937
British Journal of Health Psychology (2016), 21, 937-955 © 20/6 The British Psychological Society			The British Psychological Society
	www.wileyonlinelibrary.com		
Randomized controlled trial of a messaging intervention to increase fruit and vegetable intake in adolescents: Affective versus instrumental messages
Valentina Carfora * Daniela Caso and Mark Conner
1 Department of Humanities, University of Naples "Federico II", Italy School of Psychology, University of Leeds, UK
Objective. The present research aimed to test the efficacy of affective and instrumental text messages compared with a no-message control as a strategy to increase fruit and vegetable intake (FVI) in adolescents.
Design. A randomized controlled trial was used test impact of different text messages compared with no message on FVI over a 2-week period.
Method. A total of 1,065 adolescents (14-19 years) from a high school of the South of Italy completed the baseline questionnaire and were randomly allocated to one of three conditions: instrumental messages (N = 238), affective messages (N = 300), and no messages (N = 521). Students in the message conditions received one message each day over a 2-week period. The messages targeted affective (affective benefits) or instrumental (instrumental benefits) information about FVI. Self-reported FVI at 2 weeks was the key dependent variable. Analyses were based on the N = 634 who completed all aspects of the study.
Results. Findings showed that messages significantly increased FVI, particularly in the affective condition and this effect was partially mediated by changes in affective attitude and intentions towards FVI.
Conclusion. Text messages can be used to increase FVI in adolescents. Text messages based on affective benefits are more effective than text messages based on instrumental benefits.
Statement of contribution
What is already known on this subject?
Text messages have been shown to promote positive change in health behaviours. However, the most appropriate target for such text messages is less clear although targeting attitudes may be effective.
What does this study add?
• This randomized controlled study shows that text messages targeting instrumental or affective attitudes produce changes in fruit and vegetable intake (FVI) in adolescents.
• Text messages targeting affective attitudes are shown to be more effective than text messages targeting instrumental attitudes.
• The effect of affective text messages on FVI was partially mediated by changes in affective attitudes.
'^Correspondence should be addressed to Valentína Carfora, Via Porta di Massa, I, 80133 Naples, Italy (email: Valentína. carfora@unina.it).
DON0.1 11 l/bjhp. 12208
938   Valentina Carfora et al.
The eating habits acquired during adolescence are usually maintained into adulthood (Dietz, 1997), impacting on health in both the medium and longer term (Centers for Disease Control and Prevention, 2013). Therefore, promoting the development of healthy eating behaviours during adolescence is an important challenge in relation to promoting health. The present research tested a text-based intervention to increased fruit and vegetable consumption in adolescents.
It is recommended that individuals consume at least five servings of fruits and vegetables a day (USDA/USDHHS, 2010), as this level of fruit and vegetable intake (FVT) is the minimum necessary to provide the required nutrients (Mainvil, Lawson, Horwath, McKenzie, & Reeder, 2009; Pearson, Timperio, Crawford, & Biddle, 2009) and reduce health risks such as cancer and heart diseases (e.g., Dauchet, Amouyel, & Dallongeville, 2009; He, Nowson, Lucas, & MacGregor, 2007). Current FVI levels in adolescents tend to fall well below these recommended levels (OECD, 2012). For example, among teenagers in Campania, a region of South Italy, who represent the focus of the present research, only 7.6% of 15-year-olds and 55% of 13-year-olds eat more than one portion of vegetables per day and only 14.1% of 13-year-olds and 12.3% of 15-year-olds consume more than one portion of fruit per day (Angelini, Baldacchini, & Mignani, 2010).
Using text messages to change in health behaviours
As adolescents use text messaging (or SMS) as a preferred method of communication (Hingle, Nichter, Mederiros, & Grace, 2013; Lenhart, 2012), such messaging could form the basis of an effective strategy to promote health behaviours such as FVI (e.g., Kornman et al, 2010; Woolford, Clark, Strecher, & Resnicow, 2010). Generally, text messages have shown efficacy in reaching a large number of individuals quickly and with low cost (Kharbanda, Stockwell, Fox, & Rickert, 2009). For example, the study of Suffoletto et al. (2015) provided evidence that a text message intervention can reduce young adults' alcohol consumption. Text-based messages have a variety of characteristics (such as duration of intervention, frequency of messaging, timing of delivery) that may make them useful with groups such as adolescents. A recent systematic review of reviews (Hall, Cole-Lewis, & Bernhardt, 2015) reported that mobile text messaging interventions produced positive changes in a number of health behaviours. Moreover, a meta-analysis of 38 studies using text messages to change health behaviours indicated small-sized but significant effects (Orr & King, 2015). Although reviews (e.g., Shaw & Bosworth, 2012; Siopis, Chey, & Allman-Farinelli, 2015) have reported text-based message to promote weight loss, none has focused specifically on FVI in adolescents. Studies have shown other types of message to promote FVI (e.g., Beresford et al., 2001; Churchill & Pavey, 2013; Van Duyn et al, 2001). For example, Luszczynska, Tryburc, and Schwarzer (2007) showed e-mail messages targeting self-efficacy increased FVI in adults.
Affective and instrumental attitudes
The present research focused on testing the effectiveness of text messages targeting attitudes and their underlying outcome expectancies (Ajzen, 1991). Although text message studies have successfully targeted a number of correlates of behaviour change (e.g., Louch, Dalkin, Bodansky, & Conner, 2013 targeted attitudes, norms, and self-efficacy
Messages to increase fruit and vegetable intake 939
in order to promote appropriate insulin use in young adults) in studies using a limited number of text messages, it may be more appropriate to focus on one key factor. Attitudes and outcome expectancies in particular have been the focus of a considerable proportion of health messages. We follow this tradition. A focus on attitude can be further justified based on the fact that attitude is often the strongest predictor of intentions for health behaviours (McEachan, Conner, Taylor, & Lawton, 2011). In addition, attitude is a significant predictor of both intentions and behaviours even when controlling for other variables (e.g., Albarracin, Johnson, &2anna, 2005). In relation to FVI, a number of studies using the Theory of Planned Behaviour (TPB; Ajzen, 1991) have confirmed attitudes as key determinants of intentions to eat fruit and vegetables and other food choice behaviours (e.g., Blanchard, Fisher, et al., 2009; Blanchard, Kupperman, et at, 2009; Conner, Norman, & Bell, 2002; Godin et at, 2010; Kothe & Mullan, 2014; Kothe, Mullan, & Amaratunga, 2011), with some studies (Brug, Lechner, & De Vries, 1995; Lechner, Brug, & De Vries, 1997) reporting a direct associations between attitude and FVI.
In focusing on attitudes, we were interested to test the power of text messages targeting more traditional instrumental or cognitive outcomes like health with ones targeting more affective influences like feeling good. This allowed us to compare the power of affective text messages to increase FVI when compared to more traditional instrumental messages. Correlational data have indicated the importance of affective influences on health behaviours (e.g., Blanchard, Fisher, et at, 2009; Conner, McEachan, Taylor, O'Hara, & Lawton, 2015; Glasman & Albarracin, 2006; Lawton, Conner, & McEachan, 2009; Lawton, Conner, & Parker, 2007; McEachan et at, 2016; Nasuti & Rhodes, 2013; Rhodes, Fiala, & Conner, 2009; Trafimow et at, 2004), while intervention studies have indicated the power of affective messages to change health behaviours such as physical activity (Conner, Rhodes, Morris, McEachan, & Lawton, 2011; Morris, Lawton, McEachan, Hurling, & Conner, 2015; Sirriyeh, Lawton, & Ward, 2010; see Rhodes et at, 2009 for a review).
Present research
The present research tested the impact of affective and instrumental text messages about FVI compared with a no-message control condition (three conditions) on changing FVI in a sample of Italian adolescents using a randomized controlled trial (RCT). There is no similar existing study in adolescents, although other work has focused on determinants of FVI in this group (e.g., Lien, Lytle, & Komro, 2002), interventions to promote healthy eating more generally (e.g., French & Stables, 2003; Nicklas, Johnson, Myers, Farris, & Cunningham, 1998), or interventions for teenagers with eating problems, such obesity or overweight (e.g., Lubans et at, 2012; Woolford et at, 2011). The current focus on a text messaging intervention to increase FVI in adolescents is novel. We predicted that receiving a text message (affective or instrumental) would produce greater increases in FVI compared with no-message control condition. Based on previous work (e.g., Conner et at, 2011; Morris et at, 2015; Sirriyeh et at, 2010), we further predicted that affective text messages would be more effective than instrumental text messages in increasing FVI. Finally, we explored whether any effects of affective and instrumental text messages on FVI would be mediated by changes in affective or instrumental attitudes, respectively (cp. Conner et at, 2011), and by intentions. We took advantage of recent work on testing for serial mediation (Hayes, 2013) to explore whether intervention effects were mediated by either the relevant attitude construct (e.g., affective attitudes for affective text messages compared with control) or intentions and also whether the mediation was through the
940   Valentina Carfora eí al.
serial mediation path of intervention effect to attitude to intention to behaviour as would be predicted by the TPB (Ajzen, 1991).
Method
Sample
A high school in Naples (Campania region, South Italy) contacted the University of Naples "Federico II" to help implement a project promoting FVI in their students. The study was conducted in October 2014. To participate, students in the high school were required to be between 13 and 19 years of age and in possession of a personal mobile phone with an Internet connection. Participation involved completing questionnaires on two occasions separated by 2 weeks and in the message conditions receiving daily text messages over these 2 weeks. At time 1, a total of 1,225 students of 1,348 approached met eligibility criteria (age: M = 15.41, SD =1.53 years) and completed questionnaires before being randomized to affective, instrumental, or control group. A further 169 students subsequently refused to participate leaving 1,056 (297 in the affective group, nine refused to participate; 238 in the instrumental group, 68 refused to participate; 521 in the control group, 92 refused to participate). At time 2, a total of 634 (52% of original eligible sample) completed questionnaires (221 in the affective group, 76 lost to follow-up; 205 in the instrumental group, 33 lost to follow-up; 208 in the control group, 313 lost to follow-up). In part, the numbers lost to follow-up (particularly in the control group) was attributable to a school trip restricting participation. Figure 1 shows the flow of participants through each stage of this intervention.
Procedure
This study was conducted following receipt of ethical approval by the Department of Humanities of the University of Naples "Federico II" and by the board of the school. Students were contacted face-to-face by experimenters, who explained that the study was an examination of adolescents' healthy eating. Students were informed that the World Health Organization (2015) recommended eating at least five portions of fruit and vegetable per day (three portions of fruits and two of vegetables). A portion was specified as a fruit of medium size or 150 g, at least 50 g of raw vegetables, or 250 g of cooked vegetables. Participants then completed the questionnaire about FVI, provided demographic information and a mobile telephone number for the text messages, and created a personal code to allow matching of time 1 and time 2 questionnaires. At time 2, participants completed a second questionnaire. Questionnaires were administered in classroom time.
A randomization sequence was created using Excel 2007 (Microsoft). Participants were allocated to the three conditions in the ratio 2:1:1 for control, affective, and instrumental. This ratio was selected because we anticipated higher dropout rates in control condition. These randomization ratios lead only to a modest loss in statistical power (e.g., Dumville, Hahn, Miles, & Torgerson, 2006; Pocock, 1995; Torgerson & Campbell, 1997). The same allocation ratio has been employed in other text message studies (e.g., Suffoletto et al., 2015 on binge drinking in young adults).
Messages to increase fruit and vegetable intake
941
Eligible to l>e contact
Enrolment
Assignment
Assigned to affective group (n = JOG) Refused to participate
(h-P)
Total (/) = 297)
Assigned to instrumental aroup (m = 306) Refused to participate (n-6S)
Total (ft=238)
Analysis of affective soup («-221) Lost to follow-up
(h = 76)
1	/
Follow-up	
	
Anal ysis of instrumental group (n = 205) Lost to follow-up (h=33)	
Excluded (total n ~ 123) because Did not meet inclusion criteria
Assigned to control (« = 613) Refused to participate (,1 = 92)
Total (n=521J
Analysis of control group (»=208) Lost to follow-up (h= 313)
Figure I. Flow of participants through each stage.
Intervention
For those in the two text message conditions, messages were delivered each morning for 14 days before the start of school. To reduce the likelihood of cross-contamination, students were explicitly requested not to discuss the contents of messages with other participants during the intervention. The messages were sent by WhatsApp, a well-known cross-platform mobile messaging application. Message content was created based on previous research (e.g., McEachan, Lawton, Jackson, Conner, & Lunt, 2008; Sirriyehef al., 2010) and following guidelines (e.g., Hingle et al., 2013) about the content of the text message interventions (i.e., such as simple messages, addressing the reality of adolescent lifestyles). No messages were sent in the control condition.
In the affective condition, participants received text messages that focused on positive affective consequences of FVI (e.g., 'a diet rich in fruits and vegetables promote emotional well-being being associated with a lower prevalence of anxiety and depression'; 'the hormone of a 'good mood', namely serotonin, naturally produced by the body, is stimulated by the consumption of simple sugar, such as that contained in the fruit'; 'a diet rich in fruits and vegetables can help lower irritability, nervousness, and confusion caused by stress'). In the instrumental condition, participants received text messages that focused on the physical benefits of FVI (e.g., 'diet rich in vegetable and fruit reduces by 30-40% the probability of contracting cancer'; 'the fibre of fruits and vegetables is
942   Valentino Carfora et al.
essential for the proper functioning of the intestine and to reduce the absorption of excessive fat'; 'fruits and vegetables help the supply of useful minerals for metabolism, for the proper development of the organism and for the functioning of nerve impulses in the brain and heart'). See Appendix for full list of messages for each condition. The control group did not receive messages.
After completing the time 2 questionnaire, participants received feedback on the results of the study (the control group also received further information on the benefits of FVI to match that contained in the texts).
Measures
The following measures were taken. Demographics (gender and age) at Tl. Intentions (INT) to eat five portion of fruit and vegetable per day in the next two weeks were measured using three items on 7-point Likert scales (Armitage & Conner, 1999; e.g., 'I intend to eat at least five portions of fruit and vegetables per day over the next two weeks.. .definitely do not - definitely do'; all scored 1-7; with higher scores indicating stronger intention) at both time points. Cronbach's a was .91 and .94forTl and T2, respectively. Perceived behavioural control (PBC) was measured by seven items (Armitage & Conner, 1999; a; e.g., 'Over the next two weeks, will you feel to have personal control on eating five servings of fruits and vegetables a day?... strongly disagree - strongly agree'; all scored 1-7; higher scores indicated greater control) at both time points. Cronbach's a was .85 and .87 for Tl and T2, respectively. Affective attitudes were assessed by three items (Armitage & Conner, 1999; e.g., 'Eating at least five portion of fruit and vegetable per day in the next two weeks is... unenjoyable-enjoyable; unpleasant-pleasant; boring-exciting'; all scored 1-7; higher scores indicated more positive attitude) at both time points. Cronbach's a was .75 and .73 for Tl and T2, respectively. Instrumental attitude was assessed by three items (Armitage & Conner, 1999; e.g., 'Eating at least five portion of fruit and vegetable per day in the next two weeks is... not worthwhile-worthwhile; worthless-valuable; harmful-beneficial'; all scored 1-7; higher scores indicated more positive attitude) at both time points. Cronbach's a was .78 and .80 for Tl and T2, respectively. Subjective norm was assessed by four items (Armitage & Conner, 1999; e.g., 'People who are important to me want me to eat at least five portions of fruit and vegetables per day over the next 2 weeks.. .strongly disagree-strongly agree'; all scored 1-7; higher scores indicated greater subjective norm) at both time points. Cronbach's a was .80 and .75 for Tl and T2, respectively.
Self reported behaviour was assessed with two items (one for fruit consumption and second for vegetable consumption) and averaged. Response formats were formed by 6 points ranging from 1 portion to more than 5 portions per day (e.g., 'In the last two weeks how many portions of fruit do you eat per day?'). At time 1, this was used as a measure of past behaviour, while at time 2 it was used as the key dependent variable.
Data analysis
Analysis of intervention effects on FVI and all descriptive analyses were conducted in SPSS 170 (SPSS, Inc., Chicago, IL, USA). Multivariate, univariate, and chi-square analyses were used to check for any biases in randomization or dropout. MANCOVA was used to test for differences in FVI across conditions controlling for baseline (Tl) FVI. Finally, we used multiple mediation analyses, using bootstrapping in SPSS to test whether differences
Messages to increase fruit and vegetable intake 943
between conditions were mediated by affective and instrumental attitudes and intentions controlling for past behaviour.
Results
Preliminary analysis
Multivariate and univariate analyses at baseline (Tl) showed that there were no significant differences in intention, PBC, subjective norm, affective and instrumental attitudes, FVI, and age among the three conditions (p > .54). Chi-square also indicated that there were no significant differences in gender (p > .14). These findings suggested that randomization was adequate (Table 1) and the three conditions were matched on baseline measures.
Repeating these analyses, there were no differences in age and gender between those who did not meet inclusion criteria and those who participated to interventions (ps > .79). Comparing those who responded to questionnaires at both time points and those who dropped out between time 1 and time 2 indicated that there were no significant differences in age and gender and in the TPB variables or FVI at baseline (ps > .49). These results suggested that the final analysed sample was representative of the initial sample. Further analyses indicated that there were also no significant differences between those retained and lost to follow-up within each condition on the above variables (ps > .50).
Main analysis
Table 2 reports the mean scores on TPB variables at T2. MANOVA revealed a significant overall effect of condition, F(12, 1254) = 4.93; p = .001, r\2 = .04. Univariate Fs indicated significant effects of condition for intention, F(l, 634) = 16.32; p = .001, r|2 = .05, PBC, F(l, 634) = 3.74;p = .05, r\2 = .01, subjective norm, F(l, 634) = 5.06; p = .01, r|2 = .02, affective attitude, F(l, 634) = 8.45; p = .001, r\2 = .03, and instrumental attitude, F(l, 634) = 10.46; p = .001, r\2 = .03. Post-hoc tests (Tukey) showed that intentions were significantly higher in the affective condition than either of the control or instrumental conditions (ps < .001), which did not significantly differ from one another. Similarly,/>os£-^oc tests indicated that PBC (p < .05), subjective norm (p < .001), and affective attitude (p < .001) were significantly higher in the affective condition compared with the control condition and that no other differences were significant. Post-
Table I. Means and standard deviations of measured variables at TI in each condition
	Control		Affective		Instrumental	
	(N = 208)		(N = 221)		(N =	205)
Variables	Mean	SD	Mean	SD	Mean	SD
Intention	3.78	1.75	3.87	1.61	3.68	1.72
PBC	4.54	1.26	4.65	1.29	4.63	1.38
Subjective norm	4.27	1.42	4.16	1.32	4.06	1.43
Affective attitude	3.83	1.62	4.02	1.60	3.87	1.53
Instrumental attitude	5.31	1.42	5.45	1.49	5.25	1.55
FVI	3.83	1.72	3.76	1.55	3.95	1.78
Note. FVI = fruit and vegetable intake; PBC = perceived behavioural control.
944   Valentino Carfora et al.
Table 2. Means and standard deviations of measured variables at T2 in each condition
Control Affective Instrumental
Variables	(N =	208)	(N =	221)	(N =	205)
	Mean	SD	Mean	SD	Mean	SD
Intention	3.48	1.62	4.34	1.55	3.76	1.68
PBC	4.40	1.24	4.74	1.30	4.67	1.33
Subjective norm	3.89	1.36	4.29	1.31	4.19	1.34
Affective attitude	3.82	1.45	4.41	1.56	4.09	1.49
Instrumental attitude	4.97	1.75	5.62	1.44	5.52	1.43
FVI	4.52	2.31	5.49	2.00	5.1 1	2.07
Note. FVI = fruit and vegetable intake; PBC = perceived behavioural control.
hoc tests also showed that instrumental attitude was significantly higher in the affective condition compared with the control condition (p < .001) and significantly higher in the instrumental condition compared with the control condition (p < .001), although the affective and instrumental conditions did not significantly differ from one another.
MANCOVA on FVI at T2, controlling for past behaviour (Tl FVI), showed significant effects of the covariate, past behaviour, F(l, 664) = 265.68; p = .001, r\2 = .29, and a significant effect of condition, F(2, 664) = 2.7; p = .001; r\2 = .05. Estimated marginal means for FVI at T2 were 4.56 (Std. Error = . 12; 95% CI, 4.31-4.80) for the control group, 5.51 (Std. Error = .11; 95% CI, 5.28-5.73) for the affective group, and 5.05 (Std. Error = . 1; 95% CI, 4.80-5.30) for the instrumental group. Post-hoc tests (Tukey) revealed that controlling for past behaviour, there was a significant differences in FVI at T2 between the control and affective groups (mean difference = —0.98; p = .001; i.e., affective group higher than control), between the control and instrumental groups (mean difference = —0.60;p = .01; i.e., instrumental group higher than control), and between the affective and instrumental groups (mean difference = 0.39, p = 05; i.e., affective group higher than instrumental group).
Mediation analyses
Sequential mediation analyses, using bootstrapping in SPSS, tested whether controlling for past behaviour (FVI at Tl) the effect of intervention compared to the control group on FVI at T2 was sequentially mediated via attitude and intention (both at T2). For the comparison of affective messages against control conditions (Figure 2), we tested the mediating effect of the affective attitude (i.e., the attitude measuring corresponding to the messages), controlling for past behaviour and instrumental attitude and also controlling for instrumental condition against control condition. While for the comparison of instrumental messages against control conditions (Figure 2), we tested the mediating effect of instrumental attitude (i.e., the attitude measuring corresponding to the messages), controlling for past behaviour and affective attitude and also controlling for affective condition against control condition. The indirect effects were considered significant if bootstrapped 95% confidence intervals (CI) did not include zero. Figure 2 shows that for comparisons of the affective message against control conditions, each of the three mediated pathways was significant and the total indirect effect was significant (.41; 95% CI, 0.26-0.58). The path between conditions and FVI at T2 controlling for past
Messages to increase fruit and vegetable intake
945
Tl FVI
Figure 2. Mediation effects showing paths between variables: affective group versus control group and instrumental group versus control group. Note. All paths controlled for past behaviour; *p < .05, **p < .0l,***f> < .001. TI = time l;T2 = time 2. Values in parentheses indicate effect of condition on T2 fruit and vegetable intake (FVI) only controlling for Tl FVI.
behaviour was reduced by controlling for the mediators (Figure 2) but remained significant (.50; 95% CI, 0.18-0.82), indicating any mediation was partial. Examination of the mediated paths indicated that the strongest mediated effect was for the simple mediation path from condition to FVI via intention (.32; 95% CI, 0.19-0.47). The second strongest mediation path was the simple mediation path from condition to FVI via affective attitude (.06; 95% CI, 0.01-0.12). The sequential mediation chain from condition to FVI via affective attitude and then intention was the smallest effect, although also significant (.03; 95% CI, 0.01-0.08).
Figure 2 shows the results of the tests of mediation for the comparison of the instrumental versus control conditions. The only significant bootstrapped unstandardized indirect effect was for the sequential mediation chain from condition to FVI via instrumental attitude and then intention (.04; 95% CI, 0.01-0.08). The bootstrapped indirect effects were not significant for the simple mediation path from condition to FVI via instrumental attitude (—.00; 95% CI, —0.04 to 0.03) or the simple mediation from condition to FVI via intention (.04; 95% CI, —0.19 to 0.17). In addition, the total indirect effect was significant (.07; 95% CI, —0.06 to 0.22), the direct path between condition and FVI controlling for mediators, past behaviour and instrumental against control condition remained significant (.35; 95% CI, 0.03-0.68) and was not reduced compared to the path from condition to FVI not controlling for mediators, past behaviour (Figure 2).
Finally, we compared the impact of each message condition compared with control on T2 affective and instrumental attitudes controlling for baseline levels of these variables. The impact of the affective messages was larger on affective (r|2 = .04) compared with instrumental (r|2 = .03) attitudes and the reverse was true for the effect of instrumental messages (affective attitude r\2 = .01; instrumental attitude r\2 = .04).
946   Valentino Carfora et al.
Discussion
Affective versus instrumental attitudes intervention on FVI
The present research describes a test of the effects of different persuasive messages designed to increase FVI in high school students on changes in FVI over a 14-day period using an RCT. Participants at baseline reported an FVI below that recommended (i.e., at least five portions per day). Following the intervention, both the affective and instrumental groups significantly increased their daily FVI compared with control (i.e., to a mean level above the recommended minimum). Analyses indicated that these increases were significant in both the two message groups compared with a control no-message group. Therefore, adolescents exposed to the intervention increased their FVI. These findings are consistent with research indicating the efficacy of text messages in improving health behaviours (e.g., Hall et al., 2015; Louche? al., 2013; Siopisef al,2015} and with studies on the efficacy of paper and pencil messages (Beresford et al, 2001; Churchill & Pavey, 2013; Van Duyn et al, 2001) or computer-tailored messages to promote FVI (Brug, Glanz, Van Assema, Kok, & van Breukelen, 1998; Latimer et al., 2008; Luszczynska et al., 2007; Oenema, Tan, & Brug, 2005).
Additionally, when intervention groups were compared, the affective intervention was significantly more effective in increasing FVI compared with the instrumental group. Thus, affective text messages would appear to be the most efficacious intervention for this behaviour with this sample of adolescents. These findings support results that showed the importance of affective compared with instrumental attitudes in other health behaviours (e.g., Courneya & Friedenreich, 1997), such as physical activity (e.g., Conner et al, 2011; Morris et al, 2015), driving above the speed limit, and smoking cessation (Lawton et al., 2007). This extends these findings to healthy eating. Indeed, the findings presented here are consistent with research showing that for many health behaviours, affective-based messages may be more effective in generating behaviour change (Morris et al, 2015). The current research supports the contention that affective influences may be important causal determinants of health behaviours that could be usefully targeted in interventions.
Our mediation analyses (Figure 2) indicated that the effectiveness of instrumental messages compared with no messages was partially mediated by a sequential impact of messages on instrumental attitudes and intentions as the TPB would predict. Although it is worth noting that the total mediated effect was not statistically significant. More importantly, in the present context, the impact of the affective messages compared with no messages on FVI was partially mediated through three significant pathways (Figure 2). Consistent with the TPB was a serial mediation path through affective attitudes and intentions, although this was the smallest mediated path (and the one with the most links). The affective messages also impacted FVI through effects on intentions and through effects on affective attitudes. The latter pathway, in particular, is consistent with a growing number of studies showing affective attitudes to directly impact on health behaviours (controlling for intentions) in a way that instrumental attitudes do not (McEachan et al., 2016).
Limitation and methodological issues
There are several potential limitations for this study. First, the text messages employed in the current research might be criticized as not being purely affective or instrumental (see Appendix for all items). This is a common problem in this area (Conner et al, 2011;
Messages to increase fruit and vegetable intake 947
Sirriyeh et al., 2010) with generating realistic persuasive messages for use in real-world situations that are purely affective or instrumental in nature. Nevertheless, we would argue that at the very least, our affective messages had more of an affective 'flavour' than the instrumental messages and that the instrumental messages had more of an instrumental 'flavour' than the affective messages. Particularly, the analyses (Figure 2) showed that affective messages increased both affective and instrumental attitudes, which impacted on FVI at T2 through intentions. Importantly, instrumental messages did not generate a change in affective attitude. This may be considered to be as both a limitation in the creation of affective messages and a suggestion that affective messages, but not instrumental messages, improve both instrumental and affective attitudes towards healthy eating. Nevertheless, the most salient result is that affective attitude can be changed by affective messages and exert a direct effect (not via intentions) on post-intervention behaviour. This is consistent with suggestions from various correlational studies (Lawton et al., 2009; Lowe, Eves, & Carroll, 2002; Rhodes, Blanchard, & Matheson, 2006). These evidences provide some support for the affective and instrumental content of the messages.
Second, although the messages were matched as much as possible across groups, the messages on affective consequences of FVI were generally focused on the more short-term benefits when compared to the instrumental messages. In this regard, Hall and Fong (2007) note the importance of more immediate compared with temporally distant outcomes for health behaviours. They showed that dietary behaviours are viewed as connected to many proximal costs outcomes (e.g., inconvenience and loss of pleasure) and only weakly to distal benefits (e.g., longer lifespan and decreased risk for disease). Further research could seek to assess the relative importance of proximal versus affective outcomes in determining health behaviours (cf. Morris et al., 2015).
Third, the study used a passive control group, which received no text messages and was therefore not matched to other conditions on amount of contact time or the receiving of texts. We judged that it was more appropriate to send no texts to this group rather than texts irrelevant to FVI. It might be useful to confirm that our positive effects observed in both affective and instrumental conditions are not simply attributable to receiving a text, to increased contact with the experimenter or some enhanced Hawthorne effect. The differences between two message conditions would argue against some of these alternative explanations of our findings.
Fourth, the sample was restricted to a single city in the South of Italy; therefore, the results may not be generalizable to other samples. Fifth, the measurement of behaviour used simple self-report items. Thus, the findings need confirming with more objective measures. Although truly objective measures of daily food consumption are difficult to generate, confirming the present findings using more standardized measures of intake such as food diaries (e.g., Brouwer & Mosack, 2015) or Food Frequency Questionnaires (Willett & Lenart, 1998) would be useful. Sixth, the follow-up period was only 14 days. It would be valuable to verify the findings, with a longer follow-up period, to test whether the effects of these messages persist. This is particularly important given the fact that any health consequences of fruit and vegetable consumption are associated with prolonged periods of consumption. Finally, given that messages were sent each morning, there is the possibility that adolescents in different groups discussed the content of the messages despite instructions not to do so. Although this might be expected to lead to a reduction in the differences in conditions, it might be useful for future research to use a cluster randomized controlled design (with clusters based on not interacting groups, e.g., schools) to further test these effects.
948   Valentino Carfora et al.
Future direction and practical implications
Despite these methodological concerns, from an applied perspective, the current findings could support new policy recommendations based on the use of messaging interventions to promote healthy eating in a young people. Specifically, the current research underlined the need to emphasize enjoyment and social benefits of FVI rather than pointing to physical/health risks. More generally, in addition to proximal versus distal outcomes and affective versus instrumental outcomes, Rhodes and Conner (2010) point to the importance of positive versus negative outcomes. The present research would point to text messages particularly focusing on affective, proximal, and positive outcomes as a useful focus with adolescents in relation to FVI.
Finally, there is evidence about how hyper-responsivity to immediate contingencies, which is a high response to sensory input also described as 'sensory defensiveness' (Knickerbocker, 1980), impacts on human behaviour (e.g., Loewenstein, Read, & Baumeister, 2003) and about how individual differences in time perspective (e.g., Hall & Fong, 2003) are associated with health behaviours (e.g., Fong & Hall, 2003). It could be interesting for future research to consider the impact of other possible individual difference, such as levels of self-identity as healthy eater (e.g., Carfora, Caso, & Conner, 2015) and levels of risk awareness (Mullan, Wong, & Kothe, 2013). Furthermore, as adolescents spent several hours on Internet each day (e.g., Caso, 2015), it could be newsworthy to try to replicate this intervention using a social network or to differentiate the adolescents' FVI considering people with whom they share meals (Guidetti, Cavazza, & Graziani, 2014). Finally, future research could measure the intensity with which adolescents use the mobile phone or text messaging and monitor how much they read the sent messages and if they found them helpful.
Conclusion
In summary, the present research supported the hypothesis that it could be useful to use text messages that target affective beliefs (e.g., fun and good mood) in order to increase FVI rather than instrumental beliefs (e.g., reduction of cancer risks). Further research exploring affective influences on health behaviours in intervention studies could be worthwhile.
Conflict of interest
All authors declare no conflict of interest.
References
Ajzen, I. (1991)- The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211. doi:10.10l6/0749-5978(91)90020-t
Albarracin, D.Johnson, B. T., &Zanna, M. P. (Eds.)(2005). The handbook of attitudes. Mahwah, NJ: Lawrence Erlbaum. doi:10.1093/ijpor/edhl09
Angelini, P., Baldacchini, F., & Mignani, R. (2010). Stilt di vita e salute deigiovani in eta scolare. Rapporto sui dati regionali HBSC2009-2010 [Lifestyle and health of young people of school age. Report on regional data HBSC2009-2010]. Bologna, Italy: Centra Stampa Giunta Regione Emilia-Romagna, Direzione Generale Sanita e Politiche Sociali, Regione Emilia-Romagna.
Armitage, C. J., & Conner, M. (1999). Distinguishing perceptions of control from behavior: Predicting consumption of low-fat diet using the theory of planned behavior Journal of Applied Social Psychology, 29(X), 72-90. doi:10.1111/j.l559-18l6.1999.tb01375.x
Messages to increase fruit and vegetable intake 949
Beresford, S. A. A., Thompson, B., Feng, Z., Christianson, A., McLerran, D., & Patrick, D. L. (2001).
Seattle 5 a day worksite program to increase fruit and vegetable consumption. Preventive
Medicine, 320), 230-238. doi:10.1006/pmed.2000.0806 Blanchard, C. M., Fisher, J., Sparling, P. B., Shanks, T. H., Nehl, E., Rhodes, R. E.,... Baker, F. (2009).
Understanding adherence to 5 servings of fruits and vegetables per day: A theory of planned
behavior perspective./o«m«/ofNutrition Education and Behavior, 41(1), 3-10. doi:10.10l6/
j.jneb.2007.12.006
Blanchard, C. M., Kupperman, J., Sparling, P. B., Nehl, E., Rhodes, R. E., Courneya, K. S., & Baker, F.
(2009) ■ Do ethnicity and gender matter when using the theory of planned behavior to understand
fruit and vegetable consumption? Appetite, 52(1), 15-20. doi: 10.10l6/j.appet.2008.07.001 Brouwer, A. M., &Mosack, K. E. (2015). Expanding the theory ofplanned behavior to predict healthy
eating behaviors. Nutrition & Food Science, 45(1), 39-53. doi:10.1108/NFS-06-20l4-0055 BrugJ., Glanz, K., VanAssema, P., Kok, G., &vanBreukelen, G.J. P. (1998). The impact of computer-tailored feedback and iterative feedback on fat, fruit, and vegetable intake. Health Education &
Behavior, 25(4), 517-531. doi: 10.1177/109019819802500409 Brug, J., Lechner, L., & De Vries, H. D. (1995). Psychosocial determinants of fruit and vegetable
consumption. Appetite, 25, 285-296. doi: 10.1006/appe. 1995.0062 Carfora, V., Caso, D., & Conner, M. (2015). The role of self-identity in predicting fruit and vegetable
intake. Appetite, 106, 23-29. doi:10.10l6/j.appet.2015.12.020 Caso, D. (2015). L'uso di Internet e il benessere psicosociale in adolescenza: Uno studio
correlazionale [Internet use and psychosocial well-being in adolescence: A correlational study].
Psicologia delta Salute, 2, 141-155. doi: 10.3280/PDS2015-002008 Centers for Disease Control and Prevention (2013). Youth risk behavior surveillance - United
States. Retrieved from http://www.cdc.gov/healthyyouth/data/yrbs/index.htm Churchill, S., & Pavey, L. (2013). Promoting fruit and vegetable consumption: The role of message
framing and autonomy. British Journal of Health Psychology, 18, 610-622. doi:10.1111/
bjhp. 12007
Conner, M., McEachan, R., Taylor, N., O'Hara, J., & Lawton, R. (2015). Role of affective attitudes and
anticipated affective reactions in predicting health behaviors. Health Psychology, 34, 642-652.
doi: 10.1037/heaOOOOl43 Conner, M., Norman, P., &Bell, R. (2002). The theory of planned behavior and healthy eating. Health
Psychology, 21, 194-201. doi:10.1037/0278-6l33.21.2.194 Conner, M., Rhodes, R., Morris, B., McEachan, R., & Lawton, R. (2011). Changing exercise through
targeting affective or cognitive attitudes. Psychology & Health, 26, 133-149- doi: 10.1080/
08870446.2011.531570
Courneya, K. S., & Friedenreich, C. M. (1997). Relationship between exercise pattern across the
cancer experience and current quality of life in colorectal cancer survivors. The Journal of
Alternative and Complementary Medicine, 3, 215-226. doi:10.1089/acm.l997.3-215 Dauchet, L., Amouyel, P., & Dallongeville, G. (2009)- Fruits, vegetables and coronary heart disease.
Nature Reviews Cardiology, 6, 599-608. doi:10.1038/nrcardio.2009-131 Dietz, W. H. (1997). Periods of risk in childhood for the development of adult obesity - What do we
need to learn? The Journal of Nutrition, 127, 1884S-1886S. Dumville, J. C, Hahn, J. N. V., Miles, D. J., & Torgerson, D. J. (2006). The use of unequal
randomisation ratios in clinical trials: A review. Contemporary Clinical Trials, 27(1), 1-12.
doi: 10.10l6/j.cct.2005.08.003 Fong, G. T., & Hall, P. A. (2003). Time perspective: A potentially important construct for decreasing
health risk behaviors among adolescents. In D. Romer (Ed.), Reducing adolescent risk: Toward
an integrated approach (pp. 106-112). London, UK: Sage, doi: 10.4135/978145223361 l.nl3 French, S. S., & Stables, G. (2003). Environmental interventions to promote vegetable and fruit
consumption among youth in school settings. Preventive Medicine, 37, 593-610. doi: 10.1016/
j.ypmed.2003.09.007
950   Valentina Carfora et al.
Glasman, L. R., & Albarracin, D. (2006). Forming attitudes that predict future behavior: A metaanalysis of the attitude-behavior relation. Psychological Bulletin, 132, 778-822. doi:10.10l6/ j.ypmed.2003.09.007
Godin, G., Amireault, S., Bélanger-Gravel, A., Vohl, M., Pérusse, L., & Guillaumie, L. (2010).
Prediction of daily fruit and vegetable consumption among overweight and obese individuals.
Appetite, 54, 480^84. doi:10.10l6/j.appet.2010.01.018 Guidetti, M., Cavazza, N., & Graziani, A. R. (2014). Healthy at home, unhealthy outside: Food
groups associated with family and friends and the potential impact on attitude and
consumption. Journal of Social and Clinical Psychology, 33, 343-364. doi:10.1521/
jscp.2014.33.4.343
Hall, A., Cole-Lewis, H., & Bernhardt, J. M. (2015). Mobile text messaging for health: A systematic
review of reviews. Annual Review of Public Health, 36(1), 393—415. doi:10.1l46/annurev-
publhealth-031914-122855 Hall, P. A., & Fong, G. T. (2003). The effects of a brief time perspective intervention for increasing
physical activity among young adults. Psychology and Health, 18, 685-706. doi:10.1080/
0887044031000110447
Hall, P. A., & Fong, G. T. (2007). Temporal self-regulation theory: A model for individual health
behavior. Health Psychology Review, 1(1), 6-52. doi:10.1080/1743719070l492437 Hayes, A. F. (2013). An introduction to mediation, moderation, and conditional process analysis:
A regression-based approach. New York, NY: Guilford Press. He, F. J., Nowson, C. A., Lucas, M., & MacGregor, G. A. (2007). Increased consumption of fruit and
vegetable is related to a reduced risk of coronary heart disease: Meta-analysis of cohort studies.
Journal of Human Hypertension, 21, 717-728. doi:10.1038/sj.jhh.1002212 Hingle, M., Nichter, M., Mederiros, M., & Grace, S. (2013). Texting for health: The use of
participatory methods to develop healthy lifestyle messages for teens. Journal of Nutrition
Education and Behavior, 45(1), 12-19. doi:10.10l6/j.jneb.2012.05.001 Kharbanda, E. O., Stockwell, M. S., Fox, H. W., & Rickert, V. I. (2009). Text4Health: A qualitative
evaluation of parental readiness for text message immunization reminders. American Journal of
Public Health, 99, 2176-2178. doi:10.2105/ajph.2009.l6l364 Knickerbocker, B. M. (1980). A holistic approach to learning disabilities. Thorofare, NJ: Slack. Kornman, K. P., Shrewsbury, V. A., Chou, A. C, Nguyen, B., Lee, A., O'Connor, J., ... Baur, L. A.
(2010). Electronic therapeutic contact for adolescent weight management: The Loozit study.
Telemedicine ande-Health, 16, 678-685. doi: 10.1089/tmj.2009.0180 Kothe, E. J., & Mullan, B. (2014). A randomised controlled trial of a theory of planned behaviour to
increase fruit and vegetable consumption. Fresh facts. Appetite, 78, 68-75. doi:10.10l6/
j.appet.2014.03.006
Kothe, E. J., Mullan, B. A., & Amaratunga, R. (2011). Randomised controlled trial of a brief theory-based intervention promoting breakfast consumption. Appetite, 56(1), 148-155. doi:10.10l6/ j.appet.2010.12.002
Latimer, A. E., Williams-Piehota, P., Katulak, N. A., Cox, A., Mowad, L., Higgins, E. T., & Salovey, P. (2008). Promoting fruit and vegetable intake through messages tailored to individual differences in regulatory focus. Annals of Behavioral Medicine, 35, 363-369- doi:10.1007/sl2l60-008-9039-6
Lawton, R., Conner, M., & McEachan, R. (2009). Desire or reason: Predicting health behaviors from affective and cognitive attitudes. Health Psychology, 28(1), 56-65. doi:10.1037/a0013424
Lawton, R., Conner, M., & Parker, D. (2007). Beyond cognition: Predicting health risk behaviors from instrumental and affective beliefs. Health Psychology, 26, 259-267. doi: 10.1037/0278-6133.26.3.259
Lechner, L., Brug, J., & De Vries, H. (1997). Misconceptions of fruit and vegetable consumption: Differences between objective and subjective estimation of intake. Journal of Nutrition Education, 29, 313-320. doi: 10.10l6/s0022-3182(97)70245-0
Lenhart, A. (2012). Teens, smartphones & texting. Washington, DC: Pew Internet & American Life Project.
Messages to increase fruit and vegetable intake   951
Lien, N., Lytle, L. A., & Komro, K. A. (2002). Applying theory of planned behaviour to fruit and
vegetable consumption of young adolescents. American Journal of Health Promotion, 16,
189-197. doi: 10.4278/0890-1171-16.4.189 Loewenstein, G., Read, D., & Baumeister, R. (2003). Time and decision: Pconomic and
psychological perspectives on intertemporal choice. New York, NY: Russell Sage Foundation.
doi: 10.10l6/j.socec.2004.04.003 Louch, G., Dalkin, S., Bodansky, J., & Conner, M. (2013). An exploratory randomised controlled trial
using short messaging service to facilitate insulin administration in young adults with type 1
diabetes. Psychology, Health & Medicine, 18, 166-174. doi:10.1080/13548506.2012.68984l Lowe, R., Eves, F., & Carroll, D. (2002). The influence of affective and instrumental beliefs on
exercise intentions and behavior: A longitudinal analysis. Journal of Applied Social Psychology,
32, 1241-1252. doi:10.1111/j.l559-18l6.2002.tbOl434.x Lubans, D. R., Morgan, P. J., Okely, A. D., Dewar, D., Collins, C. E., Batterham, M.,... Plotnikoff, R. C.
(2012). Preventing obesity among adolescent girls: One-year outcomes of the nutrition and
enjoyable activity for teen girls [NEAT girls] cluster randomized controlled trial. Archives of
Pediatrics and Adolescent Medicine, 166, 821-827. doi:10.10l6/j.jsams.2012.11.806 Luszczynska, A., Tryburc, M., & Schwarzer, R. (2007). Improving fruit and vegetable consumption: A
self-efficacy intervention compared with a combined self-efficacy and planning intervention.
Health Education Research, 22, 630-638. doi:10.1093/her/cyll33 Mainvil, L., Lawson, R., Horwath, C. C, McKenzie, J. E., & Reeder, A. I. (2009). Validated scales to
assess adult self-efficacy to eat fruits and vegetables. American fournal of Health Promotion,
23, 210-217. doi:10.4278/ajhp.06l221154 McEachan, R. R. C, Conner, M., Taylor, N. J., & Lawton, R. J. (2011). Prospective prediction of
health-related behaviors with the theory of planned behavior: A meta-analysis. Health
Psychology Review, 5, 97-144. doi:10.1080/17437199.2010.52l684 McEachan, R. R., Lawton, R. J., Jackson, C, Conner, M., & Lunt, J. (2008). Evidence, theory and
context: Using intervention mapping to develop a worksite physical activity intervention. BMC
Public Health, S(l), 326. doi: 10.1186/1471-2458-8-326 McEachan, R., Taylor, N., Harrison, R., Lawton, R., Gardner, P., & Conner, M. (2016). Meta-analysis of
the Reasoned Action Approach (RAA) to understanding health behaviors. Annals of Behavioral
Medicine, 50, 592-612. doi:10.1007/sl2l60-0l6-9798-4. Morris, B., Lawton, R., McEachan, R., Hurling, R., & Conner, M. (2015). Changing self-reported
physical activity using different types of affectively and cognitively framed health messages, in a
student population. Psychology, Health & Medicine, 21, 198-207. doi:10.1080/13548506.
2014.997762
Mullan, B., Wong, C. L., & Kothe, E. J. (2013). Predicting adolescent breakfast consumption in the UK and Australia using an extended theory of planned behaviour. Appetite, 62, 127-132. doi:10.10l6/j.appet.2012.11.021
Nasuti, G., & Rhodes, R. E. (2013). Affective judgment and physical activity in youth: Review and meta-analyses. Annals of Behavioral Medicine, 45, 357-376. doi: 10.1007/sl 2160-012-9462-6
Nicklas, T., Johnson, C. C, Myers, L., Farris, R. P., & Cunningham, A. (1998). Outcomes of a high school program to increase fruit and vegetable consumption: Gimme 5 - A fresh nutrition concept for students, fournal of School Health, 68, 248-253. doi:10.1111/j.1746-1561.1998. tb06348.x
OECD (2012). Health at a glance: Europe 2012. Paris, France: Author, doi: 10.1787/978926418 3896-en
Oenema, A., Tan, F., & Brug, J. (2005). Short-term efficacy of a web-based computer-tailored nutrition intervention: Main effects and mediators. Annals of Behavioral Medicine, 29iX), 54-63. doi:10.1207/sl5324796abm2901_8
Orr, J. A., &King, R.J. (2015). Mobile phone SMS messages can enhance healthy behaviour: A metaanalysis of randomised controlled trials. Health Psychology Review, 9, 397—416. doi: 10.1080/ 17437199.2015.1022847
952   Valentino Carfora et al.
Pearson, N., Timperio, A., Crawford, D., & Biddle, S. J. H. (2009)- Family correlates of fruit and
vegetable consumption in children and adolescents: A systematic review. Public Health
Nutrition, 12, 267-283. doi:10.1017/sl368980008002589 Pocock, S.J. (1995). Clinical trials: A practical approach. Chichester, UK: John. Rhodes, R. E., Blanchard, C. M., & Matheson, D. H. (2006). A multi-component model of the theory of
planned behaviour. British Journal of Health Psychology, 11, 119-137. doi:10.1348/
135910705x52633
Rhodes, R. E., & Conner, M. (2010). Comparison of behavioral belief structures in the physical activity domain. Journal of Applied Social Psychology, 40, 2105-2120. doi:10.1111/j.l559-1816.2010.00652.x
Rhodes, R. E., Fiala, B., & Conner, M. (2009). A review and meta-analysis of affective judgments and
physical activity in adult populations. Annals of Behavioral Medicine, 38, 180-204.
doi: 10.1007/s12160-009-9147-y Shaw, R., & Bosworth, H. (2012). Short message service (SMS) text messaging as an intervention
medium for weight loss: A literature review. Health Informatics Journal, 18, 235-250.
doi:10.1177/1460458212442422 Siopis, G., Chey, T., & Allman-Farinelli, M. (2015). A systematic review and meta-analysis of
interventions for weight management using text messaging. Journal of Human Nutrition and
Dietetics, 28, 1-15. doi:10.1111/jhn.l2207 Sirriyeh, R. H., Lawton, R. J., & Ward, J. K. (2010). Physical activity and adolescents: An exploratory
randomised controlled trial (RCT) investigating the influence of affective and instrumental text
messages. British Journal of Health Psychology, 15, 825-840. doi:10.1348/135910710x
486889
Suffoletto, B., Kristan, J., Chung, T., Jeong, K., Fabio, A., Monti, P., & Clark, D. B. (2015). An interactive text message intervention to reduce binge drinking in young adults: A randomized controlled trial with 9-month outcomes. PLoS ONE, 10, e0l42877. doi:10.1371/ journal.pone.0142877
Torgerson, D., & Campbell, M. (1997). Unequal randomisation can improve the economic efficiency
of clinical trials. Journal of Health Services Research & Policy, 2(2), 81-85. Trafimow, D., Sheeran, P., Lombardo, B., Finlay, K. A., Brown, J., & Armitage, C.J. (2004). Affective
and cognitive control of persons and behaviours. British Journal of Social Psychology, 43,
207-224. doi: 10.1348/0144666041501642 U.S. Department of Agriculture and U.S. Department of Health and Human Services (USDA/
USDHHS) (2010). Dietary guidelines for Americans. Washington, DC: U.S. Government
Printing Office. Retrieved from http://www.cnpp.usda.gov/DietaryGuidelines Van Duyn, M. A. S., Kristal, A. R., Dodd, K., Campbell, M. K., Subar, A. F., Stables, G., ... Glanz, K.
(2001). Association of awareness, intrapersonal and interpersonal factors, and stage of dietary
change with fruit and vegetable consumption: A national survey. American Journal of Health
Promotion, 16, 69-78. doi: 10.4278/0890-1171-16.2.69 Willett, W., & Lenart, E. (1998). Reproducibility and validity of food-frequency questionnaires. In W.
Willett (Ed.), Nutritional epidemiology (2nd ed., pp. 101-147). New York, NY: Oxford.
doi:10.1093/acprof:oso/9780195122978.003.06 Woolford, S. J., Barr, K. L., Deny, H. A., Jepson, C. M., Clark, S. J., Strecher, V. J., & Resnicow, K.
(2011). OMG do not say LOL: Obese adolescents' perspectives on the content of text messages to
enhance weight loss efforts. Obesity, 19, 2382-2387. doi: 10.1038/oby.2011.266 Woolford, S. J., Clark, S. J., Strecher, V. J., & Resnicow, K. (2010). Tailored mobile phone text
messages as an adjunct to obesity treatment for adolescents. Journal of Telemedicine and
Telecare, 16(8), 458-461. doi:10.1258/jtt.2010.100207 World Health Organization (2015). Obesity and overweight. Fact sheet No. 311- Retrieved from
http://www.who.int/mediacentre/factsheets/fs311/en/
Received 2 December 2015; revised version received 29 April 2016
Messages to increase fruit and vegetable intake 953
Appendix : Text messages for each group
Day
Affective group
Instrumental group
A diet rich in fruit and vegetables promotes emotional well-being and is associated with lower levels of anxiety and depression
Serotonin, 'the hormone of happiness', is naturally produced by the body and stimulated by the consumption of simple sugars, such as those contained in fruit
A diet rich in fruit and vegetables can help lower irritability, nervousness, and confusion caused by stress
Five servings of fruit and vegetables a day will help you feel more energetic! Fruits and vegetables slow down digestion, ensuring a steady supply of energy during the day
Five servings of fruit and vegetables a day help reducing fatigue: try to replace sweets, which cause an increasing of the blood sugar level and its collapse, leading to fatigue and bad mood
Eating five servings of fruit and vegetables a day helps in preventing apathy, drowsiness and slow down!
Eating five servings of fruit and vegetables a day is associated with optimism! People who show a positive attitude towards everyday life have higher levels of carotenoids, which can be taken by eating fruit and vegetables
Eating five servings of fruit and vegetables a day not only is good for health but also promotes relationships: People who eat them are more beautiful. This was demonstrated by a British research, which showed that eating plenty of fruit and vegetables supposedly gives people a golden glow that makes them more attractive
An increased intake of fibre (typical in those who eat at least five portions of fruit and vegetables a day) is a protective factor against both colon cancer and breast cancer
With small, daily changes, such as eating extra fruit, the cause of premature death from any physical cause can be reduced by as much as 20%
Fruit and vegetables help supply the minerals needed for metabolic processes, the proper development of the organism, and the functioning of nerve impulses in the brain and heart
The fibres contained in fruit and vegetables are essential for proper intestinal functioning and for reducing excessive fats absorption
Yellow-orange vegetables (such as oranges, lemons, tangerines, grapefruit, apricots, peaches, carrots, peppers, squashes, loquats, corn, and melons) are rich in nutrients with antioxidant properties, fundamental for the health of bones and articulations
A diet rich in vegetables and fruit reduces the probability of cancer arising by 30^0%
Five servings of fruit and vegetables per day allow you to absorb calcium: It is essential for the bone tissue, the cardiovascular system, and blood clotting. You can find it not only in milk and cheese, but also in some green leafy vegetables (arugula and spinach) and in tree nuts (almonds and hazelnuts)
Numerous studies have shown that a diet rich in vegetables and fruit protects against the arising of numerous diseases widespread in developed countries, like different forms of cancer and cardiovascular diseases, and also cataracts, diseases of the respiratory, and digestive
Continued
954 Valentina Carfora et al. Appendix. (Continued)
Day
Affective group
Instrumental group
A study on 80,000 British people found that higher consumption of fruit and vegetables coincides with a higher sense of satisfaction and well-being. A diet rich in vegetables may have a beneficial effect on mood not only in the long term, but also day-to-day, influencing the positive emotions of individuals daily
10     Eating five servings of fruit and vegetables a day can increase your self-esteem. A healthy diet, combined with physical activity, will help you feel better about yourself Excessive consumption of junk food will make you gain weight and feel sluggish and sick. Eating the right foods, such as fruit and vegetable, will give you more energy and you feel calmer. If you feel better, you judge yourself with less severity
I I     Members of the University of Leiden in the
Netherlands found out that a substance that can be found in spinach (as well as in other fruits and vegetables, such as bananas, nuts, and avocados) is likely responsible for the production of two important neurotransmitters in the brain. These neurotransmitters facilitate faster transmission of nerve impulses, improving your reflexes and making you more snappy
12     Protein consumption is closely related to
the physical and mental well-being. Rich sources of protein are not only meat, fish, eggs, milk, and cheese, but also fruit and vegetables For example, the amino acid tryptophan (found in bananas, dried fruit, and vegetables such as spinach, cabbage, asparagus) has satiating and calming properties and modulates the mood and both emotional and sexual functions
13     A proper intake of vitamins is essential for
proper psychophysical functioning. Particularly, lack of vitamins of the B group (found in
systems (such as diverticulosis and constipation)
Eating at least five servings of fruit and vegetables a day helps prolonging the lifespan. The umpteenth confirmation comes from a study that involved 71,706 people, which spanned for 13 years: People who had never eaten fruit and vegetables lived 3 years less and had a mortality rate 53% higher than those who ate five servings a day
The low consumption of fruit and vegetables is responsible, worldwide, approximately for 31 % of coronary heart disease and for about I 1% of stroke. If every citizen of the European Union consumed at least 600 g of fruit and vegetables a day, more than 135,000 deaths a year from cardiovascular disease would be avoided
The beneficial effects of the consumption of five servings of fruit and vegetables are also linked to the protective effect of some of the components found in this class of food. These components, called 'radical scavengers', have mainly and antioxidant effect, and counteract free radicals, which are responsible for the processes of premature ageing and a series of reactions at the origin of various forms of cancer
The lack of vitamins can cause serious illnesses
Vitamin A helps sight and the integrity of the mucous membranes; it is contained in yellow-orange fruits and vegetables
Vitamin C (found in kiwis, oranges, mandarins, strawberries, tomatoes, peppers, etc.) has a key role in the integrity of blood vessels and helps the immune system in defending the body against infection; it is also contained in some green leafy vegetables Do not underestimate white fruit and vegetables (like garlic, white onion, cauliflower, fennel, pears, leeks,
Continued
Appendix. (Continued)
Messages to increase fruit and vegetable intake 955
Day
Affective group
Instrumental group
oranges, lemons, mandarins, apples, nuts, artichokes, carrots, zucchini ...) is associated with anxiety, fear, mood problems, dementia, and/or psychotic disorders
14     Recent researches have shown a strong
correlation between generous consumption of fruit and vegetables and the appearance of positive feelings in the next 24 hr, related especially to the reduction of anxiety and, more generally, to a greater serenity
mushrooms, and celery)! They are rich in substances capable of reinforcing the defences of our immune system and are able to lower cholesterol levels and blood pressure Vegetables, thanks to their therapeutic properties, can give various benefits: the strengthening of the immune system, regulation of blood pressure, cholesterol, and blood sugar, and prevention of cardiovascular diseases