For example, you can ask people year on year if they have private music lessons. You can also have data for different questions. This is useful for our example, as we might have data on private music tuition in childhood, and data on occupation in adulthood, should the participant be around that long.
DCMS’ Taking Part Survey (TPS) has a longitudinal component and a cross-sectional one.
Since 2005/2006, TPS has been run on a cross-sectional basis that involves a new sample of households, which is drawn annually, and a new group of respondents who are asked the same questions. This enables researchers who use this data to say ‘last year X% of the population had music lessons’. But it cannot, therefore account for change that happens to an individual, so you won’t know that ‘the people who stopped music lessons last year are like abc’. Given that change implies impact, this is a big deal for many of the studies we encounter in this book.
The two research projects on well-being from the US and UK that we are exploring use different samples and surveys. This means that in both studies the group in ‘creative occupations’ may not necessarily map onto those with an ‘artistic practice’ as neatly as the labels used suggest. We come back to this in the next paragraph. The UK report uses the Annual Population Survey, which contains information on people’s occupation and the ‘ONS4’ questions that we keep encountering. Creative occupations were defined using DCMS’ Creative Industries Economic Estimates  and then coded using the ONS’s standard occupational classifications, called SOC codes . The authors are therefore able to look at the four ONS measures: life satisfaction, worthwhileness, happiness and anxiety for the 30 creative occupations as defined by the DCMS  However, when you imagine a town planning officer, they probably feel quite different to you from a musician. Also, realistically, the day-to-day duties of one is likely to feel very different than the other. A town planning officer will probably have more regular hours and a more secure contract than a cellist. You might also imagine that a cellist may have more capacity for self-expression, and feeling, well, artistic, than a town planner. The differences in day-to-day tasks, security, income and so on are all important external factors that will affect well-being. Therefore, these discrepancies across creative occupations (some of which may not feel that creative) may limit improved understandings of the impact these professions have on well-being, if the model treats everyone with a job defined as ‘creative’ (using occupational codes) as equivalent. What is key here is that it is that the categories used to break down the data (from the APS), and how they have been coded into professions (using the ONS’ occupational classifications) is important context to knowing what we can understand about differences in well-being.
In contrast, the US case uses data from three surveys which target different groups. The Strategic National Arts Alumni Project (SNAAP) captures data about graduates of arts institutions. The Double Major Student Survey focusses on undergraduates who have two majors from four comprehensive institutions and five liberal arts colleges. The DDB Needham Life Style Survey (DDB) is the nation’s largest and longest running annual survey of consumer attitudes. The report states that the researchers ‘look specifically at responses to creative practice, life satisfaction, and “sense of control” in one’s life’, but it is not precisely clear whether they identified ‘creatives’ or looked at everyone who answered these questions. The participants across these surveys are classified as ‘having an artistic practice’ for different reasons. In fact, most of the secondary data analysis is of responses regarding how people do cultural activities in their spare time.
Crucially, and confusingly, the participants across the three surveys do not all actually have an ‘artistic practice’, in a professional sense. In fact, the authors ‘use the terms artistic practice, creative engagement, and creative practice interchangeably throughout this report’ . So, there is no analysis of the relationship between well-being and creative occupations, per se, or necessarily any differentiation between a professional artist or an amateur who ‘engages’ in artistic practice. Similarly, the questions used to establish aspects of subjective well-being are not the same across each survey. Table 7.4 shows the subjective well-being questions and how the ‘artists’ were identified across the three US surveys, alongside the UK case. Therefore, establishing what counts as ‘an artistic practice’ is one of the issues, and the other is establishing how subjective well-being is understood. There are therefore key differences in how these concepts were operationalised in these reports.
Table 7.4 A comparison of culture and well-being questions across the four surveys used in the two case studies
|Survey name||Description of survey||Application of the survey||Culture Q||Subjective well-being question evaluative, experience/eudaimonic?|
|DDB Needham Life Style Survey (DDB)||The DDB Needham Life Style Survey (DDB) is the nation’s largest and longest running annual survey of consumer attitudes.||In polling American adults, the surveys ask questions about— among other things— attitudes, interests, opinions, activities, product use and mass media use.||Three specific questions address creative practice, including the frequency of participation in craft projects, gardening and playing a musical instrument over the last 12 months.||SWB Q: EVALUATIVE|
A series of agree/disagree statements get at the issues of life-satisfaction (e.g. ‘I’m much happier now than I ever was before’; ‘I am very satisfied with the way things are going in my life these days’). SWB Q: EXPERIENCE
To get a sense of generalised anxiety (‘loss of control’), we examine several questions that address people’s sense of personal efficacy (e.g. ‘sometimes I feel that I don’t have enough control over the direction my life is taking’).
|Double Major Student Survey||The survey, supported by the Teagle Foundation, assesses the link between creativity, interdisciplinarity and the liberal arts by focussing on undergraduates who have two majors.||The survey drew from a sample of approximately 1700 students from four comprehensive institutions and five liberal arts colleges, and asked them questions about demographics, academic choices, self-ratings on skills and competencies, and creativity and innovation.||Students were also questioned about their participation in artistic and creative practices, including ‘played a musical instrument’, ‘painted, drew a picture, or made sculpture’ and ‘made or designed clothing, costumes, etc.’ There were a total of 10 different categories of artistic and creative practices listed among the 23 activities. Students were asked to rate the frequency with which they participated in these activities.||SWB Q: EUDAIMONIC Specifically, students were asked about their positive self-image (‘please check all of the adjectives that best describe yourself’— ‘capable’, ‘confident’, ‘resourceful); their positive social outlook; and materialistic orientation (e.g. ‘it sometimes bothers me quite a bit that I can’t afford to buy all the things I’d like’).|
|Strategic National Arts Alumni Project (SNAAP)||The Strategic National Arts Alumni Project, or SNAAP, is an online survey targeted at graduates of arts institutions, which asks questions about their experiences both during and after their arts schooling.||To date, more than 100,000 alumni have been asked questions about their career path, their artistic practice (both professionally and avocationally) and their overall satisfaction with work and life. Specifically, we look at questions from the 2009 pilot survey of 4031 graduates from across 76 different arts colleges and schools.||Questions addressing personal artistic practice and the frequency with which it is undertaken. SNAAP data allow us to look at people who were once highly involved in the arts through their schooling or career, and who are no longer practising their artistic craft or are only practising it avocationally. This may reveal some information about the importance of continued artistic practice for those who valued it highly in the past and who had achieved high levels of proficiency.||SWB Q: EVALUATIVE|
Including people’s response to the questions, ‘in most ways my life is close to my ideal’ and ‘I am satisfied with my standard of living’.
|Annual Population Sur vey (APS)||The UK’s APS covers employment, unemployment, housing, ethnicity, religion, health and education.||The APS is a repeated annual cross-sectional survey of approximately 155,000 households and 360,000 individuals. Since 2011 the APS has contained the four ONS well-being questions. Waves (years) 2011–2012 and 2012–2013 are used in the analysis.||The jobs variables relate to the main job of the individual. They used the occupations as categorised by DCMS using NS-SEC (see Table 7.4).||ONS4: ‘Overall, how satisfied are you with your life nowadays?’|
ONS4: ‘Overall, how happy did you feel yesterday?’
‘Overall, how anxious did you feel yesterday?’
ONS4:‘Overall, to what extent do you feel the things you do in your life are worthwhile?’
Table 7.4 is populated with text that has largely been cut and pasted from the two reports. It contains contextual information on the nature and purpose of the surveys used (you will see that in most cases the surveys have different aims) and the wording of the questions. I have attempted to categorise the US study into Evaluative, Experience, Eudaimonic, as per the categories in Chap. 4 and Table 4.1. This was easy for the ONS4 from the UK case, as these have been categorised for us already. The US case proved more difficult. The question about what Tepper et al. call ‘positive self-image’, while not unrelated to well-being and anxiety, fell less neatly into our categories, as designated by Dolan et al. , the ONS or those recommended by the OECD .
‘So what?’ you may ask. Well, these two reports came out in subsequent years and with titles that imply they are researching the same relationship between culture and well-being. They may appear to have used a similar approach, listed as multiple regressions of cross-sectional data. However, there are key differences in the data they investigate. 1, they report on different countries; 2, one uses three data sources, the other uses one; 3, their operationalisation of the ‘cultural occupation/artistic practice’ variables are very different; 4, as are the operationalisations of subjective well-being; 5, those running the regressions (the modellers) used slightly different controls (see Table 7.5). There are numerous reasons for these differences, but mainly, remember that theories of what is good for well-being are not entirely universal, which will affect what someone wants to control for, but also the data are different, which will limit what it is possible to control for.
Box 7.6 Control Variables Controls are control variables
Say there was a positive relationship between older people and enjoying jazz music, and a negative relationship between younger people and enjoying jazz music. A study to see if there is an association between increasing funding for jazz music and enjoyment of jazz music may find no significant difference. The differences by age would be masked because the negative (younger people) relationship and the positive (older people) relationship could cancel each other out, resulting in no overall observable relationship.
Controlling for age can better establish that ‘funding jazz is likely to have a positive effect on enjoyment in older people, but not younger people’.
Table 7.5 Controls used in the two studies looking at well-being and creatives
|Controls used in the report Artful Living: Examining the Relationship Between Artistic Practice and Subjective Wellbeing Across Three National Surveys||Controls used in the report Creative Occupations and Subjective Wellbeing|
Place of residence
Children at home
Date of survey
When look back at Table 7.4, the survey questions generating the various forms of subjective well-being data are different. They do not use the same concepts of subjective well-being and the questions are not identically worded. The samples of creative practitioners appear to overlap conceptually at first, but they are far from identical. Therefore, we are not actually really looking at the relationship between identical things. Creative occupation or artistic practice do not strictly mean having a job that is creative in these studies, and the meanings and measures of subjective well-being are different in the data analysed.
Again, ‘so what?’ you may ask. Looking at the headline evidence together is the most typical way of understanding other people’s data analysis and findings to construct a body of evidence. Taking a moment to compare these two reports highlights how different two studies which may appear comparable really are, as well as the difficulties in finding conclusive answers to questions about the well-being of any particular group of people, and the role of culture—or work—or leisure—in this. Looking at differences in data sources, concepts, methodology, findings and motivations provides extra data that help establish how conclusions and headline findings may have been arrived at.
The studies differ in numerous ways: the questions asked, who was asked (or included), the nature of the sample—as well as the interpretation of what being creative involves. Furthermore, the research designs were analysing different subjective experience contexts: different places, and different relationships to creative cultural engagement (e.g. professional or amateur). The two reports were also commissioned by different organisations in different countries with undoubtedly different research agendas. Therefore, while in principle, these two studies are looking at the same social issue in the same ways, they have different research questions that are applied to different contexts.
While the two studies were not designed to test each other, the two headline findings can be used together in a literature or evidence review to make a statement about what is known about being a ‘creative practitioner’ well-being. Notably, the UK case states: ‘[t]o our knowledge this is the first quantitative study that specifically analyases the connection between creative jobs and wellbeing’ . The US case notes that ‘[a]s of yet, no one has examined the complicated relationship between creative practice and wellbeing within the US’ and ‘preliminary work has failed to demonstrate a robust relationship between creative practice and wellbeing in part because of limited sample sizes’ . Interestingly, neither of these reports seems to have been cited much. When they are cited, for example by Tiller , the positive impacts tend to be reported. Also Tiller  reports on the benefits of ‘artistic practice’ as cultural participation, rather than being an artist, and others interpret Tepper et al.’s results as follows:
Researchers have found that the more individuals participate in artistic activity, the higher they score on a variety of wellbeing.(Kemp et al. 2018, 1)
Tepper et al. (2014) found that creating crafts, gardening, and playing a musical instrument—or personal art‐making—were positively related to life satisfaction.(Kemp et al. 2018, 3)
Part of the nonsignificant relationship between active arts participation and life satisfaction may be due to a perceived lack of time individuals feel they have to engage in creative practice. Hence, if they feel that time is constrained such that they do not have sufficient time to engage in artistic creation, benefits related to SWB may be minimal.(Kemp et al. 2018, 6)
This final point is of interest, as neither Tepper et al. nor Kemp et al. really pick up on the fact that it may not be that those engaged in active arts participation, as described, do not have enough spare time to do enough creative practice, but instead, that they could be—like our friend in the Disney movie—dissatisfied with the job they have. Tepper et al. say that it may be better for some graduates to walk away from their artistic practice , but leaving ‘the industry’ seems to be attributable to a lack of time for ‘robust artistic life’ versus ‘simply dabbling in the arts’. This analysis does not incorporate what we know of the hardships of those who are full-time artists and those who are still aspiring . Given that the authors state: ‘this report represents an initial exploration of the thesis that the arts are essential to a high quality of life’ , we might question whether they were ready for an interpretation of the arts and their labour markets as bad for well-being in various ways.
Tepper et al.’s title Artful Living: Examining the Relationship Between Artistic Practice and Subjective Wellbeing Across Three National Surveys was misleading to some audiences, particularly in the UK, where artistic practice tends to mean working as a professional artist. Instead, it was more broadly defined to include practising an art as a hobby. Similarly, not all the creative occupations in Fujiwara et al.’s report were as closely aligned to having an artistic practice as you might assume by the term creative occupation. Ultimately, it can be more difficult to compare or synthesise studies than is obvious by the title of a report, or its headline findings. This is often not acknowledged and can limit the validity of comparisons when evidence is reviewed and synthesised.
The way that the idea of culture and well-being are operationalised in these two cases differs more than to be expected: the data and the contexts in which they were collected, or the surveys or questions which generate the variables, are not always as similar as might be assumed. When we describe findings from apparently comparable studies, it is just as important to account for the motivations and methods of these studies (their contexts) as it would be our own. This is because when we synthesise the research of others, we create new knowledge that is able to make grander claims as it appears more generalisable.
- Tepper et al. 2014
- The three national surveys were the DDB Needham Life Style Survey (DDB), the Double Major Student Survey and the Strategic National Arts Alumni Project (SNAAP). Full details of sampling can be found in the report.
- e.g. Fujiwara and MacKerron 2015
- In this book the spelling of well-being is used, unless it is a direct quote, and then the spelling of the author is used.
- Tepper et al. 2014, 7
- Confusingly, what is called the Annual Population Survey is actually not one survey, but a conglomerate of other surveys, as explained in Table 7.4.
- Fujiwara et al. 2015, 1
- Gross and Musgrave 2020
- See TNS 2011 for more information on the longitudinal element.
- DCMS 2011
- ONS 2010b
- This is detailed in the report, however, for more explanation on SOC codes and the cultural sector, please also see Oman (2019).
- 2011)) in Table 7.3.
Table 7.3 Occupations in the creative industries
Creative industry Creative occupations Advertising and marketing Description
Marketing and sales directors
Advertising and public relations directors
Public relations professionals
Advertising accounts managers and creative directors
Marketing associate professionals
Town planning officers
Chartered architectural technologists
Architectural and town planning technicians
Crafts Smiths and forge workers
Weavers and knitters
Glass and ceramics makers, decorators and finishers Furniture makers and other craft woodworkers
Other skilled trades not elsewhere classified
Design: Product, graphic and fashion design Graphic designers Film, TV, video, radio and photography Arts officers, producers and directors
Photographers, audio-visual and broadcasting equipment operators
IT, software and computer services Information technology and telecommunications directors
IT business analysts, architects and systems designers
Programmers and software development professionals
Web design and development professionals
Publishing Journalists, newspaper and periodical editors Authors, writers and translators Museums, galleries and libraries Librarians
Archivists and curators
Music, performing and visual arts Artists
Actors, entertainers and presenters
Dancers and choreographers
Adapted from DCMS (2011)
There are many discussions over what counts as a creative occupation using these classifications that we won’t get too caught up in here.((A prominent recent example is Campbell et al. (2017): one of the biggest problems the author identify is the disproportionate role of IT.
- Tepper et al. 2014, 8
- As described in Chaps. 2 and 4, Eudaimonia is most often understood as purpose or flourishing.
- 2011a, 2011b
- OECD 2013; Smith and Exton 2013
- Fujiwara et al. 2015, 2
- Tepper et al. 2014, pp. 8, 10
- Google scholar searches show that Tepper et al. has been cited 15 times, and Fujiwara et al., 17 times. However, of course, that does not include all the non-academic places where these reports are cited.
- 2014, 43
- Tepper et al. 2014, 28
- refer to Brook et al. 2020 for discussion on this