Policy Memo for Masters in Public Health

profilereneo.kena87
1.Softdrinksforlunch.pdf

British Food Journal Soft drinks for lunch? Self-control, intentions and social influences Elisabeth Lind Melbye, Merete Hagen Helland,

Article information: To cite this document: Elisabeth Lind Melbye, Merete Hagen Helland, (2018) "Soft drinks for lunch? Self-control, intentions and social influences", British Food Journal, Vol. 120 Issue: 8, pp.1735-1748, https://doi.org/10.1108/ BFJ-11-2017-0605 Permanent link to this document: https://doi.org/10.1108/BFJ-11-2017-0605

Downloaded on: 01 November 2018, At: 06:28 (PT) References: this document contains references to 69 other documents. To copy this document: [email protected] The fulltext of this document has been downloaded 96 times since 2018*

Users who downloaded this article also downloaded: (2018),"Exploring the visual appeal of food guide graphics: A compositional analysis of dinner plate models", British Food Journal, Vol. 120 Iss 8 pp. 1682-1695 <a href="https://doi.org/10.1108/ BFJ-02-2018-0112">https://doi.org/10.1108/BFJ-02-2018-0112</a> (2018),"Alternative food networks: sustainable business models for anti-consumption food cultures", British Food Journal, Vol. 120 Iss 8 pp. 1776-1791 <a href="https://doi.org/10.1108/ BFJ-12-2017-0731">https://doi.org/10.1108/BFJ-12-2017-0731</a>

Access to this document was granted through an Emerald subscription provided by emerald- srm:464943 []

For Authors If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service information about how to choose which publication to write for and submission guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.

About Emerald www.emeraldinsight.com Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online products and additional customer resources and services.

Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

*Related content and download information correct at time of download.

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

Soft drinks for lunch? Self-control, intentions and

social influences Elisabeth Lind Melbye

Norwegian School of Hotel Management, Faculty of Social Sciences, University of Stavanger, Stavanger, Norway, and

Merete Hagen Helland Department of Education and Sports Science, Faculty of Arts and Education,

University of Stavanger, Stavanger, Norway

Abstract Purpose – The purpose of this paper is to explore associations between food-related self-control, intentions, descriptive peer norms, parents’ healthy eating guidance and adolescents’ consumption of sugar-sweetened beverages (SSB) in a school lunch setting. An additional aim was to evaluate the psychometric properties of the measure used to assess food-related self-control in order to reveal potential multi-dimensionality. Design/methodology/approach – A web-based survey was conducted among 694 Norwegian high school students. Multiple logistic regression was used to explore associations between the independent variables and SSB consumption. Psychometric evaluation of the self-control measure included factor analysis and internal consistency reliability. Findings – Factor analysis resulted in two food-related self-control dimensions: resistance and avoidance. Multiple logistic regression showed that intentions was the strongest predictor of SSB consumption in the sample. Avoidance and descriptive peer norms appeared as weaker predictors. Research limitations/implications – Based on the findings, the authors suggest that future studies may consider developing guiding principles on how to create health-promoting eating intentions in adolescents, how to deal with peer norms related to foods and beverages and how to avoid tempting stimuli in the environment. Such strategies may be helpful when structural changes in the environment are not feasible in the near future. Originality/value – An original aspect of the present study is that it includes a psychometric analysis of a supposedly one-dimensional self-control measure. Further, it adds to the knowledge about variables associated with adolescent SSB consumption in a school lunch context. Keywords Intentions, Adolescents, Social influence, Soft drinks, Self-control Paper type Research paper

Introduction A low intake of foods and beverages with added sugar is recommended by national and international health authorities (Nordic Council of Ministers, 2014; WHO, 2003). In Norway, which is the setting of the present study, a recent decline in the consumption of sugar-sweetened beverages (SSB) has been reported. Still, Norwegian adolescents have higher intakes than recommended (Fismen et al., 2016). According to a national survey among high schools (n ¼ 447) conducted by the Norwegian Directorate of Health (2013), the share of high school students drinking SSB with lunch on a daily basis is estimated to be more than 20 per cent. The majority (92 per cent) of the surveyed schools reported that they have a school cafeteria or other facilities were students can buy various foods and beverages, and that approximately 60 per cent of these facilities offer sugar-sweetened soft drinks, while 80 per cent offer milkshake/chocolate milk and iced coffee (of which most are sweetened with sugar) on a daily basis. Furthermore, students at most schools (96 per cent) have easy access to local shops and restaurants offering a wide variety of unhealthy foods and beverages. The survey also revealed that schools find it challenging to offer foods and beverages that are in accordance with national nutritional advice when having to compete with neighbouring facilities. With this easy access to SSB, both at school and in the school

British Food Journal Vol. 120 No. 8, 2018

pp. 1735-1748 © Emerald Publishing Limited

0007-070X DOI 10.1108/BFJ-11-2017-0605

Received 3 November 2017 Revised 24 April 2018

6 June 2018 Accepted 6 June 2018

The current issue and full text archive of this journal is available on Emerald Insight at: www.emeraldinsight.com/0007-070X.htm

1735

Self-control, intentions and

social influences

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

neighbourhood, it seems important to reveal other key influences on adolescents’ SSB consumption during school time.

Several studies have explored correlates of soft drink consumption in adolescence, and environmental factors such as availability and accessibility have consistently been identified as important predictors (De Coen et al., 2012; Gebremariam et al., 2016; Verloigne et al., 2012). Also family rules, food-related lifestyle, parenting styles and parenting practices have been found to be associated with SSB consumption in adolescents (Gebremariam et al., 2016; van der Horst et al., 2007; Verzeletti et al., 2009). Although parents and the home environment appear as important predictors of adolescents’ soft drink intake, their influence has to compete with the influence of friends, peers at school and media (Chan et al., 2012). Furthermore, eating behaviours seem (like many other types of behaviour) to be linked to the construction and expression of identity both at a personal and a social level (Bisogni et al., 2002; Vartanian et al., 2007), and particularly during adolescence, they may fulfil a function of self-expression (Guidetti and Cavazza, 2008). Apart from extrinsic (social) factors like parent and peer influences, also intrinsic (personal) factors such as the levels of self-control and impulsivity have been shown to correlate with adolescents’ consumption of SSB (Melbye et al., 2016) and other unhealthy foods (Honkanen et al., 2012). Since extrinsic and intrinsic factors are hypothesised to act in conjunction to influence human behaviour (Fox and Calkins, 2003; Ryan and Deci, 2000), they should not be seen in isolation from each other. Thus, it appears relevant to lean on elements from well-established socio-cognitive models such as Bandura’s (1977) social learning theory (SLT) and Ajzen’s (1991) theory of planned behaviour (TPB), as well as aspects of self-control theories, when exploring correlates of adolescents’ SSB consumption. While the socio-cognitive theories have their primary focus on extrinsic factors such as the context of social interactions, experiences and outside influences, self-control theories also include intrinsic factors such as temperament and inhibitory control (Vohs and Baumeister, 2016).

Based on the discussion above, the following variables were included in the current study: food-related self-control, intentions to consume SSB with school lunch, perceived descriptive peer norms related to SSB consumption with school lunch and perceived parental healthy eating guidance (HEG). Each of these variables is further described below.

Food-related self-control Individual differences in self-control have been associated with dietary behaviours in youth, whereby individuals with high levels of self-control exhibit healthier eating habits than those with lower levels of self-control (Junger and van Kampen, 2010; Luszczynska et al., 2013; Stok et al., 2015). Generally, self-control is described as the ability to resist temptations and impulses. However, the ability to resist temptations and impulses seems to depend on an individual’s overall self-controlling capacity, which has been shown to fluctuate across time (Baumeister et al., 1994). According to Baumeister et al.’s (2007) strength model of self-control, the fluctuation of self-controlling capacity may be caused by depletion of willpower (defined as a type of strength or energy), which is regarded as a limited resource. With other words: self-control appears vulnerable to deterioration over time from repeated efforts, like a muscle that gets tired. Thus, according to this model, if people rely on their willpower alone, they are likely to fail repeatedly because some temptations will inevitably turn up when one’s level of resistance is low (i.e. the muscle is tired). Interestingly, recent research suggests that self-control is also linked to avoiding temptations (Ent et al., 2015). Avoiding temptations may prevent instances of self-control failure caused by depleted willpower. Hence, avoidance appears as an important element of effective self-control. Yet, widely used measures of self-control such as Tangney et al. (2004) 36-item self-control scale, and the corresponding 13-item brief self-control scale (BSCS), do not explicitly distinguish between resistance and avoidance dimensions. Rather, they tend to include items reflecting

1736

BFJ 120,8

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

both resistance and avoidance (and other possible self-control dimensions) in calculating one total self-control score, thus presuming a unidimensional construct. According to Maloney et al. (2012), there is a general lack of well-validated measures of self-control, and despite the widespread use of Tangney et al.’s BSCS, questions about its unidimensionality and validity still remain. Maloney et al. (2012) have addressed these concerns, and their study is one of very few examining the dimensionality of the BSCS. Results from their analyses failed to support a one-factor structure, indicating that self-control is a construct with various facets. Thus, it seems important to further explore the dimensionality of this and other self-control measures.

Intentions According to traditional socio-cognitive models, such as Ajzen’s (1991) TPB, intention is the immediate precursor of behaviour. The TPB proposes that intentions result from the joint impact of attitudes, subjective norms and perceived behavioural control. According to Gollwitzer (1996), however, intentions are the end result of the deliberations of wishes and desires in the pre-decisional phase of decision making. Thus, wishes and desires are seen as important precursors of intention. There are several possible motivations underlying adolescents’ intentions to consume SSB with school lunch. Among them are hedonic motives and impression-management motives such as appearing autonomous (i.e. self-directed or independent) (Stok et al., 2010), “fitting in” or being popular among peers (De la Haye et al., 2010).

Descriptive peer norms For adolescents, the consumption of specific food products may help them create and present a desired identity and to express friendship (Stead et al., 2011). As a result, individual snack and soft drink consumption is shown to be higher when friends and peers have a high consumption (Wouters et al., 2010). Significant peer effects have also been found for the frequency of eating at fast food restaurants, suggesting that an individual is more likely to engage in these behaviours if his or her friends do (Ali et al., 2011). Further, a review by Salvy et al. (2012) reports that peers’ (actual) intake of snacks has been found to be a predictor of youths’ snack consumption. Interestingly, perceived peer behaviour (i.e. descriptive norms) is also shown to be important, and in some cases even more strongly associated with adolescents’ intake than the actual behaviour of the members of the peer group (Perkins et al., 2010). According to Lally et al. (2011), adolescents have a tendency to overestimate their peers’ intake of snacks and soft drinks and underestimate their consumption of fruits and vegetables. Furthermore, adolescents often feel a significant social pressure from their peers, and presenting a positive social profile is important to them (Stevenson et al., 2007). Accordingly, influence from peers seems as an important element for understanding adolescents’ SSB consumption.

Parents’ HEG A variety of food-related parenting practices are used to influence children and adolescents’ diet (Birch and Davidson, 2001; Blissett, 2011; Patrick and Nicklas, 2005). Previous studies have indicated that overly strict practices may have adverse effects on eating behaviours by increasing the consumption of unhealthy foods. For younger children, it is suggested that this may be the result of an increased appeal of the restricted foods that lead to over-consumption in situations where restrictions are removed (Birch et al., 2003; Fisher and Birch, 1999). For adolescents, it may represent an opposition to parental control, indicating an increased need for autonomy and self-presentation (Melbye et al., 2016; Parkin and Kuczynski, 2012). However, parenting practices involving positively framed eating

1737

Self-control, intentions and

social influences

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

guidance are shown to be associated with increased consumption of healthy food items and decreased consumption of unhealthy ones (Melbye and Hansen, 2015).

The aim of the present research was to explore the impact of relevant intrinsic and extrinsic factors on adolescents’ consumption of SSB with school lunch. More specifically, and based on the preceding literature review, we aimed to reveal the unique contribution of adolescents’ food-related self-control, behavioural intentions, perceived parent HEG and perceived descriptive peer norms in explaining the variance in SSB consumption in a school lunch context. An additional aim was to evaluate the psychometric properties of the self-control measure applied in the current study in order to reveal potential multi-dimensionality (i.e. resistance and avoidance dimensions).

Materials and methods Participants and procedures The first draft of the survey questionnaire was pre-tested by 15 second grade high school students from different schools in Rogaland. These students were recruited from the authors’ network (i.e. colleagues’ and friends’ children) and did not take part in the main survey. The pre-test was conducted in two steps. Initially, four students were asked to answer the questions orally, and to discuss all questions with an interviewer (one of the authors). Revisions were made according to their comments. Next, the questionnaire was pre-tested for clarity and length among 11 (other) students. The students filled in the questionnaire at home and sent it back to the researchers with comments. Based on these comments, a few minor revisions were made before the final draft was ready. The Norwegian Centre for Research Data, which is the Norwegian data protection officer for all Norwegian universities and several hospitals and research institutes, was notified about the study and approval was given before data collection started.

All high schools (n ¼ 25) in Rogaland County, Norway, were invited to participate in the web-based survey. Four schools situated in different parts of the county agreed to participate (two schools close to the city centre and two rural schools). The response rates among eligible first and second grade students at these schools were 14, 36, 68 and 72 per cent, giving 694 completed questionnaires. The school administration gave consent for their students’ participation in the survey, and participants were recruited through an invitation on the school website, with a link to the survey. Students were given information about the study, that participation was voluntary and anonymous and that completing the survey questionnaire implied their consent to participate. The survey was conducted using the EyeQuestion software version 3.15.3 (Logic8 BV, Wageningen, The Netherlands). The two schools with the highest response rates allowed the students to answer the survey during class (which was highly recommended by the investigators). If the survey was not finished in one session, it was possible for the respondent to return to the link and finish it at a later time. However, it was only possible to take it once. Students who missed answering some questions (n ¼ 142) were excluded from data analyses. The population of interest for this study was late adolescents (i.e. 16–19 years old), thus respondents younger than 16 and older than 19 years were excluded (n ¼ 20). The mean age in our sample was 17.1 years (SD ¼ 0.8), 67 per cent of the respondents were female and 94 per cent were of Norwegian or other Nordic origin.

Measures Consumption of SSB with school lunch was measured in days per week. Participants were asked to report their usual intake: “How many days per week do you normally consume SSB as part of your school lunch?” (SSB was defined as sugar-sweetened drinks such as carbonated soft drinks, lemonade, iced tea, sweet coffee drinks, energy drinks, milkshake/ chocolate flavoured milk, etc.). Diet drinks (i.e. drinks that do not contain sugar) were not

1738

BFJ 120,8

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

included in this measure. Since all Norwegian schools are closed during weekends, response alternatives ranged from 0 (never) to 5 days (every school day) per week. The variable was coded as follows: never ¼ 0, seldom¼ 0.5, one day per week¼ 1, two days per week¼ 2, three days per week¼ 3, four days per week¼ 4, five days per week¼ 5. SSB consumption with school lunch was later dichotomized into two categories to separate between students who do not (never or seldom) drink and students who drink SSB with school lunch: never/seldom ( ¼ 1) and 1 to 5 days per week ( ¼ 2).

Food-related self-control was measured by five items taken from Tangney et al.’s (2004) BSCS and made related to food by Honkanen et al. (2012): “I have a hard time breaking bad food habits”, “I wish I had more self-discipline when it comes to unhealthy food”, “Sometimes I can’t stop myself from eating unhealthy food, even if I know it’s wrong”, “I avoid food that is not good for me” and “People would say that I have good self-discipline when it comes to unhealthy food”. The first three items were reverse-coded so that higher numbers indicate stronger self-control.

Contemporary measurement research has suggested that for doubly concrete constructs, i.e., constructs that consist of concrete singular matters and concrete aspects related to these matters (here: SSB consumption with school lunch and intentions and descriptive peer norms related to SSB consumption with school lunch), tailor-made single-item measures may be advantageous (Ang and Eisend, 2017; Bergkvist, 2015; Bergkvist and Rossiter, 2007). Thus, adolescents’ intentions and (perceived) descriptive peer norms related to SSB consumption with school lunch were measured by one tailor-made item each: “I wish to drink SSB with my school lunch every day” and “My friends often drink SSB as part of their school lunch”.

Adolescents perceptions of their parents’ HEG were measured by an adapted version of Haszard et al.’s (2013) seven-item HEG scale, which is one of five subscales resulting from factor analysis on the Comprehensive Feeding Practices Questionnaire originally developed by Musher-Eizenman and Holub (2007). This scale has previously been adapted to a Norwegian context in Bjelland et al.’s (2014) research on family processes and dietary habits among Norwegian 13–15 years old and their parents, and further adapted to reflect adolescents’ perceptions of parents’ HEG in the present study. The HEG items read: “Most of the food my parents keep in the house is healthy”, “My parents encourage me to try new foods”, “My parents discuss with me why it is important to eat healthy foods”, “My parents tell me that healthy food tastes good”, “My parents discuss with me the nutritional value of foods”, “ My parents encourage me to eat a variety of foods”, “My parents model healthy eating for me by eating healthy foods themselves”, “My parents try to show enthusiasm when eating healthy foods” and “My parents show me how much they enjoy eating healthy foods”. All items measuring self-control, intentions, descriptive peer norms and HEG were scored on a five-point Likert scale ranging from 1 (totally disagree) to 5 (totally agree).

Participants were also asked to report their age in years, their gender (1 ¼ female, 2 ¼ male), their ethnicity (1 ¼ Norwegian/Nordic origin, 2 ¼ all other origins) and the highest completed level of education for their mothers and fathers (1 ¼ elementary school, 2 ¼ high school, 3 ¼ up to 4 years of college/university, 4 ¼ more than 4 years of college/university, 5 ¼ did not complete any education, 6 ¼ do not know). Parental education was dichotomized into two categories: “high school or less” ( ¼ 1) and “college/university education” ( ¼ 2).

Statistical analyses Statistical analysis started with psychometric evaluation of the self-control measure to reveal potential multi-dimensionality. Next, descriptive analyses and bivariate correlations were run. Finally, logistic regression analysis was run to explore associations between the independent variables and SSB consumption with school lunch.

1739

Self-control, intentions and

social influences

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

Psychometric evaluation of the self-control measure. Psychometric evaluation of the self-control items included factor analysis and calculation of internal consistency reliability (Cronbach’s α). We applied Tabachnick and Fidell’s (2007) recommendations for assessment of the appropriateness of our data for factor analysis: i.e. the presence of correlation coefficients greater than 0.3 in the correlation matrix, a KMO value of 0.6 or greater and a significant Bartlett’s test ( po0.05). Initial factor analysis using the Kaiser criterion (i.e. eigenvalue W1) and a scree plot test followed by factor analyses using different numbers of fixed factors were used to decide the number of factors to retain. Since the items were likely to be significantly correlated (i.e. they were all aiming to tap food-related self-control), oblique rotation was chosen to clarify the data structure (Costello and Osborne, 2005). Factor loadings of 0.4 or higher on allocated scale were used as a criterion for convergent validity, while cross-loadings of less than 0.4 on any other scale were used as a criterion for discriminant validity (Hair et al., 2010). Additionally, substantive evaluation based on the items’ face validity and previous research on the dimensionality of the self-control construct was used to arrive at the final factor solution. Because of a small number of items in our self-control measure, Cronbach’s α was classified as ⩾ 0.6¼ acceptable and ⩾ 0.07 ¼ good (Field, 2013). Before running further analyses, items were added up to produce averaged scores.

Descriptive analyses. Mean scores, standard deviations, skewness and kurtosis values were calculated for all study variables. As suggested by Kline (2005), we chose to apply cut-off values of 3.0 and 8.0 for skewness and kurtosis, respectively. Prior to model analyses, bivariate correlations were run to test for multicollinearity between independent variables. We applied a cut-off value of 0.80 or greater for multicollinearity, as suggested by Haerens et al. (2008).

Model analysis. Multiple logistic regression was used to explore the impact of the measured intrinsic (self-control dimensions and intentions) and extrinsic (perceived descriptive peer norms and parental HEG) factors on the consumption of SSB with school lunch. Since adolescent gender and parental educational level are well-known correlates of children and adolescents’ dietary behaviours, including soft drink consumption (Forshee and Storey, 2003; Vereecken et al., 2005), the regression model included adjustments for these variables.

Results Psychometric evaluation of the self-control measure Inspection of the correlation matrix for the five self-control items revealed the presence of several correlation coefficients of 0.3 and above. The KMO value was 0.69, and Bartlett’s test of sphericity showed statistical significance ( po0.001), supporting the factorability of our data. The Kaiser criterion suggested a two-factor solution, while the scree plot test suggested a three-factor solution. Based on this, we compared forced two- and three-factor solutions to decide how many factors to retain. In our sample, a two-factor solution reflecting the hypothesised subcomponents of self-control, resistance and avoidance, seemed more reasonable than a three-factor solution. The three items loading on the resistance factor made conceptual sense, as they all seemed to reflect the individual’s perception of his/her self-discipline regarding unhealthy eating/unhealthy food consumption (i.e. ability to resist temptations). However, the two items loading on the avoidance factor appeared conceptually different from each other as one item seemed to be a clear cut reflection of the individual’s avoidance of unhealthy foods (I avoid food that is not good for me), while the other item seemed to be a reflection of the individual’s perception of other people’s view on his/her self-discipline regarding unhealthy food (People would say that I have good self-discipline when it comes to unhealthy food). Thus, the latter item was removed, and a final factor solution with three items reflecting

1740

BFJ 120,8

D ow

nl oa

de d

by T

ex as

A &

M U

ni ve

rs it

y A

t 06

:2 8

01 N

ov em

be r

20 18

( P

T )

resistance and one item reflecting avoidance was reached (see Table I). This two-factor solution explained 79 per cent of the variance in our data, and Cronbach’s α for the resistance dimensions was 0.80 (it was not possible to calculate α for the avoidance dimension after removing one of two items in the final solution). A total (averaged) score for the resistance dimension was calculated, and the resulting new self-control variables named “resistance” and “avoidance” were used as independent variables in the logistic regression model.

Descriptive analyses Mean scores, standard deviations, skewness and kurtosis for the study variables are presented in Table II. Bivariate correlations are presented in Table III. All variables had values within the range of chosen cut-offs for skewness and kurtosis, and no multicollinearities were found for the independent …