Carlijn B. M. KamphuispersName, Tessa JansenpersName, Johan P. MackenbachpersName, Frank J. van LenthepersName (2015date). “Bourdieu's Cultural Capital in Relation to Food Choices: A Systematic Review of Cultural Capital Indicators and an Empirical Proof of Concept.” PLOS ONEtitle·j.
Object of study — quantitative measures of Bourdieuterm·key's cultural capitalterm·dom, and cultural capitalterm·dom, socioeconomic positionterm·dom and food choicesterm·key among Dutch adults
Research question — whether cultural capitalterm·dom differs between socioeconomic groups and relates to healthy and unhealthy food choicesterm·key
Contribution — a 56-item questionnaire measuring food-choice relevant cultural capitalterm·dom in its three states, with first empirical associations
phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | cultural capital as a possible explanation of socioeconomic inequalities in food choices — “Cultural capital may be a new, potentially powerful explanation for inequalities in food choices, which may lead to new entry points for the development of interventions to promote healthy food choices among low socioeconomic groups.” | |
| claim-type | method — “We developed a set of questionnaire items that capture cultural capital of the three different types in relation to food choices (see Table 2 ).” | |
| method — evidence | — “The questionnaire was empirically tested in a postal survey completed by 2,953 adults participating in the GLOBE cohort study, The Netherlands, in 2011.” | |
| object of study — evidence | — “We aim 1) to carry out a systematic review to identify existing quantitative measures of cultural capital, 2) to develop a questionnaire to measure cultural capital for food choices, and 3) to empirically test associations of socioeconomic position with cultural capital and food choices, and of cultural capital with food choices.” | |
| research question — evidence | — “The aim of the present paper was to develop a set of questionnaire items in order to investigate to what extent the possession of cultural capital differs between socioeconomic groups, and to what extent cultural capital relates to healthy and unhealthy food choices.” | |
| contribution — evidence | — “We developed a set of questionnaire items that capture cultural capital of the three different types in relation to food choices (see Table 2 ).” | |
| mechanism | material deprivation Material deprivation partly mediates the link between socioeconomic position and health behaviours. — “Material deprivation, which is related to a lower household income and poorer living conditions (i.e. poorer housing conditions, more financial problems), has been shown to partly mediate the association between socioeconomic position and health behaviours [ 13 ].” | mechanisms[0] |
| mechanism | class distinction High socioeconomic groups differentiate themselves by adopting healthy lifestyles, low groups by behaviours of freedom from convention. — “The influential French sociologist Pierre Bourdieu described how health- and lifestyle behaviours may therefore be subject to class distinction; high socioeconomic groups differentiate themselves from low socioeconomic groups by adopting healthy lifestyles, and low socioeconomic groups distinguish themselves from high socioeconomic groups with behaviours that give them a sense of ‘freedom from convention’ [ 16 ].” | mechanisms[1] |
| sense | cultural capital - for food choices, the resources an individual possesses and employs to make healthy food choices — “The practice or strategy of action ([ 30 ], p. 39) would be healthy food choice, and the currency or cultural capital would be the resources an individual possesses and employs to make healthy food choices.” | senses[0] |
| relation | socioeconomic position —operationalised-as→ highest attained educational level — was operationalised by — “In this 2011-survey, socioeconomic position was operationalised by the respondent’s highest attained educational level, which is considered a good indicator of socioeconomic position in the Netherlands [ 34 ]” | concept_edges[0] |
| weight | socioeconomic position —operationalised-as→ highest attained educational level — 0.9 — “In this 2011-survey, socioeconomic position was operationalised by the respondent’s highest attained educational level, which is considered a good indicator of socioeconomic position in the Netherlands [ 34 ]” | concept_edges[0] |
| relation | objectivised cultural capital —operationalised-as→ cultural possessions — was consistently measured in the literature by — “Objectivised cultural capital was consistently measured in the literature by cultural possessions and we translated this to a list of possessions related to food choice behaviour.” | concept_edges[1] |
| weight | objectivised cultural capital —operationalised-as→ cultural possessions — 0.85 — “Objectivised cultural capital was consistently measured in the literature by cultural possessions and we translated this to a list of possessions related to food choice behaviour.” | concept_edges[1] |
| relation | incorporated cultural capital —contrasted-with→ economic capital — It is not possible to convey ... as would be possible with — “It is not possible to convey incorporated cultural capital to someone else, as would be possible with economic capital or objectivised cultural capital.” | concept_edges[2] |
| weight | incorporated cultural capital —contrasted-with→ economic capital — 0.75 — “It is not possible to convey incorporated cultural capital to someone else, as would be possible with economic capital or objectivised cultural capital.” | concept_edges[2] |
| replication | original | |
| limitation | Restricting the search to 'cultural capital' may have missed studies with similar measures under other labels. — “Since we restricted the search term to ‘cultural capital’ we may have missed studies that included similar measures but did not label them as cultural capital.” | limitations[0] |
| limitation | The questionnaire is mostly food-specific. — “A limitation of the cultural capital questionnaire is that it is mostly food-specific; only the questions regarding family institutionalised cultural capital and the general values may also be applicable to other health-behaviours, like physical activity.” | limitations[1] |
| limitation | The items suit Western cultures like the Dutch and may need modification elsewhere. — “Another limitation may be that the selected questionnaire items are especially applicable to Western cultures similar to the Dutch culture; for application in other cultures, some questionnaire items may need some modification (e.g. regarding nutrition knowledge, and possession of cooking equipment).” | limitations[2] |
| limitation | Food intake was self-reported, with possible social desirability bias. — “A limitation of the empirical test is that food intake was based on self-reports, with social desirability bias as a potential problem.” | limitations[3] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | No prior study had empirically tested cultural capital in relation to healthy and unhealthy food choices. — “Our systematic review showed that the concept of cultural capital has not yet been empirically tested in relation to healthy and unhealthy food choices.” | content_tags.findings[0] |
| finding | Lower socioeconomic groups possessed significantly less cultural capital of all three types. — “As reported in Table 3 , lower socioeconomic groups possessed significantly less cultural capital of all three types than higher socioeconomic groups.” | content_tags.findings[1] |
| finding | Low total cultural capital was associated with unhealthy food choices, after adjustment for age, sex and socioeconomic position. — “Controlled for age, sex, and socioeconomic position, low levels of total cultural capital were associated with unhealthy food choices.” | content_tags.findings[2] |
| finding | The incorporated state showed the most consistent associations with unhealthy food choices. — “Of the three subtypes of cultural capital, the incorporated state showed the most consistent associations with unhealthy food choices, and in the expected direction: low levels of incorporated cultural capital were associated with more unhealthy food choices.” | content_tags.findings[3] |
| finding | Most reviewed studies (57) came from educational research. — “The vast majority of studies (n = 57) was conducted in the field of educational research, mainly related to educational achievement of children (e.g. with Grade Point Average, or math and reading performance as outcome measures).” | content_tags.findings[4] |
| quantity | 113 — — “Consequently, 113 articles were included in the review (see Table 1 ).” | content_tags.quantities[0] |
| quantity | 2,953 — — “These 30 respondents were excluded from the analysis, resulting in an analytic sample of N = 2,953.” | content_tags.quantities[1] |
| quantity | 56 — — “The questionnaire consisted of 56 items with an acceptable internal consistency (Cronbach’s alpha of .654).” | content_tags.quantities[2] |
| variable | family institutionalised cultural capital — “Therefore, to measure institutionalised cultural capital, we focused on the socialisation processes in which acquisition of cultural capital takes place, and operationalised family institutionalised cultural capital by educational level of the father, mother and partner of the respondent (each with four categories, i.e. 1 = no education or primary education; 2 = lower vocational and intermediate general education; 3 = intermediate vocational and higher general education; and 4 = higher professional education and university).” | content_tags.variables[0] |
| definition | social gradient — “For instance, a robust finding in health inequalities research is the social gradient, which illustrates that poorer health and health behaviours are not only found among the worst off; a surplus in income relates to a better health status at each level of the social hierarchy, also when comparing the highest with the second-highest income level [ 14 ].” | content_tags.definitions[0] |
| definition | openness to change — “The ten general values were aggregated to four higher-order values [ 52 ], i.e. openness to change (i.e. pursuing whatever intellectual or emotional directions one wishes, however unpredictable or uncertain the outcomes)” | content_tags.definitions[1] |
| key-terms | cultural capital, Bourdieu, food choices, socioeconomic position, incorporated cultural capital, objectivised cultural capital, economic capital, social capital, taste, nutrition knowledge |
| concept | cultural capital — a non-material resource that accumulates throughout the life course — “He states that taste is developed through cultural capital , a non-material resource that accumulates throughout the life course.” | concepts[0] |
| concept | objectivised cultural capital — Objectivised cultural capital, i.e. books, equipment, or a work of art, can be transferred to other people, just as money can be transferred. — “Objectivised cultural capital, i.e. books, equipment, or a work of art, can be transferred to other people, just as money can be transferred.” | concepts[1] |
| concept | incorporated cultural capital — not defined in this paper's own words; the paper operationalises it through participation, skills, knowledge and values — “It is not possible to convey incorporated cultural capital to someone else, as would be possible with economic capital or objectivised cultural capital.” | concepts[2] |
| concept | institutionalised state — official (school) diplomas, ‘guarantees’ properties of incorporated capital — “The third type of cultural capital is the institutionalised state, i.e. official (school) diplomas, ‘guarantees’ properties of incorporated capital” | concepts[3] |
| concept | economic capital — Economic capital includes all sources of income — “Economic capital includes all sources of income; social capital encompasses the “aggregate of the actual or potential social resources deriving from group membership” ([ 26 ], p. 51). Economic and social capital are linked to cultural capital by the access they provide to education and social networks” | concepts[4] |
| concept | social gradient — illustrates that poorer health and health behaviours are not only found among the worst off — “For instance, a robust finding in health inequalities research is the social gradient, which illustrates that poorer health and health behaviours are not only found among the worst off; a surplus in income relates to a better health status at each level of the social hierarchy, also when comparing the highest with the second-highest income level [ 14 ].” | concepts[5] |
| concept | material deprivation — related to a lower household income and poorer living conditions (i.e. poorer housing conditions, more financial problems) — “Material deprivation, which is related to a lower household income and poorer living conditions (i.e. poorer housing conditions, more financial problems), has been shown to partly mediate the association between socioeconomic position and health behaviours [ 13 ].” | concepts[6] |
| concept | taste — not defined in this paper's own words; in Bourdieu, the means by which high socioeconomic groups distinguish themselves, “such as musical, artistic and culinary tastes” — “Bourdieu depicted how high socioeconomic groups distinguish themselves from low socioeconomic groups through their ‘taste’ for various lifestyle attributes, such as musical, artistic and culinary tastes [ 17 ].” | concepts[7] |
| concept | socioeconomic position — operationalised by the respondent’s highest attained educational level — “In this 2011-survey, socioeconomic position was operationalised by the respondent’s highest attained educational level, which is considered a good indicator of socioeconomic position in the Netherlands [ 34 ]” | concepts[8] |
| concept | highest attained educational level — not defined in this paper's own words; measured with four categories — “In this 2011-survey, socioeconomic position was operationalised by the respondent’s highest attained educational level, which is considered a good indicator of socioeconomic position in the Netherlands [ 34 ]” | concepts[9] |
| concept | cultural possessions — not defined in this paper's own words; translated here to cooking-related possessions — “Objectivised cultural capital was consistently measured in the literature by cultural possessions and we translated this to a list of possessions related to food choice behaviour.” | concepts[10] |
| concept | openness to change — pursuing whatever intellectual or emotional directions one wishes, however unpredictable or uncertain the outcomes — “The ten general values were aggregated to four higher-order values [ 52 ], i.e. openness to change (i.e. pursuing whatever intellectual or emotional directions one wishes, however unpredictable or uncertain the outcomes)” | concepts[11] |
| concept | self-transcendence — transcending one's selfish concerns and promoting the welfare of others and of nature — “self-transcendence (i.e. transcending one's selfish concerns and promoting the welfare of others and of nature)” | concepts[12] |
| Cronbach's alpha α | as written: Cronbach’s alpha of .654 — “The questionnaire consisted of 56 items with an acceptable internal consistency (Cronbach’s alpha of .654).” | content_tags.statistics[0] |
| odds ratio OR | as written: OR 3.22 (95% CI: 1.77 — “Low socioeconomic groups were more likely to eat more unhealthy than healthy bread products (OR 3.22 (95% CI: 1.77–5.84)), and to eat more unhealthy than healthy meat products (OR 1.96 (95% CI: 1.37–2.80)) than higher socioeconomic groups (controlled for age and sex).” | content_tags.statistics[1] |
Full text — tagged
The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.
Bourdieuterm·key's Cultural Capitalterm·dom in Relation to Food Choicesterm·key: A Systematic Review of Cultural Capitalterm·dom Indicators and an Empirical Proof of Concept Abstract Unhealthy food choicesterm·key follow a socioeconomic gradient that may partly be explained by one’s ‘cultural capitalterm·dom’, as defined by Bourdieuterm·key. We aim 1) to carry out a systematic review to identify existing quantitative measures of cultural capitalterm·dom, 2) to develop a questionnaire to measure cultural capitalterm·dom for food choicesterm·key, and 3) to empirically test associations of socioeconomic positionterm·dom with cultural capitalterm·dom and food choicesterm·key, and of cultural capitalterm·dom with food choicesterm·key.
We systematically searched large databases for the key-word ‘cultural capitalterm·dom’ in title or abstract. Indicators of objectivised cultural capitalterm·dom and family institutionalised cultural capitalterm·var, as identified by the review, were translated to food choiceterm·key relevant indicators. For incorporated cultural capitalterm·dom, we used existing questionnaires that measured the concepts underlying the variety of indicators as identified by the review, i.e. participation, skills, knowledge, values. The questionnaire was empirically tested in a postal surveyterm·mth completed by 2,953 adults participating in the GLOBE cohort study, The Netherlands, in 2011.
The review yielded 113 studies that fulfilled our inclusion criteria. Several indicators of family institutionalised (e.g. parents’ education completed) and objectivised cultural capitalterm·dom (e.g. possession of books, art) were consistently used. Incorporated cultural capitalterm·dom was measured with a large variety of indicators (e.g. cultural participation, skills). Based on this, we developed a questionnaire to measure cultural capitalterm·dom in relation to food choicesterm·key. An empirical test of the questionnaire showed acceptable overall internal consistency (Cronbach’s alphaterm·stat of .654; 56 items), and positive associations between socioeconomic positionterm·dom and cultural capitalterm·dom, and between cultural capitalterm·dom and healthy food choicesterm·key.
Cultural capitalterm·dom may be a promising determinant for (socioeconomic inequalities in) food choicesterm·key.
Introduction Poor dietary intake is a major risk factor for morbidity and mortality [ 1 , 2 ]. Compared to high socioeconomic groups, low socioeconomic groups typically engage more in unhealthy behaviours, such as unhealthy food choicesterm·key [ 3 – 8 ]. Therefore, the behavioural explanation is an important mechanism underlying socioeconomic inequalities in health [ 9 – 11 ]. Our understanding of why low socioeconomic groups typically engage in unhealthy behaviours is still insufficient [ 12 ].
Unhealthy behaviours are thought not to be the result of entirely voluntary choices, but rather influenced by structures in the daily context, such as material resources [ 13 ]. Material deprivationterm·dom, which is related to a lower household income and poorer living conditions (i.e. poorer housing conditions, more financial problems), has been shown to partly mediate the association between socioeconomic positionterm·dom and health behaviours [ 13 ]. However, disparities in material resources are not the only explanation. For instance, a robust finding in health inequalities research is the social gradientterm·dom, which illustrates that poorer health and health behaviours are not only found among the worst off; a surplus in income relates to a better health status at each level of the social hierarchy, also when comparing the highest with the second-highest income level [ 14 ]. Similarly, healthy dietary intake also continues to improve above the level of poverty [ 3 ].
Health behaviours may also depend on one’s sociocultural resources. By means of socialisation processes, dispositions and behaviours of individuals become similar within each social class and generation[ 15 ]. The influential French sociologist Pierre Bourdieuterm·key described how health- and lifestyle behaviours may therefore be subject to class distinction; high socioeconomic groups differentiate themselves from low socioeconomic groups by adopting healthy lifestyles, and low socioeconomic groups distinguish themselves from high socioeconomic groups with behaviours that give them a sense of ‘freedom from convention’ [ 16 ]. Bourdieuterm·key depicted how high socioeconomic groups distinguish themselves from low socioeconomic groups through their ‘tasteterm·dom’ for various lifestyle attributes, such as musical, artistic and culinary tastesterm·dom [ 17 ]. He states that tasteterm·dom is developed through cultural capitalterm·dom , a non-material resource that accumulates throughout the life course. Although tasteterm·dom is also linked to economic capitalterm·dom, it mainly relates to cultural capitalterm·dom, as tasteterm·dom remains stable, also when people’s income increases over time, thereby reflecting certain cultural norms and values [ 17 , 18 ].
Possession of cultural capitalterm·dom may thus be a missing link in the relation between socioeconomic positionterm·dom and food choiceterm·key behaviour. Although the relation between cultural capitalterm·dom and health, healthy lifestyles, and health behaviours is a subject of growing interest [ 19 – 24 ], the concept has found little empirical application in studies regarding socioeconomic inequalities in health behaviours so far (e.g. [ 23 , 25 ]). Likely, research into the role of cultural capitalterm·dom is hampered by the unfamiliarity of researchers with the concept, and by the lack of clear indicators and measures of cultural capitalterm·dom that can be applied to study (inequalities in) health behaviours. Abel (2008) noted that new indicators need to be developed for studying cultural capitalterm·dom in relation to health and health behaviour [ 20 ].
The aim of the present paper was to develop a set of questionnaire items in order to investigate to what extent the possession of cultural capitalterm·dom differs between socioeconomic groups, and to what extent cultural capitalterm·dom relates to healthy and unhealthy food choicesterm·key. We first wanted to learn about currently used measures of cultural capitalterm·dom. Therefore, (1) we systematically reviewed the literature in order to identify currently used quantitative measures of cultural capitalterm·dom. Subsequently, based on these findings (2) we developed a set of questionnaire items to study cultural capitalterm·dom in relation to food choicesterm·key. Finally, (3) we piloted our questionnaire to assess its applicability in empirical research, and investigated a) associations of socioeconomic positionterm·dom (i.e. respondent’s highest attained educational levelterm·mth) with unhealthy food choicesterm·key (adjusted for age and sex), b) associations of socioeconomic positionterm·dom with cultural capitalterm·dom, and c) associations of cultural capitalterm·dom with unhealthy food choicesterm·key (adjusted for socioeconomic positionterm·dom, age and sex).
First, we will give an introduction into the concept of cultural capitalterm·dom, the three forms of cultural capitalterm·dom, and its relation to economic and social capitalterm·key.
Cultural capitalterm·dom theory The cultural capitalterm·dom notion provides an explanation of social stratification mechanisms and was originally conceptualised by Bourdieuterm·key to explain class differences in academic achievement [ 26 – 28 ]. Cultural capitalterm·dom is acquired through education and socialisation and includes “the distinctive forms of knowledge and ability that students acquire […] from their training in the cultural disciplines” (Bourdieuterm·key, 2010, p. xviii) [ 17 ]. Through available cultural capitalterm·dom in the family, one is more inclined to ‘inherit’ cultural resources that can be mobilised to accumulate incorporated cultural capitalterm·dom [ 26 , 29 ]. Cultural capitalterm·dom is not available to everyone, and like other forms of capital (e.g. monetary assets), it can serve as a currency to obtain other resources. Scarcity and exclusiveness determine its value. Individuals are usually not aware of the cultural capitalterm·dom they possess and put into action. Rather it takes the form of unconscious customs and beliefs and the feel for what is the right thing to do [ 17 , 24 ].
Cultural capitalterm·dom emerges in three different states: incorporated (“embodied”; e.g. dispositions and competencies), objectivised (“objectified”; e.g. possession of books, dictionaries, instruments), and institutionalised (e.g. educational qualifications) [ 20 , 26 ]. Incorporated cultural capitalterm·dom, “the form of long-lasting dispositions of the mind and the body”, entails socialisation, personal effort, and time investment, and becomes a part of the individual’s habitus (Bourdieuterm·key, 1986, p. 47) [ 26 ]. It is not possible to convey incorporated cultural capitalterm·dom to someone else, as would be possible with economic capitalterm·dom or objectivised cultural capitalterm·dom. Lareau and Weininger ([ 28 ], p. 156) refer to incorporated cultural capitalterm·dom as “the legitimate cultural attitudes, preferences and behaviours […] that are internalized during the socialization process”. For socioeconomic inequalities in food choiceterm·key behaviour, we expect incorporated cultural capitalterm·dom to have the largest potential to be on the causal chain between socioeconomic positionterm·dom and food choicesterm·key. In our study, we hypothesize to find the strongest associations of cultural capitalterm·dom with food choicesterm·key for the incorporated state. Also Abel emphasised the importance of incorporated cultural capitalterm·dom for health behaviour [ 19 , 20 ], consistent with Bourdieuterm·key [ 17 , 26 ].
Objectivised cultural capitalterm·dom, i.e. books, equipment, or a work of art, can be transferred to other people, just as money can be transferred. Nevertheless, proper use of objectivised cultural capitalterm·dom is closely tied to incorporated cultural capitalterm·dom: a work of art becomes meaningful only to those who know how to evaluate its value. The third type of cultural capitalterm·dom is the institutionalised stateterm·dom, i.e. official (school) diplomas, ‘guarantees’ properties of incorporated capital, and was described by Bourdieuterm·key as “legally guaranteed qualifications, formally independent of the person of their bearer” ([ 26 ], p. 50).
Aside from cultural capitalterm·dom, Bourdieuterm·key defined two other forms of capital to describe the composition of lifestyle attributes related to social class: economic capitalterm·dom and social capitalterm·key [ 18 ]. Economic capitalterm·dom includes all sources of income; social capitalterm·key encompasses the “aggregate of the actual or potential social resources deriving from group membership” ([ 26 ], p. 51). Economic and social capitalterm·key are linked to cultural capitalterm·dom by the access they provide to education and social networks (e.g. tuition and club-membership fees). The other way around, cultural capitalterm·dom determines accumulation and deployment of economic and social capitalterm·key; for instance, education may give access to better-paid jobs, and shared norms and values are necessary to enter certain social networks [ 20 , 26 ].
A fundamental question underlying our search for indicators of cultural capitalterm·dom is: when does culture become capital, i.e. when does culture take the form of a currency that can be deployed to get ahead in life, when is it valued as giving an individual a social advantage? For cultural capitalterm·dom in relation to socioeconomic inequalities in food choiceterm·key behaviour, the advantage would be good health. The practice or strategy of action ([ 30 ], p. 39) would be healthy food choiceterm·key, and the currency or cultural capitalterm·dom would be the resources an individual possesses and employs to make healthy food choicesterm·key. Thus, how exactly can we operationalise this cultural capitalterm·dom to make healthy food choicesterm·key?
Methods Systematic review—methods We conducted a review of the literature in a broad range of databases, in order to get a complete overview of existing quantitative measures of cultural capitalterm·dom. We searched the databases PubMed, PsychInfo, Web of Science, and JSTOR with the key-word ‘cultural capitalterm·dom’ as a search term in ‘title’ and/or ‘abstract’ (e.g. PubMed search strategy: (cultural*[tiab] AND capital*[tiab])). Inclusion criteria for studies were: cultural capitalterm·dom was quantitatively measured in an empirical study, studies were published as research articles in peer-reviewed journals, in the English language, and published between 1979 (first publication year of Bourdieuterm·key’s book ‘Distinction’ [ 17 ]) and October 2010. TJ scanned the retrieved articles for the term ‘cultural capitalterm·dom’ in either title and/or abstract, and assessed abstracts for quantitative measurement of cultural capitalterm·dom; the other criteria were used as search limits in the databases. If the measurement of cultural capitalterm·dom could not be derived from the abstract, we labelled the article as ‘potentially relevant’ and included the article for viewing the full text. When, after reading full texts, it turned out a study did not include at least one quantitative measure of cultural capitalterm·dom, we excluded the article.
Articles that met all inclusion criteria were classified according to their research field. Further, we extracted indicators of cultural capitalterm·dom from the included articles and classified them according to the three states of cultural capitalterm·dom, i.e. objectivised, incorporated, and institutionalised. We distinguished between indicators measured at the individual level (i.e. cultural capitalterm·dom measured as activities and/or knowledge of the respondent) and family level (cultural capitalterm·dom measured as activities and/or knowledge of the respondent’s parents), or a combination of the two. Some studies drew from the same data source; nevertheless we evaluated all studies separately, since the applied indicators differed.
Questionnaire development—methods The review of existing measures of institutionalised, objectivised and incorporated cultural capitalterm·dom was used as main input for the development of the questionnaire to measure food choiceterm·key relevant cultural capitalterm·dom. The most often applied indicator(s) were used for translation into measures of food choiceterm·key relevant cultural capitalterm·dom, using existing questionnaires where possible. If a predominant measure did not come forth from the review, we used the main concepts underlying the different indicators, and searched the literature to find existing questionnaires to measure these concepts. The questionnaire was pilot-tested for readability, comprehensibility and completion time among a convenience sample of 41 adults (30% male, mean age 46 years, 49% high educated), partly consisting of friends of the second author, and partly of patients to a physiotherapy practice. Pilot participants filled in the questionnaire (including four additional questions on completion time and questions that were unclear, difficult, or unpleasant to answer) at home, and returned it to the second author directly, or on their next visit to the physiotherapy practice. Based on the results of the pilot studies, several questions were rephrased or excluded.
Empirical study—methods The developed questionnaire was included in a large-scale postal surveyterm·mth in 2011, administered in a new wave of data collection for the longitudinal cohort study called GLOBE (more information in [ 31 – 33 ]). Of the respondents to the previous surveyterm·mth in 2004 (N = 4,784) which formed a stratified sample of the 25–75 years old population in the city of Eindhoven and surrounding cities in 2004, n = 249 had died, n = 76 had emigrated, and n = 14 were lost to follow-up (i.e. no correct address information available), which resulted in a sample of N = 4,437 that was sent the 2011-survey. For the total of 2,983 respondents that returned the surveyterm·mth (response 67.2%), missing values for sex (n = 21), age (n = 24), and educational level (n = 172) could largely be replaced by information from the 2004-questionnaire, resulting in only one case with a missing value for age and 29 cases with missing values for educational level. These 30 respondents were excluded from the analysis, resulting in an analytic sample of N = 2,953. The use of personal data in the GLOBE study is in compliance with the Dutch Personal Data Protection Act and the Municipal Database Act, and has been registered with the Dutch Data Protection Authority (number 1248943). No formal approval of the medical ethics committee of the Erasmus University Medical Centre was required for the study.
In this 2011-survey, socioeconomic positionterm·dom was operationalised by the respondent’s highest attained educational levelterm·mth, which is considered a good indicator of socioeconomic positionterm·dom in the Netherlands [ 34 ] (measured with four categories, i.e. 1 = no education or primary education; 2 = lower vocational and intermediate general education; 3 = intermediate vocational and higher general education; and 4 = higher professional education and university).
With a Food Frequency Questionnaire that was part of the 2011-survey [ 35 – 37 ], we obtained self-reported information on the number of days per week that respondents consumed specific food items (categories: never, less than 1 day/week; 1–2 days/week; 3–4 days/week; 5–6 days/week; and every day), based on which total frequency outcomes were constructed for: consumption of unhealthy vs. healthy bread products (= summed frequency of eating white bread and croissants, minus the summed frequency of eating wholemeal bread and rye bread), unhealthy vs. healthy snacking (= summed frequency of eating candy bars, potato chips, and salted nuts, minus the summed frequency of eating gingerbread, Japanese mix, popcorn, and unsalted nuts), and consumption of unhealthy vs. healthy meat products (= summed frequency of eating beef, mutton, and pork, minus the summed frequency of eating fish, chicken, and vegetarian products like tofu). For analytic purposes, outcomes were dichotomised in: 1 = making more unhealthy choices, vs. 0 = making more healthy choices, or as many healthy as unhealthy choices.
Statistical analyses were conducted in SPSS 20.0. Internal consistency and item correlation of cultural capitalterm·dom measures were tested by Cronbach’s alphaterm·stat. Preferably, principal component analysis was used to identify factors that represented the underlying concepts (results are available in S2 Tables Factor analyses). In case of a low Cronbach’s alphaterm·stat (<0.5), or a high rate of missing values on one of the items, we calculated mean scores or sum scores to combine several items in one measure.
In logistic regression models, odds ratiosterm·stat with 95% confidence intervals were calculated for the food choiceterm·key outcomes by socioeconomic positionterm·dom (indicated by the respondent’s highest attained educational levelterm·mth) controlled for age and sex, and by the cultural capitalterm·dom indicators, controlled for sex, age, and socioeconomic positionterm·dom. The analyses were weighted to reflect the source population (i.e. the 25–75 years of age adult population of the city of Eindhoven and surrounding cities in 2004) with respect to sex, age, and educational level.
Results Systematic review—results The literature search yielded 2,659 articles (PubMed: 48; PsychINFO 295; Web of Science: 563; JSTOR: 1,753) (see Fig 1 for the PRISMA flow chart regarding the inclusion of articles). We excluded articles when the term ‘cultural capitalterm·dom’ was not mentioned in the title or abstract, or when other inclusion criteria were not met. This strategy reduced the number of potentially relevant articles to 311, and to 230 after duplicates were removed. Close reading of full-texts resulted in exclusion of another 117 articles since these were either theoretical or methodological papers, reviews, or qualitative studies. Consequently, 113 articles were included in the review (see Table 1 ). The vast majority of studies (n = 57) was conducted in the field of educational research, mainly related to educational achievement of children (e.g. with Grade Point Average, or math and reading performance as outcome measures). Other research fields were employment/career, volunteering, parenting behaviour, and religion. Seven articles reported studies conducted in the field of (mental) health, though three ensued from the same study [ 25 , 38 , 39 ].
PRISMA flow diagram, regarding the inclusion of articles in the systematic review.
Results of the systematic review: most prevalent indicators of cultural capitalterm·dom.
A large variety of cultural capitalterm·dom measures ensued from the systematic review (see S1 Tables Systematic review). Institutionalised cultural capitalterm·dom was assessed in a limited number of studies though rather consistently, namely as highest education completed (both on the individual level and family level). Objectivised cultural capitalterm·dom was consistently operationalized with indicators of ‘cultural possessionterm·mth’, for which five measures at the individual level and four at the family level were found, e.g. possession of books, owning works of art, and library membership. For incorporated cultural capitalterm·dom , both on the level of the individual and the family, a large variety of measures was found, with cultural participation, cultural skills, cultural knowledge, and values as main underlying themes. Attending cultural events or activities (e.g. going to concerts, visiting museums) was most often employed, with n = 34 for the individual level, and n = 19 for the family level (i.e. parents attending cultural activities and parents taking child to cultural activities together). Other often-applied operationalisations of incorporated cultural capitalterm·dom were cultural skills acquired through participation in classes (e.g. painting, playing a musical instrument), reading skills, cultural knowledge (e.g. of literature, music, arts), and important life values, like religion. The few studies that investigated associations of incorporated cultural capitalterm·dom with (mental) health outcomes also applied these types of indicators.
Questionnaire development—results We developed a set of questionnaire items that capture cultural capitalterm·dom of the three different types in relation to food choicesterm·key (see Table 2 ). The questionnaire consisted of 56 items with an acceptable internal consistency (Cronbach’s alphaterm·stat of .654).
Questionnaire items to measure cultural capitalterm·dom related to food choicesterm·key.
The review showed that institutionalised cultural capitalterm·dom was consistently measured by educational level, either own educational level as indicator of individual institutionalised cultural capitalterm·dom, or parents’ highest attained educational levelterm·mth as indicators of family institutionalised cultural capitalterm·var. The current study is part of a broader research project into socioeconomic inequalities in food choicesterm·key, in which we use the respondent’s own educational level as indicator of socioeconomic positionterm·dom, which therefore cannot be used as indicator of individual institutionalised cultural capitalterm·dom. Therefore, to measure institutionalised cultural capitalterm·dom, we focused on the socialisation processes in which acquisition of cultural capitalterm·dom takes place, and operationalised family institutionalised cultural capitalterm·var by educational level of the father, mother and partner of the respondent (each with four categories, i.e. 1 = no education or primary education; 2 = lower vocational and intermediate general education; 3 = intermediate vocational and higher general education; and 4 = higher professional education and university). The educational level of the respondent’s father and mother may capture family cultural capitalterm·dom influences during childhood and adolescence to which the respondent has been exposed, and which may have influenced the respondent’s current capital levels, and current food choicesterm·key. The educational level of the respondent’s partner may capture current family cultural capitalterm·dom influences on capital levels and food choicesterm·key. The Cronbach’s alphaterm·stat for these three items was .762. We calculated the mean score of these three items as overall measure of family institutionalised cultural capitalterm·var. For n = 713 respondents, this mean value was based on less than three items, due to missing values. Respondents with missing values on all three items (n = 394) received a missing value for family institutionalised cultural capitalterm·var. The mean score was dichotomised in 0 = low (1–2.51) vs. 1 = high (2.52–4) family institutionalised cultural capitalterm·var.
Objectivised cultural capitalterm·dom was consistently measured in the literature by cultural possessionsterm·mth and we translated this to a list of possessions related to food choiceterm·key behaviour. We asked respondents whether they owned several cooking-related possessions, i.e. stove, cook book(s), set of knives, kitchen scale, juicer, bread maker, and frying pan (yes/no) (Cronbach’s alphaterm·stat: .391). The items regarding the possession of a bread maker and frying pan showed a weak correlation with the other five items. After exclusion of these two items, the Cronbach’s alphaterm·stat raised to .545. A sum score was created based on the remaining five items, and divided in two categories: 0 = low (0–4) vs. 1 = high (5) objectivised cultural capitalterm·dom.
The measures of incorporated cultural capitalterm·dom that came forth from the literature review were wide ranging, however, we selected the main underlying themes of participation , skills , knowledge , and values as the constructs that form one’s incorporated cultural capitalterm·dom, and searched the literature to find existing questionnaires to measure these.
Participation was measured with two food-related participation items, namely (categorisation between brackets): “In the last month, how many times have you met with people in a public place to have some food?” (not at all, vs. at least once) [ 40 ], and “In the last month, how many times have people visited you in your home to have dinner, or have you visited people for dinner in their home?” (not at all, vs. at least once) [ 40 ]. A sum score of the two variables was created (ranging from 0–2), and was divided in 1 = no (0) and 2 = some (1–2) food-related participation.
For food choiceterm·key related skills we distinguished three types of abilities: cooking skills [ 41 , 42 ], grocery shopping skills [ 43 ], and skills to find and process information about nutrients and food preparation (adapted from [ 44 ]). Cooking skills were measured with statements about the preparation of vegetables and fish, and the use of recipes (with answers on a 5-point Likert scale ranging from totally agree to totally not agree). Cronbach’s alphaterm·stat for these three items was .773, and a factor analyses resulted in one factor that explained 70% of the variance. Scores for this factor were interpreted and categorised as: 1 = low cooking skills (negative scores) vs. 2 = high cooking skills (zero and positive scores). Grocery shopping skills were measured with two items on making a shopping list or deciding in store what to buy (with 5 answering options ranging from always to never). Cronbach’s alphaterm·stat for these two items was .561, and a factor analyses resulted in one factor that explained 69% of the variance. Scores for this factor were interpreted and categorised as: 1 = low grocery shopping skills (negative scores) vs. 2 = high (zero and positive scores). Food information skills were measured with five items, for instance: “Do you read nutrition information on the labels of food products?” and “Do you use information on the labels of food products to decide what to buy?” (with 5 answering options ranging from always to never). Cronbach’s alphaterm·stat for these five items was .529, and particularly the item for understanding nutrition information loaded low on all factors. After excluding this item, Cronbach’s alphaterm·stat was .537, and a factor analysis on the remaining four items resulted in two factors that together explained 69% of the variance. Scores for these factors were interpreted as follows: 1 = low skills for use of nutrition information on food packages (negative scores) vs. 2 = high (zero and positive scores) skills, and 1 = low skills for use of nutrition information and recipes from magazines and the internet (negative scores) vs. 2 = high (zero and positive scores).
Nutrition knowledgeterm·key was measured with an existing questionnaire including 16 items (Cronbach’s alphaterm·stat: .519), namely four different questions (e.g. do these products contain high or low levels of added sugar? do these products contain high or low levels of protein?), that were asked with regard to four products each (e.g. bananas, chicken, chocolate, red meat) (three answer categories: high, low, don’t know) [ 45 , 46 ]. A sum score of all correct answers was made (ranging from 0–16), and divided in 1 = low (0–10) and 2 = high (11–16) nutrition knowledgeterm·key.
We created a mean score based on the participation, skills and knowledge variables as measure of total incorporated cultural capitalterm·dom , and divided the mean score into 0 = low (1.00–1.49), vs. 1 = high (1.50–2.00).
Instead of food values-which would be too close to the tasteterm·dom of food, and therefore to the outcome of interest- ten distinctive general human values were constructed [ 47 , 48 ] by means of the Portrait Values Questionnaire of Schwartz (also applied in previous studies on food choicesterm·key [ 49 , 50 ]), including 21 items (Cronbach’s alphaterm·stat: 0.801). Values were computed excluding respondents who had more than 5 missing values or 16 similar answers on the value items, according to Bilsky, Janik, and Schwartz (2011) [ 51 ]. The ten general values were aggregated to four higher-order values [ 52 ], i.e. openness to changeterm·dom (i.e. pursuing whatever intellectual or emotional directions one wishes, however unpredictable or uncertain the outcomes), conservation (i.e. preserving the status quo and the certainty it provides in relationships with close others, institutions, and traditions), self-transcendenceterm·dom (i.e. transcending one's selfish concerns and promoting the welfare of others and of nature), and self-enhancement (i.e. enhancing one's own personal interests, even at the expense of others) and dichotomised into high = 1 and low = 0. Although considered part of one’s incorporated cultural capitalterm·dom, these general values were too distinct in nature from the other variables to capture in the overall sum score for incorporated cultural capitalterm·dom, and therefore, were analysed separately.
A variable for total cultural capitalterm·dom (excluding the general values) was created by computing the mean score of the family institutionalised, objectivised, and total incorporated cultural capitalterm·dom variables, which was dichotomised into low total cultural capitalterm·dom (mean score of 0–0.50) vs. high (mean score of 0.51–1.00). A factor analysis of all cultural capitalterm·dom items combined (leaving out the general values, and with a sum score for food knowledge) showed that the measures of family institutionalised cultural capitalterm·var loaded on one factor with the items for nutrition information skills, and defined separate scales from those of objectivised cultural capitalterm·dom, cooking skills, shopping skills, and participation (results available in S2 Tables Factor analyses).
Empirical study–results Mean age of the sample was 56.4 years (SD 13.0) and 56.7% was female. As reported in Table 3 , lower socioeconomic groups possessed significantly less cultural capitalterm·dom of all three types than higher socioeconomic groups. For instance, where 63.6% of the high socioeconomic group possessed a high level of incorporated cultural capitalterm·dom, this was 32.8% among the low socioeconomic groups. The proportion of missing values on the variables for family institutionalised and objectivised cultural capitalterm·dom was higher among low than high socioeconomic groups.
Odds ratiosterm·stat for unhealthy food choicesterm·key by socioeconomic positionterm·dom a (adjusted for age and sex) and by cultural capitalterm·dom (adjusted for age, sex, and socioeconomic positionterm·dom), and prevalence rates of family institutionalised, objectivised and incorporated cultural capitalterm·dom by socioeconomic positionterm·dom a .
Low socioeconomic groups were more likely to eat more unhealthy than healthy bread products (OR 3.22 (95% CI: 1.77–5.84)), and to eat more unhealthy than healthy meat products (OR 1.96 (95% CI: 1.37–2.80)) than higher socioeconomic groups (controlled for age and sex). The second-lowest group was more likely to consume more unhealthy than healthy snacks (OR 1.52 (95% CI: 1.11–2.07) compared to the highest socioeconomic group.
Controlled for age, sex, and socioeconomic positionterm·dom, low levels of total cultural capitalterm·dom were associated with unhealthy food choicesterm·key. Of the three subtypes of cultural capitalterm·dom, the incorporated state showed the most consistent associations with unhealthy food choicesterm·key, and in the expected direction: low levels of incorporated cultural capitalterm·dom were associated with more unhealthy food choicesterm·key. Bread consumption was associated with all cultural capitalterm·dom variables, whereas meat consumption and snacking were only significantly associated with total and incorporated cultural capitalterm·dom. Regarding the general values, low educated were less open to change and more conservative than high educated, and had lower scores for self-transcendenceterm·dom and self-enhancement. A low openness to changeterm·dom and low self-transcendenceterm·dom seemed to be associated with making less healthy food choicesterm·key, especially for bread consumption.
Discussion Our systematic review showed that the concept of cultural capitalterm·dom has not yet been empirically tested in relation to healthy and unhealthy food choicesterm·key. Several indicators of institutionalised (e.g. educational level) and objectivised cultural capitalterm·dom (e.g. possession of books, art) have consistently been applied in other research fields, whereas incorporated cultural capitalterm·dom has been measured with a large variety of indicators (e.g. cultural participation, skills). During questionnaire development, indicators of institutionalised and objectivised cultural capitalterm·dom were translated to food choiceterm·key relevant indicators. For incorporated cultural capitalterm·dom, we used existing questionnaires that measured the larger concepts that covered the variety of indicators as identified by the review (i.e. participation, skills, knowledge, values). An empirical test of the cultural capitalterm·dom questionnaire showed an acceptable overall internal consistency (Cronbach’s alphaterm·stat of .654; 56 items). Higher socioeconomic groups significantly retained more cultural capitalterm·dom of all three states than those from lower socioeconomic groups, and having more cultural capitalterm·dom was associated with healthier food choicesterm·key.
The main strength of the present study is our comprehensive approach for the development of a set of measures for food choiceterm·key relevant cultural capitalterm·dom. The literature review resulted in an inclusive overview of all cultural capitalterm·dom indicators and measures that have been applied in different fields of research, from the early conceptualisations to recent understandings of the concept of cultural capitalterm·dom. We were thereby more inclusive than a previous review of selected studies that applied cultural capitalterm·dom in educational research [ 28 ]. Since we restricted the search term to ‘cultural capitalterm·dom’ we may have missed studies that included similar measures but did not label them as cultural capitalterm·dom. A limitation of the cultural capitalterm·dom questionnaire is that it is mostly food-specific; only the questions regarding family institutionalised cultural capitalterm·var and the general values may also be applicable to other health-behaviours, like physical activity. However, the advantage of this approach is that the food-choice specific cultural capitalterm·dom variables are more likely to have causal associations with food choicesterm·key than general cultural capitalterm·dom indicators (like museum visits, and possession of art), which make them more useful as entry points for intervention development to promote healthy food choicesterm·key. Another limitation may be that the selected questionnaire items are especially applicable to Western cultures similar to the Dutch culture; for application in other cultures, some questionnaire items may need some modification (e.g. regarding nutrition knowledgeterm·key, and possession of cooking equipment). A limitation of the empirical test is that food intake was based on self-reports, with social desirability bias as a potential problem.
We combined the results of the literature review with existing questionnaires on the underlying concepts of the three states of cultural capitalterm·dom. Especially with regard to incorporated cultural capitalterm·dom, clearly a gap exists between Bourdieuterm·key’s abstract ideas and the operationalisation into measurable constructs of incorporated cultural capitalterm·dom, as the measures of incorporated cultural capitalterm·dom that came forth from the literature review were wide ranging. Therefore, we selected the main underlying concepts of participation, skills knowledge, and values-which are also prominent in the original writings of Bourdieuterm·key [ 17 , 26 ] and others [ 20 , 28 ]. Measures for skills and knowledge that were selected from the food choiceterm·key literature sometimes needed some adaption to the Dutch situation, as, for instance, the questions on food knowledge included particular British food items not well known in the Netherlands.
We followed Abel [ 20 ] by appreciating values as reflecting incorporated cultural capitalterm·dom, since values are typically ‘long lasting dispositions of the mind’ [ 26 ], and guide what individuals appraise as important in their lives [ 53 – 55 ]. Values are both individual characteristics and cultural patterns of social class [ 56 ] and have a major influence on lifestyle and consumption patterns [ 57 ]. We measured general values instead of food-choice related values since the food values would be too close to the tasteterm·dom of food, and therefore to the outcome of interest itself. However, the general values were rather distinctive in nature from the other food-choice specific incorporated cultural capitalterm·dom items so that these could not be summed in the overall food-choice related incorporated cultural capitalterm·dom sum score.
In additional univariate analyses, we separately investigated the association of each specific variable (that was used to compose the cultural capitalterm·dom variables) with food choicesterm·key, adjusted for sex, age and socioeconomic positionterm·dom (see S3 Tables Cultural capitalterm·dom and food choicesterm·key). Results indicated that the educational level of the respondent’s father, mother and partner had strong positive associations with bread choices, while for the other two outcomes associations were weaker. The possession of a juicer (one of the items used as indicator of objectivised cultural capitalterm·dom) had consistent significant associations with all three food choiceterm·key outcomes. Skills for preparing vegetables and fish, reading nutrition information on food packages, and food knowledge (which were all used as indicators of incorporated cultural capitalterm·dom) also showed consistent positive associations with healthy food choicesterm·key. The two food participation indicators were positively related with healthier meat choices, but less strongly with the other two food choiceterm·key outcomes.
The review showed that own educational level was the most often used indicator of institutionalised cultural capitalterm·dom at the individual level. However, as own educational level was used as indicator of socioeconomic positionterm·dom in our study, and as we were looking for factors that can explain socioeconomic inequalities in food choicesterm·key, we focused on the socialisation processes in which acquisition of cultural capitalterm·dom takes place, and operationalised family institutionalised cultural capitalterm·var by educational level of the father, mother and partner of the respondent. What the best way is to operationalise institutionalised cultural capitalterm·dom and socioeconomic positionterm·dom when both are needed from a conceptual point of view deserves attention in future studies.
For most cultural capitalterm·dom items, the proportion of missing values was acceptable, ranging between 1–5%. Only for the family institutionalised cultural capitalterm·var items, a large number of respondents missed data on educational level of their partner (19.1%), father (26.7%) and/or mother (26.1%), with the largest proportion of missing values among respondents with a low socioeconomic positionterm·dom. In further analyses, we will apply multiple imputation on the separate cultural capitalterm·dom items, before combining them into mean scores or sum scores.
We were conscientious in our attempt to develop new indicators of cultural capitalterm·dom that did justice to Bourdieuterm·key’s ideas. Empirical assessment of the cultural capitalterm·dom questionnaire suggested promising associations with socioeconomic positionterm·dom and several food choicesterm·key, and will soon be explored further for other food choiceterm·key outcomes. Also, we will further explore associations of food-related cultural capitalterm·dom with Bourdieuterm·key’s general indicators of cultural capitalterm·dom (e.g. possession of books and works of art), and associations of cultural capitalterm·dom with social and economic capitalterm·dom in relation to inequalities in food choicesterm·key. Also, future research may study cultural capitalterm·dom in relation to other health behaviours, such as physical activity. Questionnaire items regarding objectivised and incorporated cultural capitalterm·dom should then be adjusted accordingly.
Conclusions This study’s systematic review identified a large variety of cultural capitalterm·dom indicators which served as main input for the development of a cultural capitalterm·dom questionnaire. An empirical application of the questionnaire demonstrated that higher educated people significantly possessed more cultural capitalterm·dom, and those with a high level of cultural capitalterm·dom more often made healthier food choicesterm·key as opposed to those with low cultural capitalterm·dom. Cultural capitalterm·dom may be a new, potentially powerful explanation for inequalities in food choicesterm·key, which may lead to new entry points for the development of interventions to promote healthy food choicesterm·key among low socioeconomic groups.
Supporting Information (DOC) Click here for additional data file.
(XLS) Click here for additional data file.
(DOC) Click here for additional data file.
(DOC) Click here for additional data file.
(DOC) Click here for additional data file.