Note: Correction added on 11 April 2013 after first publication online on 25 January 2013. The author, Samantha Millar, was originally included in the author list with her maiden name, Samantha Connor. This error has been corrected in this version of the article.
Evaluation of a theory of planned behaviour–based breastfeeding intervention in Northern Irish Schools using a randomized cluster design
Article first published online: 25 JAN 2013
© 2013 The British Psychological Society
British Journal of Health Psychology
Volume 19, Issue 1, pages 16–35, February 2014
How to Cite
Giles, M., McClenahan, C., Armour, C., Millar, S., Rae, G., Mallett, J. and Stewart-Knox, B. (2014), Evaluation of a theory of planned behaviour–based breastfeeding intervention in Northern Irish Schools using a randomized cluster design. British Journal of Health Psychology, 19: 16–35. doi: 10.1111/bjhp.12024
- Issue published online: 8 JAN 2014
- Article first published online: 25 JAN 2013
- Manuscript Accepted: 18 DEC 2012
- Manuscript Received: 23 MAY 2012
- R&D Office
The aim of this research was to evaluate the effectiveness of a school-based intervention designed to enhance young people's motivations to breastfeed.
A cluster randomized controlled trial was conducted involving 50 post-primary schools from across Northern Ireland. However, dropout and exclusion criteria utilized for the current study resulted in an effective sample size of 42 schools.
The intervention was delivered in two 35-min classroom sessions targeting those beliefs identified by the theory of planned behaviour (TPB) as significant in predicting motivation to breastfeed. Questionnaires incorporating the key components of the TPB were administered to all intervention and control schools at baseline, 1 and 6 months post-intervention. Multi-level modelling was used to analyse the data.
Findings suggest that the intervention was effective in that it increased females' intentions to breastfeed, expanded their knowledge and led to more favourable attitudes and perceptions of subjective norms. Notably, females' knowledge increased more in secondary schools than in grammar schools irrespective of whether they were control or intervention schools.
The research has provided evidence to support the use of the TPB in the design and evaluation of an intervention to increase females' intentions to breastfeed.
Statement of contribution
What is already known on this subject? The beneficial effects of breastfeeding are widely documented, and exclusive breastfeeding for the first 6 months of an infant's life continues to be encouraged. However, Northern Ireland still has the lowest rates of breastfeeding in Europe and one of the lowest rates of breastfeeding worldwide. The school system has the potential to positively promote breastfeeding, but work here is sporadic and uncoordinated. The theory of planned behaviour (TPB) is a useful framework for designing and evaluating behavioural change interventions, but as yet, no study has designed and evaluated the effectiveness of an intervention specifically targeted at those beliefs identified as important in predicting uptake of a particular behaviour.
What does this study add? This study provides evidence to support: The effectiveness of a randomized controlled trial in health psychology. The efficacy of the TPB as a useful framework for the design and evaluation of a behavioural change intervention. The school system as a means of promoting positive attitudes to breastfeeding.
The beneficial effects of breastfeeding are widely documented (Dyson et al., 2005), and exclusive breastfeeding for the first 6 months of an infant's life continues to be encouraged (WHO, 2010; World Health Organization/UNICEF, 2003). However, Northern Ireland still has the lowest rates of breastfeeding in Europe (HSC Public Health Agency, 2010) and one of the lowest rates of breastfeeding worldwide. For example, according to the Infant Feeding Survey conducted in 2010, 64% of mothers initiated breastfeeding compared to 81% in the rest of the UK. Whilst there is some evidence to suggest that there have been significant improvements in initiation rates in Northern Ireland in recent years (Bolling, 2006), the incidence of breastfeeding appears to have levelled off. For example, within the last 5 years rates have only risen from 63% to 64%. Initiation rates in England, Scotland and Wales, however, have continued to improve (NHS Information Centre, 2011).
Various interventions designed to enhance the initiation and duration of breastfeeding have been implemented. Fairbank et al. (2000) conducted a systematic review of the effectiveness of 59 such interventions and concluded that several activities appeared successful. These activities included small group health education classes, peer support programmes, media campaigns and structural changes in hospital practices. Perhaps surprisingly, none of these interventions focused specifically on the school setting despite numerous recommendations to promote breastfeeding through health classes (Dykes & Griffiths, 1998; Leffler, 2000; Lister-Sharpe et al., 1999; Littman, Medendorp, & Goldfarb, 1994; Lockey & Hart, 2003; Ross & Goulet, 2002; Spear, 2007, 2010; Stockley, 2002; Tarrant & Dodgson, 2007; Wambach & Cole, 2000). Indeed, the World Health Organization recommends that accurate information is provided through schools and other educational channels on the basis that breastfeeding is a learned behaviour (World Health Organization/UNICEF, 2003). These sentiments are echoed in several other policy documents where it is argued that breastfeeding education should formally be included as part of the national curriculum so that future parents have access to impartial information that is essential for informed decision-making (Health Promotion Agency, 2005).
Breastfeeding policy has been influenced to a large extent by research evidence to suggest that a dearth of observable role models leads to a neutral or negative attitude to breastfeeding (Health Promotion Agency, 2005). This is particularly pertinent in Northern Ireland where the majority of mothers have never fed outside their own homes (Campbell, 2000). This coupled with the finding that a child/young person who was breastfed themselves is likely to have more positive attitudes and intentions to breastfeed than one with limited experience and/or exposure has led to numerous recommendations to foster positive perceptions in the classroom. Indeed, according to Goulet, Lampron, Marcil, and Ross (2003), there is little time to challenge misconceptions during prenatal classes, and this may well be too late given there is also evidence to suggest that decisions about infant feeding are often made prior to pregnancy (Earle, 2000; Hoddinott & Pill, 1999; Losch, Dungy, Russell, & Dusdieker, 1995; Stewart-Knox, Gardiner, & Wright, 2003; Wambach & Cole, 2000).
Whilst there is some evidence to suggest that the school system has an important role to play in this respect and thus could be used to challenge the cultural view of breasts and promote the idea that breastfeeding is normal, natural and beneficial (Stockley, 2002), work in this area is limited (Health Promotion Agency, 2005). Many studies have highlighted the importance of various theoretical constructs, for example, interventions focused on enhancing self-efficacy have resulted in an increase in breastfeeding rates (Noel-Weiss, Rupp, Cragg, Bassett, & Woodend, 2006; Schlickau & Wilson, 2005). However, most interventions have lacked a theoretical focus.
Elsewhere, it is acknowledged that social cognition models including the theories of reasoned action and planned behaviour (Ajzen, 1985; Ajzen & Fishbein, 1980) have much to offer in the context of health promotion. Indeed, there is some evidence to suggest that the theory of planned behaviour (TPB) in particular is a useful framework for designing and evaluating behavioural change interventions (Hardeman et al., 2002). Moreover, numerous studies have supported the application of the TPB to breastfeeding and have shown that its core constructs are useful predictors of one's decision to breastfeed (Dodgson, Henley, Duckett, & Tarrant, 2003; Duckett et al., 1998; Giles, Connor, McClenahan, & Mallett, 2010; Giles et al., 2007; McMillan et al., 2008, 2009; Rempel, 2004; Wambach, 1997). However, as yet, no study has designed and evaluated the effectiveness of an intervention specifically targeted at those beliefs identified as important in predicting uptake.
To use the TPB in this context, it is recommended that one-first conducts an elicitation study to identify the most salient beliefs underlying the behaviour and uses these to construct a questionnaire that can then be administered to a sample of people drawn from the target population to identify the most important determinants of the target behaviour (Hardeman et al., 2002; Sutton, 2002). The intervention can then be designed based on the questionnaire components most strongly associated with the target behaviour, and the questionnaire can be used pre- and post-intervention to evaluate its effectiveness. The intervention described here was designed to enhance young people's motivations to breastfeed and is based on elicitation work previously conducted (Giles et al., 2007, 2010). Consistent with some previous research (Armitage & Conner, 2001), this work employed separate measures of perceived control and self-efficacy and provided evidence to suggest that the TPB constructs of attitude, subjective norm and self-efficacy are useful predictors of young people's intentions to breastfeed. These findings appear to be consistent with those of others (Dodgson et al., 2003; Duckett et al., 1998; McMillan et al., 2008, 2009; Rempel, 2004; Wambach, 1997) and, as such, together with knowledge served as targets for this school-based intervention.
Whilst there does not appear to be a consensus regarding the best age at which to introduce breastfeeding into the curriculum (Health Promotion Agency, 2005), with some suggesting a focus on primary-school-aged children given that teenagers may already have made up their minds (Fletcher Williams & Tappin, 2002), others have advised that interventions should be directed towards young adolescents as this is an impressionable age during which they can be persuaded to think and act differently (Goulet et al., 2003; Greene, Stewart-Knox, & Wright, 2003). Given the practicalities of conducting research in educational settings (e.g., in terms of the restraints imposed by the national curriculum for examinations and assessment) and the cultural issues associated with research of this nature, early adolescence was deemed the most appropriate time to intervene. This project has been a short-term longitudinal evaluation of this work, with post-intervention follow-ups at 1- and 6-month intervals.
More specifically, the focus of the current study was to (1) test a breastfeeding programme with a cohort of 13-/14-year-old school children and (2) evaluate the programme, exploring key changes in social cognition variables such as attitude, confidence and motivation to breastfeed. To fulfil these objectives, all analyses were implemented using multi-level modelling (MLM) (also referred to as hierarchical linear modelling – HLM). The data were consistent with a three-level model, whereby level 1 represented a number of repeated measures, which were nested in pupils (level 2), who were nested in schools (level 3). Given the primary aim of our study, to test the effectiveness of the intervention, the focal predictor was a level 3 dummy-coded variable that denoted whether or not a particular school received the intervention (0 = control school; 1 = intervention school). We hypothesized that the intervention would significantly improve students intention to breastfeed, their feeling and moral-based attitudes towards breastfeeding, their self-efficacy, subjective norm and knowledge about breastfeeding.
The secondary aim of this study was to determine whether the intervention was more or less effective given the type of school. Thus, we aimed to assess school type (grammar vs. secondary), as well as the two-way interaction between school type and treatment (intervention vs. control). Given lower rates of breastfeeding initiation and duration are often reported among women from lower SES backgrounds (Bolling, 2006; Bolling, Grant, Hamlyn, & Thornton, 2007; Hamlyn, Brooker, Oleinikova, & Wands, 2002; Misra & James, 2000) and children enrolled in the secondary school system in Northern Ireland are more likely to be from lower income families (HSC Public Health Agency, 2010), we hypothesized that the intervention would be more effective in secondary schools compared with grammar schools. For example, given the lower base rate of knowledge expected in secondary schools (based on a higher prevalence of students from less socially advantaged backgrounds), we hypothesized that there would be a significantly greater increase over time in knowledge levels for secondary school students compared with grammar school students.
Participants and procedures
Forty-two post-primary schools including 31 secondary and 11 grammar from across Northern Ireland participated in this study. One class of 13-/14-year-old pupils was selected from within each school (average class size was 17), comprising 698 participants in total. Although a randomized control trial with individual pupils as the unit of analysis was the preferred study design, this was not possible given the likely contamination effects within schools and classes. Therefore, a cluster randomized design was employed using schools as the unit of randomization.
All 228 post-primary schools in Northern Ireland were contacted by letter and invited to participate. Fifty schools initially agreed to take part, a response rate of 22%. Allocation of schools was managed by the York Clinical Trials Unit who carried out independent randomization to either an intervention or a control group using a 2:3 ratio. A computer program used a randomly generated number to randomly order the list of schools in the group. It then assigned the first 40% of the schools to the intervention allocation and the other 60% to the control allocation. A 2:3 ratio was chosen due to time and financial constraints. Minimization (also known as restricted allocation) was employed as the method of allocating schools to conditions. Baseline characteristics used in the minimization strategy were Education and Library Board Area, type of school (secondary or grammar) and gender of school (mixed sex, single sex girl, single sex boy). The aim of minimization was to balance the groups with respect to baseline factors and also to increase statistical power as the stratifying factors were strong predictors of outcomes (Torgerson & Torgerson, 2007).
One intervention school later withdrew from the study due to a change of principal; another school failed to maintain contact with the research team. Two control schools also withdrew from the study due to time constraints. Attrition bias at the school level was therefore 8% as four clusters (schools) dropped out after allocation. A further two schools, both of which were control schools, were excluded at this stage of the analyses as they were ‘all boys’ schools and the focus here was on females. This left the total number of participating schools as 44:18 in the intervention group and 26 in the control group. Schools allocated to the intervention group received two 35-min educational sessions delivered by a member of the research team, as part of the Year 10 Home Economics education curriculum. Control schools continued to receive the standard Home Economics curriculum without the additional input (Figure 1).
Designing the intervention
An elicitation study was first conducted, comprising a series of six semi-structured focus groups with 48 young people. Following the advice of Ajzen (1985), a series of nine questions were used to structure the focus group sessions (Francis et al., 2004). The beliefs elicited were used to construct a questionnaire, which was subsequently piloted on a sample of 121 school children (Giles et al., 2007). The resulting questionnaire was then administered to 2021 year 10 children from 36 post-primary schools in Northern Ireland. Data were analysed using correlation and regression analyses and the results used to inform the design of the intervention (Giles et al., 2010).
The intervention was delivered in two 35-min classroom sessions involving a series of PowerPoint slides originally developed by the Health Promotion Agency (NI). Whilst feedback from 39 teachers and 90 pupils suggested that these materials were both appealing and appropriate (Health Promotion Agency, 2008), they were subsequently modified by the research team to incorporate the specific cognitions identified as significant during the design phase. Specifically, given the importance attached to attitudes, self-efficacy, knowledge and the role of one's significant others in the questionnaire survey (Giles et al., 2010), the materials were modified to emphasize the naturalness of breastfeeding, the importance of the family network in encouraging and supporting a woman to breastfeed and provided information about the consequences of the behaviour (e.g., in terms of the health benefits to both mother and child). This was mostly achieved via visual images and text incorporated in the PowerPoint slides but also included other materials, for example worksheets, all of which were designed to be interactive where possible to engage participants and encourage discussion. A more detailed description of the intervention using the taxonomies of behaviour change techniques (Abraham & Michie, 2008) is displayed in Table 1. The resulting intervention was delivered by a member of the research team as part of the year 10 Home Economics curriculum. This researcher was involved in all aspects of the research process including the design phase and had been briefed to deliver the intervention in a standardized way. Given the sample was designated a vulnerable group, it was also a requirement that the researcher was police-checked. Further, a responsible adult (e.g., teacher) was present in the classroom during the delivery of the intervention.
|Behaviour change technique||Target variables||Procedures and materials|
|Provision of general information linking behaviour to health||GN, ATT||Session 1: PP slides covering general introduction and history of infant feeding, environmental issues, and worldwide facts and figures|
|Session 2: PP slides covering health benefits, composition of breast versus bottled milk and the physiology of how breastfeeding works|
|Provide information on consequences||GN, ATT||Session 1: PP slides covering general health benefits to both mother and child, practical and financial benefits|
|Session 2: PP slides covering specific health benefits to both mother and child. Facts and figures provided. Also the importance of skin-to-skin contact, bonding, etc. Short audio extract (local radio advertisement) highlighting the benefits of breastfeeding|
|Provide information about others approval||SN||Session 1: PP slides covering ways in which fathers, family and wider social circle can be involved in breastfeeding, further supported by visual images of significant others|
|Session 2: Two short audio extracts (local radio advertisements) highlighting the need for breastfeeding to be accepted in public places|
|Prompt intention formation||INT||Session 1: Encouragement provided throughout to persuade participants that breastfeeding is best for both mother and child|
|Session 2: Positive images depicting the naturalness of breastfeeding and the bonding process|
|Participants asked to indicate whether they would consider breastfeeding if they have a baby|
|Prompt barrier identification||ATT||Session 1: PP slides covering potential barriers including social attitudes (embarrassment), lack of public facilities and availability of bottled milk. Images depict the discrete way in which breastfeeding can be performed|
|Session 2: Discussion on potential barriers and ways they may be overcome (e.g., avoiding use of dummies). Demonstration of breast pump|
|Provide general encouragement||INT, SE||Session 1: Verbal encouragement provided throughout the session together with 1-min television advertisement summarizing the positive aspects (e.g., bonding, health benefits, father involvement, ease and discrete way it can be performed)|
|Session 2: Verbal encouragement provided to support instruction provided (see below) and to highlight issues depicted in visual images (see intention formation above)|
|Set graded tasks||INT||Session 2: Discussion surrounding what to expect at different stages, recommended time to breastfeed, etc.|
|Short breastfeeding quiz completed with participants|
|Provide instruction||SE||Session 2: PP slides on how to breastfeed successfully in terms of positioning and attachment, mother support groups, etc.|
|Model/demonstrate the behaviour||SN, SE||Session 1: Visual images and posters of positive role models breastfeeding in a variety of locations. One-minute television advertisement showing a woman breastfeeding with the father at her side|
|Session 2: Visual images demonstrating successful attachment in a variety of positions|
|Prompt identification as role model||SN||Session 1: Images portrayed in PP slides of young girls' breastfeeding and seeming to enjoy the experience|
|Plan social support/social change||SN||Session 1: Discussion focused on role of the extended family and various support groups available. The role fathers can play is emphasized. Information, images and discussion about breastfeeding in public places|
|Session 2: Advice given reuse of support groups to help make breastfeeding work|
Pre-testing the intervention
The intervention was piloted in six schools to ensure that it was acceptable to both teachers and pupils and could be implemented in a reasonably standardized way. This also involved asking teachers a range of questions including, for example, ‘What are your initial impressions of the breastfeeding materials?’ and ‘Is the content appropriate for this age group?’ The intervention was generally well received at this stage. For example, several teachers reported that the session had good visual appeal and made specific reference to the graphics and cartoons employed. Some focused specifically on certain aspects of the intervention: ‘I think it is vital that this topic is addressed early in school education so that the embarrassment element is reduced’ and ‘I particularly liked the clear photographs of mothers breastfeeding in public which are not embarrassing’. However, some suggestions were made as to how the sessions could be made more effective, and as a result, some minor changes were implemented. Specifically, some additional materials were included to stimulate thought and discussion (e.g., use of a breast pump), and minor amendments were made to the PowerPoint presentation (e.g., in terms of language used). Perhaps more importantly, it became apparent that two school visits would be difficult given the already extensive nature of the Home Economics curriculum. Following some consultation with teachers, it was therefore agreed to deliver the two 35-min sessions at one point in time.
Evaluating the intervention
Questionnaires incorporating the direct measures of the Ajzen and Fishbein (1980) framework were administered to 18 intervention and 26 control schools at baseline (Figure 1) and again at 1 and 6 months post-intervention to evaluate its effectiveness. The primary outcome measure was intention to breastfeed, which was assessed using the three items typically employed in TPB-based studies (Francis et al., 2004). Participants were required to consider the possibility that they would have a baby at some time in the future and respond to items accordingly (Table 2). Secondary outcome measures included both affective and instrumental versions of attitude, knowledge and a measure of subjective norm and self-efficacy. Although not a direct focus of the intervention, a measure of perceived control was also included.
|Construct||n items||Sample items||Cronbach's alpha coefficient|
|Intention||3||I intend (will try/have decided) to breastfeed if I ever have a baby: extremely likely (7)/extremely unlikely (1); strongly agree (7)/strongly disagree (1)||.94|
|Feeling-based attitude||3||Breastfeeding a baby would be: extremely pleasant; positive; enjoyable (7)/extremely unpleasant; negative; unenjoyable (1)||.90|
|Moral-based attitude||3||Breastfeeding a baby would be: extremely good; wise; right (7)/extremely bad; foolish; wrong (1)||.91|
|Subjective norm||3||Most people who are important to me think I should (would approve/would want me to) breastfeed if I ever have a baby: extremely likely (7)/extremely unlikely (1)||.87|
|Perceived control||3||Whether or not I breastfeed if I ever have a baby is entirely up to me: strongly agree (7)/strongly disagree (1); How much personal control do you feel you have over breastfeeding a baby?: complete control (7)/no control (1); How much do you feel that breastfeeding is beyond your control: not at all (1)/very much (7)||.55|
|Self-efficacy||3||I believe I have the ability/am capable/am confident that I will be able to breastfeed if I ever have a baby: definitely do (7) - do not (1)/capable (7) – not at all (1)/confident (7) – not at all (1)||.83|
|Knowledge||14||Babies who are bottle-fed have more illnesses than babies who are breastfed; Small breasts will not produce enough milk; Women who breastfeed should avoid certain foods||.76|
Preliminary analyses examined the dropout rate in the present sample and the correlations between level 1 variables across the waves of data collection.
Initially, 698 participants completed baseline measures. Of these participants, 540 (77.6%) also completed the 1-month follow-up assessment, whilst 158 (22.6%) did not. Of the 540 participants who completed both the baseline and 1-month follow-up assessment, 434 (80.4%) also completed the 6-month assessment, whilst 106 (19.6%) dropped out. Of the participants who completed all three assessments, 153 (35.3%) were 13 years old, 275 (63.4%) were 14 years old, and 6 (1.4%) were 15 years old.
Correlation across waves
All dependent variables were significantly correlated from Wave 1 to Wave 2 and Wave 3, as shown in Table 3.
|Wave 1||Wave 2/Wave 3|
|Perceived Control (PC)||.397/.356||.561/.478||.537/.464||.402/.377||.419/.417||.367/.374||.317/.320|
|Moral Attitude (MATT)||.712/.681||.466/.407||.576/.560||.660/.671||.762/.717||.631/.645||.395/.373|
|Subjective Norm (SN)||.654/.637||.398/.412||.472/.486||.579/.603||.654/.649||.742/.738||.414/.382|
Multi-level modelling was implemented to evaluate the programme by exploring key changes in social cognition variables such as attitude, confidence and motivation to breastfeed. The intra-class correlations, which signify the degree of resemblance between randomly drawn micro-level units belonging to a single macro-unit (Snijders & Bosker, 2012), were calculated from the intercept only model, containing no explanatory variables. Intra-class correlations were .120 for intention, .079 for feeling-based attitudes, .081 for moral-based attitudes, .060 for self-efficacy, .017 for perceived control, .090 for subjective norm and .170 for knowledge. These indicated that the analysis needed to take account of the clustered nature of the data. MLM is able to take account of this dependency by estimating the variance associated with group differences (intercepts) and group differences in associations (slopes) between predictors and dependent variables (Snijders & Bosker, 2012; Tabachnik & Fidell, 2007). MLM was preferred over repeated measures analysis of variance methods as it does not require sphericity (Quené & van den Bergh, 2004) and is better able to deal with missing data as participants are not required to provide data at all time points to be included in the analysis.
Level 1 variables
First-level units were the repeated measures of the total scores of each of the constructs: intention to breastfeed, feeling-based attitudes towards breastfeeding, moral-based attitudes towards breastfeeding, perceived control over breastfeeding, self-efficacy regarding breastfeeding, subjective norm and knowledge about breastfeeding. The explanatory variable at this level was time. Three points in time were included – baseline and 1- and 6-month follow-ups, coded as 0, 1 and 2. Table 4 shows descriptive statistics for each dependent variable on the control and intervention groups. Effect sizes for the intervention were computed from the HLM model, rather than from these analyses, as the HLM model takes the nested structure of the data into account.
|Variable||Control M (SD)||Intervention M (SD)|
|Intentions||4.72 (1.85)||5.08 (1.88)|
|Perceived control||5.57 (1.00)||5.67 (1.11)|
|Self-efficacy||4.93 (1.43)||5.38 (1.34)|
|Attitude||4.91 (1.25)||5.07 (1.25)|
|Moral attitude||5.66 (1.21)||5.74 (1.36)|
|Subjective norms||5.19 (1.37)||1.59 (5.22)|
|Intentions||4.71 (1.81)||5.43 (1.71)|
|Perceived control||5.64 (1.06)||5.80 (0.98)|
|Self-efficacy||5.11 (1.39)||5.64 (1.19)|
|Attitude||4.94 (1.26)||5.46 (1.22)|
|Moral attitude||5.53 (1.27)||6.03 (1.11)|
|Subjective norms||5.16 (1.40)||5.60 (1.41)|
|Knowledge||0.58 (0.23)||0.82 (0.17)|
|Intentions||4.78 (1.83)||5.51 (1.65)|
|Perceived control||5.58 (1.13)||5.82 (1.09)|
|Self-efficacy||5.25 (1.44)||5.68 (1.19)|
|Attitude||5.00 (1.33)||5.58 (1.25)|
|Moral attitude||5.52 (1.29)||6.06 (1.15)|
|Subjective norms||5.21 (1.40)||5.54 (1.42)|
|Knowledge||0.60 (.24)||0.82 (0.16)|
Level 2 variables
Second-level variables were the cases, with each case representing a participant/pupil in the study. Whilst no person-level explanatory variables were included in the model, it was necessary to include this level to maintain the hierarchical structure that was inherent in this nested data. The sample was restricted to females who had complete data at all three time points, resulting in a sample size of 698 individuals. Participants came from a restricted age range (13–15), so age was not entered as a level 2 predictor.
Level 3 variables
The study was conducted in 44 schools (18 intervention and 26 control). However, when the sample was restricted to cases with complete data across all three time points, one intervention and one control school were removed, resulting in a sample of 42 schools (17 intervention and 25 control). These 42 schools were the third-level units. The treatment condition (i.e., allocation to intervention or control group) formed the explanatory variable at this level. School type was also an explanatory variable at this level, categorized by secondary school or grammar school, as was the two-way treatment by school-type interaction. The equations for the model at each level can be described as follows:
- Level 1 model:
- Level 2 model:
- Level 3 model:
To determine whether the sample size was adequate in terms of the number of level 3 units and the number of individuals within these schools, a power analysis was conducted using the Optimal Design software package (Spybrook et al., 2011).
Multi-level modelling was conducted using the HLM 6 package (Raudenbush, Bryk, Cheong, & Congdon, 2004). A three-level hierarchical model was employed to assess the effects of time, treatment condition and school type on intention to breastfeed, feeling-based attitudes towards breastfeeding, moral-based attitudes towards breastfeeding, perceived control over breastfeeding, self-efficacy regarding breastfeeding, subjective norm and knowledge about breastfeeding. Measurement scales were keyed so that higher scores indicated more of the construct measured by the scale. The primary analyses focused on the effects of the intervention. Thus, to assess the effects of the intervention, hierarchical models included a cross-level interaction between intervention and time. In evaluation studies, such interactions can be particularly informative. Further secondary analyses examined the question of whether the intervention was differentially effective depending on school type. Treatment effects were analysed in the context of school type and the two-way interaction of treatment by school type. In these analyses, the cross-level interaction of school type by time was included, as well as the two-way interaction at level three of intervention by school type.
The power analysis revealed that the present sample size had power in excess of .90 to detect moderate effect sizes (delta = .4) in level 1 slopes between treatment and comparison schools. Thus, the sample size was deemed adequate for the analysis.
The primary analyses found significant effects for the intervention on females' Intention to Breastfeed, B = 0.208, t(1275) = 2.715, p = .007, Attitudes, B = 0.223, t(1275) = 4.655, p < .001, Moral Attitude, B = 0.231, t(1275) = 4.211, p < .001, Subjective Norm, B = 0.118, t(1275) = 2.521, p = .012, and Knowledge, B = 0.109, df(1275) = 7.843, p < .001. However, no significant intervention effects were found for Perceived control or for Self-efficacy. Changes in outcomes occurred between baseline (T1) and the 1-month (T2) follow-up and were sustained at the 6-month follow-up (T3), as shown in Table 5.
|Outcome||Intercept Mean (SE)||Slope||Treatment effect size (%)|
|Time (SE)||Treatment (SE)|
|Intention to breastfeed||4.834 (0.123)***||0.012 (0.038)||0.208 (0.077)**||14.85|
|Attitudes||4.943 (0.072)***||0.040 (0.0270)||0.223 (0.048)***||23.47|
|Moral attitude||5.656 (0.080)***||−0.068 (0.024)**||0.231 (0.046)***||19.42|
|Subjective norm||5.173 (0.010)***||0.008 (0.033)||0.118 (0.047)*||6.68|
|Knowledge||0.540 (0.014)***||0.037 (0.010)***||0.109 (0.014)***||49.06|
|Perceived control||5.573 (0.072)***||0.006 (0.043)||0.089 (0.065)||na|
|Self-efficacy||5.062 (0.097)***||0.146 (0.035)***||0.022 (0.051)||na|
|Outcome||Intercept||Variance component Level 1||Intercept 1/Intercept 2|
|Intention to breastfeed||2.527***||0.516||0.208***|
Follow-up analyses assessed the level 3 (school level) effect size for significant treatment effects. The level 3 effect is assessed by computing the percentage reduction in the Tau (Beta) parameter for the level 2 intercept. The resulting effect size indicates the proportion of between-school variance in the outcome that is explained by the addition of the level 3 treatment effect. The effect size was computed only when the treatment effect was significant. The analysis indicates that the treatment accounted for a substantial proportion of the variance between schools in Knowledge (49.06%), Attitudes (23.47%), Moral Attitudes (19.42%), Intentions to Breastfeed (14.85%) and Subjective Norms (6.68%).
The next stage of the analysis considered the possibility that the effectiveness of the intervention might differ between secondary and grammar schools. To test for such an effect, we assessed school type (grammar vs. secondary), as well as the two-way interaction between school type and treatment (intervention vs. control). No significant interactions were found between school type and treatment, suggesting that the intervention was equally effective in secondary and grammar schools in increasing Intention to Breastfeed, Attitudes, Moral Attitude and Subjective Norms. However, school type had significant effects on Knowledge, β = 0.050, t(1273) = 4.900, p < .001. Thus, a significantly greater increase over time in Knowledge levels occurred in secondary schools compared with grammar schools (regardless of whether they were intervention or control schools).
The cumulative pattern of the findings from the MLM analysis is consistent with the view that the intervention was effective. That is, it increased female students' intentions to breastfeed, expanded their knowledge about/awareness of breastfeeding and led to more favourable attitudes and perceptions of social norms. Therefore, our primary aim was met, and our associated hypothesis was confirmed for the effectiveness of the intervention on the outcome measures of intention to breastfeed, feeling and moral-based attitudes towards breastfeeding, subjective norm, and knowledge about breastfeeding. As such, this study has not only provided evidence to suggest that the school system can play a role in health education and promotion and should expose all pupils to the art of breastfeeding through health classes but has also provided support for the efficacy of the TPB as a framework for informing the design and evaluation of interventions.
Finding that the intervention led to more positive attitudes is particularly encouraging given that numerous studies have identified attitude as a significant predictor of one's intention to breastfeed (Avery, Duckett, Dodgson, Savik, & Henley, 1998; Duckett et al., 1998; Giles et al., 2007; McMillan et al., 2008; Rempel, 2004; Wambach, 1997), and attitude has been identified as an important precursor of future breastfeeding. Interestingly, both moral and feeling-based attitudes were enhanced as a result of the intervention, suggesting that the attempt to promote the potential benefits of breastfeeding (e.g., it is the right thing to do as it serves to enhance bonding) and to address the potential negative reaction towards breastfeeding (e.g., do not need to feel embarrassed as it can be done discretely) had some success.
Addressing the issue of embarrassment would seem to be particularly important given that negative attitudes to the idea of breastfeeding in public places are consistently reported in Northern Ireland (Stewart-Knox et al., 2003). However, whilst the present attempt to depict positive images of women breastfeeding in public places together with recent initiatives like the ‘Breastfeeding Welcome Here’ scheme, in which a number of businesses and public facilities have signed up to support and welcome breastfeeding families (HSC Public Health Agency, 2010), has gone some way to address the issue of public support for breastfeeding, the majority of mothers have never breastfed outside their own homes. This has prompted the suggestion that a breastfeeding mother in Northern Ireland is ‘an isolated figure in a bottle-feeding culture’ (Campbell, 2000). Given the finding that those directly exposed to breastfeeding have more positive attitudes (Friel, Kelleher, Campbell, & Nolan, 1999; Purtell, 1994) and in turn are more likely to intend to breastfeed (Tarrant & Dodgson, 2007), health promotion interventions would be wise to address the potential negative reaction towards breastfeeding in public.
The finding that the students involved in this study were more likely to perceive their significant others as supporting their decision to breastfeed following the intervention was also encouraging given social influences on breastfeeding have been shown to be particularly important. Zimmerman and Guttman (2001), for example, reported that the limited involvement of significant others was an important factor in deterring women from breastfeeding and recommended that interventions address how others can participate in the care of the baby. Particular emphasis in this study was on peer and family support given the findings of our elicitation study (Giles et al., 2007, 2010) and those reported elsewhere (HSC Public Health Agency, 2010; Stewart-Knox et al., 2003; Tarrant et al., 2010) and in particular the role of the father (Khassawnwh, Khader, Amarin, & Alkafajei, 2006; Swanson & Power, 2005; Tarrant & Kearney, 2008).
Increasingly, it is acknowledged that fathers have a key role to play, particularly in relation to the support and encouragement they can provide (Acker, 2009; Henderson et al., 2011; Rempel & Rempel, 2011). For example, in Rempel and Rempel's (2011) qualitative study with 21 fathers of breastfeeding babies in Canada, the fathers described the types of support they provided from completing household tasks and being involved in childcare to giving emotional support and encouragement and reinforcing breastfeeding behaviour in their partners as well as through gaining knowledge about breastfeeding or, as the authors suggest, being ‘breastfeeding savvy’ (p.120). As a consequence, in Northern Ireland, fathers have been targeted in recent information campaigns (Northern Ireland Assembly, 2011). However, it is also acknowledged that issues around sexuality, sexism and embarrassment on the part of the father can discourage the mother from breastfeeding (Acker, 2009; Henderson et al., 2011), hence the need to expose males as well as females to intervention programmes.
Lack of knowledge has also been identified as a potential barrier to breastfeeding. Research not only suggests that parents of breastfed infants appear more knowledgeable about the potential health benefits compared with parents of bottle-fed infants (Shaker, Scott, & Reid, 2004) but has shown that increasing the knowledge of both mothers and fathers can significantly increase the chances of exclusive breastfeeding (Susin et al., 1999). It was therefore encouraging to find that the intervention employed here successfully enhanced knowledge and that the non-grammar school children had lower baseline scores, and these scores showed a significantly greater increase over time. Given the absence of a significant interaction between treatment and school type, we can only speculate that this difference may be partly explained by the way in which breastfeeding is being discussed in the schools already. Whilst efforts were made to ensure that schools did not discuss the topic during the research phase, in reality the research team had little control over the curriculum, and it is possible that there are variations within school type. However, irrespective of this, the increase in knowledge for those in secondary schools is a positive development particularly given the significantly lower rates of breastfeeding initiation and duration often reported among women from less socially advantaged backgrounds (Bolling, 2006; Bolling et al., 2007; Hamlyn et al., 2002; Misra & James, 2000), and the calls for interventions targeting women on low incomes (Fairbank et al., 2000).
It is also posited that mothers who are better educated are more likely to breastfeed (Doyle & Kelleher, 2010; McMillan et al., 2009; Tarrant, Younger, Sheridan-Pereira, & Kearney, 2010). A UK-based study (McMillan et al., 2009), using an extended TPB questionnaire among pregnant women living in economically deprived areas, showed that age of leaving full-time education predicted intention to breastfeed beyond that accounted for by TPB constructs. The positive influence of maternal education on breastfeeding is also reported in a study by Doyle and Kelleher (2010) among mothers in Ireland and Northern Ireland and by Tarrant, Younger, Sheridan-Pereira, and Kearney (2010) with mothers in Dublin, Ireland.
Our hypothesis that the intervention would be effective in increasing students' self-efficacy regarding breastfeeding was not supported. In other words, the intervention did not increase the students sense that they would be able to breastfeed a baby, but this was perhaps unrealistic given ‘mastery experience’ is the most important factor determining self-efficacy (Bandura, 1986), and breastfeeding is obviously not a behaviour the students have been able to try. Whilst efforts were made to ensure that the images depicted in the PowerPoint presentations would be perceived as similar in age and appearance to the student group as the modelling process is more effectual when a person sees himself/herself as similar to the role model (Bandura, 1986), it is likely that a significant proportion of the sample had limited previous exposure to breastfeeding role models (Stewart-Knox et al., 2003) and will require direct exposure if self-efficacy is to be enhanced.
Although perceived control was not a direct focus of the intervention, it is perhaps important to note that the Cronbach's alpha coefficient for this scale was relatively low. Concerns in the literature regarding the conceptualization of perceived behaviour control have led some researchers to operationalize the construct in different ways (Armitage & Conner, 2001). Whilst these issues were considered at the time of item construction, we note that one of our items did employ a different response option from the other two. This could partly explain the low reliability as scales with all response options clearly labelled have been shown to yield higher test–retest reliability than those with only end points labelled (Weng, 2004). However, given the distal nature of the behaviour, the low alpha coefficient might also be explained by the fact that the items did not have much meaning for females of this age.
This study has provided evidence to support the efficacy of the TPB in the design and evaluation of a school-based intervention to promote positive attitudes to breastfeeding. A major strength of the intervention is that it is based on a sound theoretical framework. Furthermore, there was little research conducted on breastfeeding education when the present, large-scale study was initiated, and as such, it has also addressed a gap in the literature. However, it is important to acknowledge that not only were we unable to measure actual breastfeeding behaviour, but this is also likely to be a (very) distal behaviour for this group of females. Given the intention–behaviour relationship is likely to be weakened across time (Ajzen & Fishbein, 1980), it is likely that the effects of the intervention might not be sufficiently long lasting. It would therefore be particularly beneficial if the intervention could be repeated at regular intervals (e.g., yearly) and evaluated in the longer term to assess any potential change over time in the adolescents' attitudes to breastfeeding and their perceptions of social influence, control and self-efficacy in relation to breastfeeding.
The authors would like to thank the R&D Office for funding this project, the Public Health Agency for access to materials, the York Clinical Trials Unit who carried out the randomization and the schools and children who participated in the study.
- 2008). A taxonomy of behaviour change techniques used in interventions. Health Psychology, 27, 379–387. doi:10.1037/0278-6126.96.36.1999 , & (
- 2009). Breast is best but not everywhere: Ambivalent sexism nd attitudes towards private and public breastfeeding. Sex Roles, 61(7–8), 476–490. doi:10.1007/011199-009-9655 (
- 1985). From intentions to actions: A theory of planned behaviour. In J. Kuhl & J. Beckman (Eds.), Action-control: From cognition to behavior (pp. 11–39). Heidelberg, Germany: Springer. (
- 1980). Understanding attitudes and predicting social behavior. Englewood-Cliffs, NJ: Prentice-Hall. , & (
- 2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40, 471–499. doi:10.13481014466601164939 , & (
- 1998). Factors associated with very early weaning among primiparas intending to breastfeed. Maternal and Child Health Journal, 2, 167–179. doi:10.10231A:1021879227044 , , , , & (
- 1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall. (
- 2006). Infant feeding survey 2005: Early results. Prepared for the Information Centre for Health and Social Care and the UK Health Departments by BMRB Social Research. Leeds, UK: The Information Centre. (
- 2007). Infant feeding 2005. London, UK: The Information Centre. , , , & (
- 2000). A breastfeeding strategy for Northern Ireland. Ulster Medical Journal, 69(1), 1–2. (
- 2003). Theory of Planned Behaviour-based models for breastfeeding duration among Hong Kong mothers. Nursing Research, 52, 148–158. , , , & (
- 2010). Comparative analysis of breastfeeding practices in Ireland and Northern Ireland. Irish Journal of Medical Science, 179(Supp. 11), 444–445. , & (
- 1998). A theory of planned behavior-based structural model for breastfeeding. Nursing Research, 47(6), 325–336. , , , , , , & (
- 1998). Research: Societal influences upon initiation and continuation of breastfeeding. British Journal of Midwifery, 6(2), 76–80. , & (
- 2005). Promotion of breastfeeding initiation and duration: Evidence into practice briefing. National Institute for Health and Clinical Excellence. Retrieved from http://www.nice.org.uk/niceMedia/pdf/EAB_Breastfeeding_final_version.pdf , , , , , & (
- 2000). Why some women do not breastfeed: Bottle feeding and the fathers role. Midwifery, 16, 323–330. doi:10.1054/midw.2000.0222 (
- 2000). A systematic review to evaluate the effectiveness of interventions to promote the initiation of breastfeeding. Health Technology Assessment, 4(25), 1–171. , , , , , & (
- 2002). Breastfeeding education for Scottish schools. Paper No. 14, University of Glasgow/PEACH (Paediatric Epidemiology and Community Health) Unit. , & (
- 2004). Constructing questionnaires based on the theory of planned behaviour: A manual for health services researchers. Newcastle, UK: Centre for Health Services Research. , , , , , , , , & (
- 1999). Evaluation of the Nutrition Education at Primary School (NEAPS) programme. Public Health Nutrition, 2(4), 549–555. doi:10.1017/1368980099000737 , , , & (
- 2010). Attitudes to breastfeeding among adolescents. Journal of Human Nutrition and Dietetics, 23, 285–293. doi:10.1111/1365277201001048 , , , & (
- 2007). Measuring young people's attitudes to breastfeeding using the theory of planned behaviour. Journal of Public Health, 29(1), 17–26. doi:10.1093/pubmed/fd1083 , , , , , & (
- 2003). Attitudes and subjective norms of male and female adolescents toward breastfeeding. Journal of Human Lactation, 19(4), 402–410. doi:10.1177/0890334403258337 , , , & (
- 2003). Feeding preferences and attitudes to breastfeeding and its promotion among teenagers in Northern Ireland. Journal of Human Lactation, 19(1), 53–61. doi:10.1177/0890334402239735 , , & (
- 2002). Infant feeding 2000. London, UK: The Stationery Office. , , , & (
- 2002). Application of the theory of planned behaviour in behaviour change interventions: A systematic review. Psychology and Health, 17(2), 123–158. doi:10.1080/08770440290013644 , , , , , & (
- Health Promotion Agency (2005). Breastfeeding in Northern Ireland: A summary report on knowledge, attitudes and behaviour. Belfast, UK: The Health Promotion Agency for Northern Ireland.
- Health Promotion Agency (2008). Breastfeeding awareness in schools: Evaluation of a CD ROM and teaching materials. Belfast, UK: HPA.
- 2011). Men and infant feeding: Perceptions of embarrassment, sexuality and social conduct in white low income British men. Birth Issues in Perinatal Care, 38(1), 61–70. doi:10.1111/1523-536201000442 , , , & (
- 1999). Qualitative study of decisions about infant feeding among women in the East End of London. British Medical Journal, 318, 30–34. doi:10.1136/bmj.318.7175.30 , & (
- HSC Public Health Agency (2010). Breastfeeding rates in Northern Ireland increasing. Retrieved from http://www.publichealth.hscni.net/news/breastfeeding-rates-northern-ireland-increasing
- 2006). Knowledge, attitude and practice of breastfeeding in the north of Jordan: A cross-sectional study. International Breastfeeding Journal, 23, 1–17. doi:10.1186/17464358117 , , , & (
- 2000). U.S. high school age girls may be receptive to breastfeeding promotion. Journal of Human Lactation, 16(1), 36–40. doi:10.11771089033440001600107 (
- 1999). Health promoting schools and health promotion in schools. Health Technology Assessment, 3, 22. , , , & (
- 1994). The decision to breastfeed: The importance of fathers' approval. Clinical Pediatrics, 33, 214–219. doi:10.11771000992289403300405 , , & (
- 2003). Addressing inequalities in health: Breast benefits project. British Journal of Midwifery, 11(5), 281–285. , & (
- 1995). Impact of attitudes on maternal decisions regarding infant feeding. The Journal of Pediatrics, 126(4), 507–514. doi:10.1016/002234769570342 , , , & (
- 2009). Using an extended theory of planned behaviour to inform interventions aimed at increasing breastfeeding uptake in primiparas experiencing material deprivation. British Journal of Health Psychology, 14, 379–403. doi:1011348/135910708336112 , , , , , & (
- 2008). Predicting breastfeeding in women living in areas of economic hardship: Explanatory role of the theory of planned behaviour. Psychology and Health, 23, 767–788. doi:10.1080/08870440701615260 , , , , , , … (
- 2000). Factors associated with breastfeeding duration and initiation in Missouri. Journal of the American Dietetic Association, 100, 1071–1073. , & (
- NHS Information Centre (2011). Infant feeding survey 2010: Early results. The Health and Social Care Information Centre, University of York.
- 2006). Randomised controlled trial to determine effects of prenatal breastfeeding workshop on maternal breastfeeding self-efficacy and breastfeeding duration. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 35(5), 616–624. doi:10.1111/15526909200600077 , , , , & (
- Northern Ireland Assembly (2011). Official Report (Hansard): Committee for Health, Social Services and Public Safety (2011/2012). Breastfeeding rates in Northern Ireland. Retrieved from http://www.niassembly.gov.uk/Assembly-Business/Official-Report/Meetings-of-Assembly-Committee
- 1994). Teenage girls' attitudes to breastfeeding. Health Visitor, 67, 156–157. (
- 2004). On multi-level modeling of data from repeated measures designs: A tutorial. Speech Communication, 43, 103–121. doi:10.1016/j.specom200402004 , & (
- 2004). HLM 6: Hierarchical linear and nonlinear modelling. Chicago, IL: Scientific Software International. , , , & (
- 2004). Factors influencing the breastfeeding decisions of long-term breastfeeders. Journal of Human Lactation, 20, 306–318. doi:10.1177/0890334404266969 (
- 2011). The breastfeeding team: The role of involved fathers in the breastfeeding family. Journal of Human Lactation, 27, 115–121. doi:10.1177/0890334410390045 , & (
- 2002). Attitudes and subjective norms of Quebecian adolescent mothers towards breastfeeding. Canadian Journal of Public Health, 93(3), 198–202. , & (
- 2005). Breastfeeding as health promoting behaviour for Hispanic women: Literature review. Journal of Advanced Nursing, 52(2), 200–210. doi:10.1111/136526482005-03579 , & (
- 2004). Infant feeding attitudes of expectant parents: Breastfeeding and formula feeding. Journal of Advanced Nursing, 45(3), 260–268. doi:10.1046/1365-2648200302887 , , & (
- 2012). Multilevel analysis: An introduction to basic and advanced multilevel modeling. (2nd ed.). London, UK: SAGE. , & (
- 2007). College students' experiences and attitudes regarding middle and high school-based breastfeeding education. The Journal of School Nursing, 23(5), 276–282. doi:10.1177/105984050702300500601 (
- 2010). School nurses and teachers: Attitudes regarding inclusion of breastfeeding education in school curricula. Journal of School Nursing, 26(2), 137–146. doi:10.1177/1059840509350739 (
- 2011). Optimal design for longitudinal and multilevel research: Documentation for the optimal design software version 3.0. Retrieved from www.wtgrantfoundation.org or from sitemaker.umich.edu/group-based , , , , , & (
- 2003). What is the problem with breastfeeding? A qualitative analysis of infant feeding perceptions. Journal of Human Nutrition and Dietetics, 16, 265–273. doi:10.1046/1365277200300446 , , & (
- 2002). Consolidation and updating the evidence base for the promotion of breastfeeding. Retrieved from www.wales.gov.uk. Cardiff, UK: National Assembly of Wales. (
- 1999). Does parental breastfeeding knowledge increase breastfeeding rates? Birth, 26(3), 149–156. doi:10.1046/15235361999.00149 , , , , , & (
- 2002). Using social cognition models to develop health behaviour interventions: Problems and assumptions. In D. Rutter & L. Quine (Eds.), Changing health behaviour: Intervention and research with social cognition models (pp. 193–208). Buckingham, UK: Open University Press. (
- 2005). Initiation and continuation of breastfeeding: Theory of planned behaviour. Journal of Advanced Nursing, 50(3), 272–282. doi:10.1111/1365-2648.200503390 , & (
- 2007). Using multivariate statistics. (5th ed.). Boston, MA: Allyn & Bacon. , & (
- 2007). Knowledge, attitudes, exposure and future intentions of Hong Kong university students toward infant feeding. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 36(3), 243–254. doi:10.1111/15526909200700144 , & (
- 2008). Public health Nutrition, Breast feeding practice in Ireland in the Symposium on “The challenge of translating nutrition research into public health nutrition”. Proceedings of the Nutrition Society, 67, 371–380. doi:10.1017/50029665108008665 , & (
- 2010). Factors associated with duration of breastfeeding in Ireland: Potential areas for improvement. Journal of Human Lactation, 27(3), 262–271. doi:1011177/0890334411413097 , , , & (
- 2010). The prevalence and determinants of breastfeeding -initiation and duration n a sample of women in Ireland. Public Health Nutrition, 13, 760–770. doi:10.1017/51368980009991522 , , , , & (
- 2007). The use of minimization to form comparison groups in educational research. Educational Studies, 33(3), 333–337. doi:10.1050/03055690701423267 , & (
- 1997). Breastfeeding intention and outcome: A test of the theory of planned behavior. Research in Nursing and Health, 20(1), 51–59. (
- 2000). Breastfeeding and adolescents. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 29(3), 282–294. doi:10.1111/1552-6909.2000.02050 , & (
- 2004). Impact of the number of response categories and anchor labels on coefficient alpha and test-retest reliability. Educational and Psychological Measurement, 64, 956–972. doi:10.117710013164404268674 (
- WHO (2010) Retrieved from http://www.who.int/nutrition/topics/infantfeeding/en/index.html
- World Health Organization/UNICEF (2003). Global strategy for infant and young child feeding. Geneva, Switzerland: World Health Organization.
- 2001). “Breast is best”: Knowledge among low-income mothers is not enough. Journal of Human Lactation, 17(1), 14–19. doi:10.1177/089033440101700104 , & (