A health equity critique of social marketing: Where interventions have impact but
Rebecca Langforda,*, Catherine Panter-Brickb
aDepartment of Anthropology, Durham University, Dawson Building, South Road, Durham DH1 3LE, UK
bDepartment of Anthropology & Jackson Institute, Yale University, 10 Sachem Street, New Haven, CT 06511, USA
a r t i c l e i n f o
Available online 11 February 2013
a b s t r a c t
Health interventions increasingly rely on formative qualitative research and social marketing techniques
to effect behavioural change. Few studies, however, incorporate qualitative research into the process of
program evaluation to understand both impact and reach: namely, to what extent behaviour change
interventions work, for whom, in what contexts, and why. We reflect on the success of a community-
based hygiene intervention conducted in the slums of Kathmandu, Nepal, evaluating both maternal
behaviour and infant health. We recruited all available mothereinfant pairs (n ¼ 88), and allocated them
to control and intervention groups. Formative qualitative research on hand-washing practices included
structured observations of 75 mothers, 3 focus groups, and 26 in-depth interviews. Our intervention was
led by Community Motivators, intensively promoting hand-washing-with-soap at key junctures of food
and faeces contamination. The 6-month evaluation period included hand-washing and morbidity rates,
participant observation, systematic records of fortnightly community meetings, and follow-up interviews
with 12 mothers. While quantitative measures demonstrated improvement in hand-washing rates and
a 40% reduction in child diarrhoea, the qualitative data highlighted important equity issues in reaching
the ultra-poor. We argue that a social marketing approach is inherently limited: focussing on individual
agency, rather than structural conditions constraining behaviour, can unwittingly exacerbate health
inequity. This contributes to a prevention paradox whereby those with the greatest need of a health
intervention are least likely to benefit, finding hand-washing in the slums to be irrelevant or futile. Thus
social marketing is best deployed within a range of interventions that address the structural as well as
the behavioural and cognitive drivers of behaviour change. We conclude that critiques of social mar-
keting have not paid sufficient attention to issues of health equity, and demonstrate how this can be
addressed with qualitative data, embedded in both the formative and evaluative phases of a health
? 2013 Elsevier Ltd. All rights reserved.
It is well established that many of the most intractable public
health issues in poorer parts of the world could be prevented or
ameliorated simply by changing people’s behaviour: for example,
hand-washing to prevent diarrhoea, bednets to protect against
malaria, or condoms to stop the spread of sexually transmitted
infections (Briscoe & Aboud, 2012). As highlighted in a recent
special issue on behaviour change in developing countries (Aboud
& Singla, 2012), the key to achieving improved health outcomes
lies in designing interventions grounded in theoretical models of
behaviour change, good quality evidence, and an in-depth under-
standing of the target audience.
However, changing people’s behaviour is notoriously difficult to
do and there remain remarkably few examples of truly successful,
Briceno-Leon, & Johnson, 2001). Many interventions are poorly
theorised, often based on the premise that educating people about
potential threats to health will be sufficient to motivate a change in
risk practices. While this approach may result in changes in
knowledge and attitude, there is little evidence to suggest it
translates into actual behavioural change (Loevinsohn, 1990). As
more sophisticated behavioural models have developed, it is clear
that new approaches to changing behaviour are needed, for
example to extend beyond rational, cognitive drivers of behaviour.
* Corresponding author. Current address: School of Social and Community
Medicine, University of Bristol, Canynge Hall, 39 Whatley Road, Bristol BS8 2PS, UK.
Tel.: þ44 0 117 928 7353.
E-mail address: Beki.email@example.com (R. Langford).
Contents lists available at SciVerse ScienceDirect
Social Science & Medicine
journal homepage: www.elsevier.com/locate/socscimed
0277-9536/$ e see front matter ? 2013 Elsevier Ltd. All rights reserved.
Social Science & Medicine 83 (2013) 133e141
It is here that social marketing potentially has much to offer. A
dominant paradigm in health promotion, social marketing “pro-
motes the voluntary behaviour of target audiences by offering
benefits theywant, reducing barriers theyare concerned about, and
using persuasion to motivate their participation in program activ-
ity” (Kolter & Roberto, 1989: 24). It moves away from traditional
health education models that simply tell people what to do, to-
wards an approach that seeks to ‘sell’ the behaviour in question by
convincing the target audience that it provides a solution to
a problem they believe is important, and/or offers them a benefit
they value (Grier & Bryant, 2005: 323). The first step in this process
is uncovering and understanding people’s perspectives, prefer-
ences, and aspirations. As Curtis (2001: 76) asserted for hygiene
prevention, ‘health is not the only motivation for healthy behav-
iour; other goals may be far more important.’ Identifying these
other goals is a crucial part of discovering how best to promote
novel social norms and behaviours.
Social marketing has been embraced by the public health
community and widely utilised in low-income countries, partic-
ularly in relation to infectious disease control (e.g., sexual health,
malaria, and hygiene). However, there remains a paucity of evi-
dence regarding whether this approach is more effective than other
types of behavioural intervention. Comprehensive and convincing
evaluations of social marketing interventions undertaken in low-
income countries are rare. Those that do exist tend to be nar-
rowly focussed and exclusively quantitative in nature, relying
heavily on sales data (e.g., condoms purchased) or self-reported
changes in knowledge or behaviour (Price, 2001).
Intervention success is judged by many criteria: these include
impact, reach, sustainability, cost effectiveness, acceptability, and
equity. The process of evaluation is thus multi-dimensional,
requiring a mixed-methods, ‘qual-quant-itative’ approach to fully
capture all relevant dimensions of evidence. Careful and detailed
qualitative research is at the heart of a social marketing approach:
to identify what needs to be communicated, inwhat way, to whom,
and through which channels in order to achieve a change in key
behaviours (Biran et al., 2005: 213). But to-date, qualitative
research is recognised as crucial only to the formative stages of
intervention development; it is valued as a critical step in under-
standing how to shape and deliver a health intervention, but its
importance beyond such formative stages is often overlooked.
In-depth contextual data in the evaluative stages are often con-
spicuously absent. When so many interventions are focussed on
changing norms and behaviours, such limited qualitative analysis
are embedded within a range of social, political and economic con-
texts” (Moffat, White, Mackintosh, & Howell, 2006: 28). Despite calls
for the inclusion of qualitative data (Donnovan et al., 2002), few
evaluations provide details on the implementation process or wider
context that might be critical to its impact (Roen, Arai, Roberts, &
Popay, 2006). Evaluations remain largely focussed on ‘hard out-
comes’, with a paucity of qualitative work that seeks to capture local
responses to interventionprogrammes. Nor has there been adequate
exploration and critical reflection of the unintended consequences
and potential harms that may arise from interventions (Kleinman,
2010), especially those that specifically aim to shift social norms.
In the field of behavioural health, researchers have thus accu-
mulated expertise with respect to measuring attitudinal or
behavioural change following a specific health intervention; how-
ever, theyare often left with a ‘black box’ as to how exactly this was
(or was not) achieved. Some key questions e what works, for
whom, in what circumstances, in what respects, and how (Pawson
& Tilley, 2006) e often remain poorly answered.
In this paper, wecriticallyreflecton the success of a community-
based hygiene intervention and the insights gained through long-
term qualitative research embedded in programme evaluation.
Our intervention targeted maternal hand-washing behaviours in
the slums of Kathmandu, Nepal. We capitalized on the ‘lessons’
learnt of previous community-based hand-washing interventions
in developing countries (especially Curtis et al., 1997, 2001). Thus
we focussed attention on the psychosocial determinants of
behavioural change, informed by the Theory of Planned Behaviour
(Fishbein & Ajzen, 1975). Additionally, we capitalized on insights
derived from careful studies of the drivers of hygiene behaviour,
targeting what Aunger et al. (2010: 384) called ‘motivated behav-
iours’ e behaviours that “occur in response to a need, or perceived
discrepancy between an aspect of a person’s current state and an
ideal state” to create a demand for hygiene. In this way, we incor-
porated strategies from social marketing to ‘sell’ hand-washing
behaviours to mothers caring for young children. The impact of
this intervention, regarding ‘hard outcomes’ (namely, maternal
hand-washing practices, child morbidity, and growth) is reported
elsewhere (Langford, Lunn, & Panter-Brick, 2011). We focus this
paper on qualitative data collected in the formative and evaluation
phases of the intervention. We present these data to evaluate both
the power of a social marketing approach and its limitations.
Ethical approval was formally obtained from both the Nepal
Health Research Council and Durham University’s Research Ethics
Committee. Verbal informed consent was gained from all partici-
pants and slum community leaders.
The study was conducted in the eight largest slums of Kath-
mandu, randomised to either intervention or control groups, on the
basis of the most recent demographic data available (Shrestha &
Shrestha, 2005). We purposively chose the largest slum settle-
ments in order to maximise sample size, leaving aside small set-
tlements where intensive observation and repeated measures
would be too logistically difficult over the year of study. Our target
population consisted of mothers caring for infants 3e12 months
old, the first year of life being the age range most relevant to
monitor changes in child growth and morbidity outcomes. Our
sample was small, but comprised of all available mother/infant
pairs (n ¼ 88) living in the slums. Eligible participants were iden-
tified from house-to-house surveys, intensively recruited, and
invited to an information meeting; all agreed to participate.
Mothers were also offered a small gift (200 rupees, approximately
£1.50) for growth measurements and time compensation. Having
recruited a total sample of mothers with infants, we divided our
work between areas assigned to intervention (n ¼ 45) and control
(n ¼ 43). We worked intensively with this sample over one year, as
the primary aim of our study was to produce good-quality repeated
measures on infant health outcomes, pre- and post-intervention.
The study was carried out during 2005. We conducted formative
research for four months, and implemented the intervention for six
months with continuous evaluation. The lead author (RL) con-
ducted participant observation throughout this period, and led the
intervention with the help of two research teams: one responsible
for the intervention’s design and implementation, the other
responsible for survey evaluation. The first team included two
Nepali research assistants, to assist with focus groups and in-
terviews, and five well-known and respected women from the
slums, recruited to be Community Motivators (CMs) taking prime
responsibility for program implementation. The second team
comprised ten Nepali field workers, trained to conduct structured
observations and administer weekly child morbidity surveys; to
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
minimize bias, they were never involved in any elements of the
The aim of the formative research was to assess mothers’ hy-
giene practices and identify their motivations and barriers to
changing hand-washing behaviour, in order to inform the devel-
opment of the intervention message. Following a research protocol
for hygiene interventions established by Curtis et al. (1997), we
focussed on hand-washing-with-soap at five key junctures: after
going to the toilet, after cleaning-up child faeces, and before
cooking, eating food, or feeding the baby. We undertook (i) struc-
tured observations and (ii) self-reports of hand-washing behaviour,
as well as (iii) focus group discussions, and (iv) in-depth interviews,
as described below.
Structured observations were implemented, after piloting, with
a random selection of households (n ¼ 75): our field workers
recorded mothers’ hand-washing behaviour, noting whether or not
soap was used, for 3-h periods in the morning. Mothers were
informed that our observations focussed on their daily work, rather
than hygiene. Post-intervention, we chose not to repeat direct ob-
servations, due to likely ‘reactivity’: mothers would be aware that
observations focussed on hand-washing behaviour. For self-
reports, surveys were undertaken with all mothers (n ¼ 88), both
pre- and post-intervention.
Focus group discussions focussed on local perceptions of
cleanliness and hygiene. We conducted three groups, randomly
selecting participants (6e8 mothers/group) from intervention
communities. The groups, lasting 2 h, were moderated in Nepali by
a research assistant specifically trained for this task, with com-
prehensive notes taken by a second Nepali assistant. The moder-
ator, note-taker, and lead author met after each focus group to
We focussed semi-structured interviews in intervention com-
munities, selecting a random sample of respondents (n ¼ 26, from
a total 45) having developed a discussion guide to prompt for emic
perspectives on health and hygiene. The lead author conducted
interviews in the privacy of mothers’ own homes, in Nepali, with
assistance from a Nepali research assistant where necessary. In-
terviews lasted approximately 1 h and were not recorded; notes
were taken throughout and written up into comprehensive field
notes immediately after.
The intervention activities were developed by the research
team, and implemented in intervention communities (mothers in
the control group continued habitual practices). They aimed to (i)
promote a positive attitude towards hand-washing, (ii) establish
hand-washing as a social norm, and (iii) remove barriers that might
hinder this practice. Thus mothers were encouraged to wash hands
with soap at the five key junctures mentioned above. They were
supplied with a free bar of soap every two weeks, to remove eco-
nomic constraints on purchasing soap and initiate behavioural
change. CMs held launch meetings to introduce mothers to the
programme, and conducted regular home visits (daily, then taper-
ing to once a week) to each and every mother throughout the
intervention period. Fortnightly mothers’ group meetings were
facilitated by CMs, providing an opportunity to discuss the inter-
vention messages in an informal and sociable setting. CM meetings
were also held every two weeks with the lead author to share ideas
or concerns and assess the intervention’s progress. Posters were
designed and prominently displayed throughout the communities,
while drama performances, a ‘hand-washing song,’ and dancing
were composed and facilitated at group meetings and other com-
We assessed the program’s impact, both quantitatively and
qualitatively. Quantitative indicators encompassed (i) self-reported
hand-washing rates, (ii) child morbidity reports, assessed week-by-
week using a simple symptom checklist for main illnesses, and (iii)
measures of infant growth (reported elsewhere, Langford et al.,
Qualitatively, we assessed attitudinal and behavioural change, as
well as constraints on hygiene behaviour, with (i) participant obser-
vation, and (ii) in depth interviews. The lead author visited slums on
a daily basis, taking up opportunities for informal observations and
conversations, attended fortnightly mothers’ group meeting, and
convened regular meetings with CMs. Post-intervention, she con-
ducted in-depth interviews with participants from intervention
communities (n ¼ 12, from total 45), purposively chosen to reflect
relative poverty and engagement in the programme.
Formative data were analysed collaboratively by the lead author
with Nepali research assistants, to inform the design of the inter-
vention. In-depth qualitative analysis built upon this first phase.
This involved content analysis of all field notes, interviews, and
focus group discussions, in English and Nepali, coded by hand to
identify salient thematic categories, using an iterative process of
comparison between all sources of ethnographic data. All names
have been changed.
Conditions in the slums
The socio-demographic characteristics of our sample are shown
in Table 1. We found no differences between intervention and
control groups with respect to demographic variables or composite
socio-economic score; participants in intervention communities
were ‘poorer’ on two counts, namely, living in a dwelling with just
one room, and using the cheapest form of cooking fuel.
Half-the sample (55%) owned their own house (but not deeds to
the land on which it was built); the rest lived in rented accom-
modation. Most houses were simple brick constructions roofed
with corrugated iron sheets, while ultra-poor households lived in
shelters of woven bamboo or plastic sheeting. Overhalf (57%) of the
sample lived in just one room, which served as kitchen, bedroom
and general living area for the entire family.
Access to drinking water was limited in all areas: three settle-
ments had public taps that provided water for just a few hours on
alternate days; the remaining settlements relied primarily onwater
frompublic or private tube wells and deep wells. Water forhygienic
purposes was always available from these wells, even during the
dry season. Soap was available in every household. The majority of
families (82%) did not have access to a private toilet, but instead
shared sanitary facilities with several families or used public toilets.
The number of people sharing access and the cleanliness and state
of repair of the toilets varied dramatically within the settlements.
Baseline hand-washing practices
Structured observation showed that hand-washing-with-soap
was not routinely practised. Only a fifth (21%) of mothers washed
hands with soap after defaecation, and only 14% did so when
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
cleaning the baby after defecation. Just two were seen to wash
hands with soap before handling food, and none did so before
feeding the child (data not shown). Mothers’ self-reports, on the
other hand, indicated that hand-washing was a normative (if not
actual) practice, at least after defecation. At baseline, 96% mothers
reported hand-washing-with-soap after defecation, 82% reported
using soap after cleaning the baby’s bottom, with just a few using
soap before cooking food (12%), feeding the baby (26%), or eating
a meal (14%).
Drivers motivating hand-washing behaviour
Four main reasons were identified, from thematic analyses, for
hand-washing-with-soap: health vigilance, disgust, personal ben-
efits, and social norms (Fig. 1).
Health vigilance was key. Mothers felt responsible for preparing
food, looking after children, and thus protecting their family from
ill-health. Cleanliness was important to prevent sickness e hand-
washing was part of a suite of activities aimed to maintain hy-
giene and health in the home.
You must keep the house clean, give clean food, wash your hands
after going to the toilet, keep surroundings clean. You must do this
to stay healthy and avoid disease. (Interview data).
Hand-washing was also strongly motivated by notions of dis-
gust, especially around defaecation, when failure to wash hands
with soap would make mothers feel ‘disgusting’, ‘wrong’, ‘dirty’,
I have to wash my hands after going to the toilet. Even if they are
clean and don’t smell, I still want to wash them. I know in my heart
that they’re still dirty and I feel wrong. (Interview data).
By contrast, using soap to clean hands made them feel ‘nice’,
‘clean’, ‘fresh’, ‘light’, ‘at ease’. Only soap could offer such a ‘really
clean’ feeling. The personal benefits of using soap focused on
having soft, nice-smelling hands.
[Soap] makes your hands smell nice and it makes me feel I look
good, nice. I feel light afterwards. (Interview data).
Hygienic behaviour was also driven by strong social expecta-
tions. It was important for the women to be thought of as ‘good’
mothers who looked after their families well, and achieving this
required high standards of hygiene. A strong motivation for
Fig. 1. Determinants of behaviour change: how proximate influences were targeted by the intervention (formative phase).
Household demographic and socio-economic characteristics of control and inter-
Control (n ¼ 43) Intervention (n ¼ 45)
Maternal education %
Paternal education %
Rooms in house %
Fuel type %
Income per month (Rs)
aBased on composite measures of parental levels of education, housing tenure,
access to facilities household income and ownership of valuable material
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
hygienic behaviour therefore centred on what people might say
about them in private and a desire to give the right impression.
I want them to think, ‘Oh, she’s a good mother, she keeps herself
and her house clean.’ (Interview data).
Barriers to hand-washing behaviour change
Women identified three main barriers to hand-washing: it was
unnecessary, impractical, or ineffective (Fig. 1).
Feelings of disgust drove hand-washing behaviour: this carried
weight after defaecation, but hardly applied to cooking, eating or
feeding a child, when water alone was deemed sufficient to cleanse
hands. There was clearly no social expectation to use soap in the
latter instances, unless hands were visibly soiled.
You only need to wash with water before cooking. Your hands
aren’t dirty then so no soap is necessary. (Focus group data).
Hand-washing-with-soap required time and effort: one had to
go outside rather than quickly rinse hands in a bowl at home; it
took longer to clean hands with soap; and it required greater
amounts of water to rinse away all the suds, water which had to be
fetched from a communal pump. Cost and availability of soap were
also a problem: while soap was present in every sample household,
soap for hand-washing was still mentioned as a financial burden.
If you spend ten rupees on soap, that’s ten rupees you could have
spent on food. (Interview data).
Good hygiene could prevent sickness, but could not guarantee
good health. Despite the efforts of the mothers to protect them,
children contracted illnesses for avariety of reasons, many of which
mothers had no control over. Diarrhoea, colds, fevers and other
diseases were commonly attributed to changes in the weather, the
cold (Nepali: chiso), evil spirits, or simply ‘fate’, which better hy-
giene was unable to prevent. Equally, the poor living conditions of
the slums were seen to thwart women’s efforts to keep their chil-
dren clean and healthy.
I am so careful about my children. I pay great attention to keeping
them clean, washing their hands, their faces, giving them good
food, clean clothes e and yet they still get sick. (Interview data).
Development of the intervention messages
for hand-washing practices. However, mothers also expressed
ambivalence regarding the link between hygiene and health, and
compelling; it could conflict with their own ‘lived experience’ in the
slums. A compelling intervention message therefore needed to tap
into other goals and motivations for hygienic behaviour.
Ways inwhich the intervention drew on insights from formative
research, regarding the proximate drivers of hand-washing
behaviour, are modelled in Fig. 1. We promoted hand-washing-
with-soap as a means of protecting one’s family from sickness,
but also tapped into other salient drivers of hygiene behaviour.
Thus, the intervention targeted social norms around hand-washing
by emphasising the idea that this is what ‘responsible’ mothers do.
In addition, we attempted to create a demand for hygiene by
stressing how hand-washing made one feel ‘clean,’ ‘light,’ with
‘soft’ and ‘scented’ hands, to encourage mothers to feel good about
themselves. Finally soap was provided free-of-charge to all partic-
ipants to overcome financial barriers to participation.
Impact of the intervention
By the end of the intervention period, mothers in the inter-
vention areas were more likely than counterparts to report hand-
washing-with-soap at four out of five key junctures (Table 2). In
terms of morbidity, babies in intervention communities experi-
enced fewer days of diarrhoea (p ¼ 0.023, Table 2).
Ethnographic evaluation, however, revealed a more nuanced
picture, one gained from participant observation, accounts taken at
key questions e what works, for whom, in what contexts and why?
In mothers’ meetings and in-depth interviews, mothers talked
about establishing new hand-washing behaviours. Most felt they
were now washing their hands more than before. As one woman
It’s easy now. It’s become a habit. You see the soap and it reminds
you. (Mothers’ group meeting).
The mothers had also encouraged their husbands, mothers-in-
law and older children to wash hands more often, resulting in
changes in their whole family:
I taught my children about it and now my eldest is always saying
“Shouldn’t we wash our hands now, Mummy?” (Mothers’ group
When discussing the benefits of hand-washing, mothers men-
tioned health benefits for their children, but also focussed on per-
sonal benefits thus mirroring the messages promoted through the
[Hand-washing] kills the germs on your hands. If you don’t do it,
yourchild will become sick... I think [myson] is less sick now, he has
less diarrhoea. (Mothers’ group meeting).
Those soaps [laundry/dish soap] make your hands dry and sore.
This soap is nice though. It doesn’t dry your hands, it makes them
soft again. (Interview data).
The regular home visits to each mother by the CMs were of
central importance in establishing new hand-washing habits.
Women explained that it was helpful to have someone remind
them, during the first month, when they were most likely to forget.
But even more significant was making hand-washing acts more
‘visible’ in the community. CMs identified one of the most suc-
cessful elements of the intervention to be harnessing social norms
regarding the need ‘to be seen to be clean.’ Being aware that other
people might be watching what they were doing was a powerful
driver to behaviour change. As one CM explained,
[The mothers] have touse the public toilets down by the streamand
that’s right next to the rower pump where women wash their
Impact of the intervention on hand-washing and child morbidity.
Control (n ¼ 43) Intervention (n ¼ 45)
Maternal hand-washing (% post-intervention)
Cleaning baby’s bottom
Feeding the baby
Child morbidity (median days with symptoms)
aChi-Square e hand-washing; ManneWhitney U e morbidity.
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
clothes. They come out and they know people are watching so they
make sure to come over and ask for some soap so they can wash
their hands. (CM meeting).
While another commented,
Everyone knows each others’ business here. They all want to keep
up with each other. So if so-and-so’s doing it, they want to do it too.
In sum, homevisits, group meetingsand the everyday visibilityof
hand-washing behaviour, coupled with social norms regarding hy-
Our intervention was less successful at stimulating hand-
washing before contact with food, in contrast to the key juncture
of defecation. Mothers often admitted they skipped hand-washing
When you’re at home it’s easy. But if you’re out at a friend’s house,
they often don’t have [soap] and so I don’t wash my hands then.
(Mothers’ group meeting).
This was reiterated by the CMs:
They never think about hand-washing [before contact with food]. I
remind them about it and they say ‘Yes, yes’ but you know they
don’t really think it’s important. (CM meeting).
Thus in the evaluation phase, as with the formative phase, there
simply was not the same hygienic or moral imperative to wash
hands before contact with food. Moreover, because these food-
related behaviours were less visible e hidden inside the home e
it was not possible to harness the same social pressure to shift
these behavioural norms.
It is difficult to appraise how many women disengaged with the
intervention, in terms of not changing their hand-washing practices.
CMs readily identified eight women by name (one fifth of the
intervention group); they may have known others, but not specifi-
cally brought them to attention in meetings with the lead author.
These eight women were among the poorest women enrolled in the
study. Reasons for their lack of engagement were identified through
in-depth interviews and open-ended discussion. They included
poverty, powerlessness, and competing priorities, as conceptually
represented in Fig. 2 and illustrated with case examples below.
Extreme poverty resulted in both practical and psychological
constraints on behavioural change for these women. Unlike the
majority of mothers in the study, these women often had to seek
employment outside of the home, in order tomeet baresubsistence
needs. This presented a number of practical constraints on their
ability to change hand-washing practices. A case in point was
Bishnu-Maya Gurung, mother to nine month-old Pratik. She
intermittently took up labouring work to supplement her hus-
band’s income. Having no family in the area, she had to take Pratik
with her, either carrying him on her back or leaving him some-
where nearby while she worked. She stated the obvious: when
engaged in labouring work, hand-washing simply was not possible;
there was no water, no toilet, no soap, and no way of getting access
to these resources. She knew the importance of hygiene, but since
no soap or water was available to her for much of the day, she
simply could not wash her hands at the appropriate key junctures.
Powerlessness to govern family hygiene behaviours is best
exemplified by the case of Sarmila Pariyar. She was employed to
wash dishes and clothes for richer families nearby. Her husband
was an alcoholic, out of work for several years, and thus her family
(of six) was entirely dependent on her wage of just 2000 Rs (£15)
a month. Her job took her away from the home for several hours,
both morning and afternoon. During this period, her four children
were ostensibly left in the care of their father, although Sarmila
admitted he spent much of the day drunk or asleep; mostly the
children had to take care of themselves. Her youngest son, Sujal,
was often left in the care of his five-year-old sister who was sup-
posed to feed him when he grew hungry. Thus, for a large pro-
portion of the day, managing the hygiene behaviours of her family
was largely outside of Sarmila’s control.
Of course, despite their work, both Bishnu-Maya and Sarmila
were at home for some of the time and could therefore potentially
have stepped up hand-washing practices at these times. However,
practical considerations were not the only constraints their
behaviour. Competing priorities were also at work. Sarmila’s CM
identified another important reason for her lack of engagement:
hand-washing simply was not the most important priority at that
time. In her words,
She’s just not interested... It’s very difficult for her. Her husband
does nothing, he doesn’t work, he just drinks all day and she has
Fig. 2. Determinants of health equity: how structural issues led to disengagement with the intervention (evaluation phase).
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
no-one to help her with all those children. She has other things to
worry about. (CM meeting).
For Sarmila, the potential threat of her child becoming sick as
a result of not hand-washing was far less pressing than the need to
earn enough money to survive the next week, especially as she was
rarely at home to be able to instil this new hygiene behaviour.
The CMs, for their part, were acutely aware of the difficulty of
promoting behavioural change in the face of such challenging life
circumstances. They frequently referred to certain mothers as being
‘difficult’ to engage in the programme, as typified by Bhumika
Lama. Her husband did occasionally work but spent the money he
earned on alcohol. A few weeks after giving birth, Bhumika
therefore returned to her insecure, low-paid work as a day-
labourer. As a result of her work, Bhumika often did not attend
the mothers’ group meetings, and was rarely at home when the CM
went to visit. When she was at home, the CM reported her as being
uninterested or even hostile to the hand-washing message. It was
awell-knownfact that Bhumika’s husband was an alcoholic, thathe
often beat her, and that they were struggling to survive on her low
wage.Her CM admitted that promoting hand-washing in the face of
all the other problems Bhumika faced felt irrelevant and futile.
The particular difficulties facing the ultra-poor, especially those
who had to leave the home for work, were acknowledged by other
mothers in the intervention.
If a woman is a widow or her husband is away or not working and
she has lots of children, she doesn’t have time to look after them all
properly. It’s very hard for them. (Interview data.)
However, often these women and their actions were discussed
in morally loaded terms.
Some mothers leave the house early for work and don’t take proper
care of their children. If they wanted to they could get up early in
the morning and clean the house and see to all the things, but they
are just lazy and don’t want to do this. (Interview data.)
During informal conversations, mothers would ‘name and
shame’ their neighbours who they felt were not living up to the
expected standards of hygiene.
[My neighbour] doesn’t wash her hands after she goes to the toilet
and she lets her children go to the toilet on the ground out there...
She just says ‘Let it be, I don’t care’. ... I think she’s happy being dirty.
Cleanliness thus carried strong moral connotations: rather than
admit to not hand-washing, most women talked more generally
about the difficulties of keeping clean in a highly contaminated
slum environment. As one mother stated,
We have to live next to this dirty, smelly stream and there’s nothing
we can do. You can’t keep yourself or your children clean and
healthy if you have to live in a place like this. (Interview data.)
Our study is unique among community-based hygiene in-
terventions in that ethnographic data were embedded throughout
the research process, alongside quantitative measures of the in-
tervention’s impact. We relied on qualitative research in two ways:
in designing a compelling intervention message, and in providing
a better understanding of impact and reach.
Formative qualitative research was crucial in the design of the
intervention message. While baseline observations confirmed low
rates of hand-washing practice, formative data identified that
hand-washing and hygiene were important to slum-dwelling
women, involving strong normative and moral expectations. Bor-
rowing ideas from social marketing, we sought to understand
hand-washing behaviour from the perspective of our target audi-
ence to design an intervention that targeted the most salient
drivers of behaviour change. To go beyond mere health concerns,
we sought to create a demand for hygiene, targeting what Aunger
et al. (2010) call ‘motivated behaviours’ by emphasising the posi-
tive personal benefits of hand-washing to mothers and tapping into
social expectations around maternal responsibility for hygiene.
This strategy worked well; we observed an increase in reported
maternal hand-washing rates and a 40% decrease in diarrhoeal
morbidity in intervention relative to control groups. This result falls
in line with those reported from other community-based hand-
washing interventions (Curtis & Cairncross, 2003). The intervention
was particularly effective at increasing hand-washing after contact
with faeces; by making hand-washing more ‘visible’ within the
community, we were able to harness the power of normative hy-
giene expectations and a cultural imperative ‘to be seentobe clean’.
The intervention was less effective at increasing hand-washing
before contact with food, where moral and social imperatives to
wash hands were weaker: emphasising positive personal benefits
alone was not effective in shifting behaviour. We thus demon-
strated that by targeting and harnessing the power of social norms
it is possible to change hygiene behaviours in challenging envi-
ronments such as urban slums, where hygiene promotion is both
important and difficult to achieve.
Unlike many community-based hand-washing interventions
which limit qualitative data to formative stages of research, this
study embedded ethnographic data into programme evaluation.
These data revealed a nuanced appraisal of programme effective-
ness. While the intervention engaged the majority of women,
changing hand-washing behaviour was neither easily achievable
nor a high priority for the ultra-poor, given other pressing daily
concerns. This highlights an important limitation of a social mar-
keting approach to behaviour change. Where barriers to hand-
washing are largely attitudinal e for example, where hand-
washing is deemed unnecessary and/or ineffective e a marketing
approach that taps into other salient motivators can be highly
effective. Where women face real structural and practical barriers
to change, this type of approach is unlikely to succeed.
Our intervention was informed by ideas from social marketing,
targeting motivated behaviours and social norms to create a ‘de-
mand’ for hygiene. It differed from a ‘pure’ social marketing
approach in two ways. First, because our primary aim was to assess
the impact of hand-washing on child growth and morbidity, we
provided mothers with soap free-of-charge. Second, given the small
scale of the intervention, we did not carry out detailed audience
segmentation. Had we done so, we might have identified at the
outset the poorest women for whom the intervention would not be
compelling. Nonetheless, it is difficult to see how any way of mar-
keting thismessagetothe ultra-poorcould havebeeneffectivewhen
social, economic, and physical circumstances remained unchanged.
This limitation, pertaining to the wider ‘structural’ constraints
on people’s behaviour,has
community-based hygiene studies. Aunger et al. (2010), for exam-
ple, reported that economic constraints were significant barriers to
hand-washing in Kenya. In the poorest communities, people raised
concerns over ‘wasting’ soap and water, concluding that costs of
using these valuable resources often outweighed any potential
benefits. Similarly, Jenkins and Scott (2007) reported on the pos-
sibilities of using a social marketing approach to improve house-
hold sanitation facilities. They identified important structural
issues e namely poverty and a lack of finance and credit options e
that placed significant constraints on householders’ decision-
making to adopt sanitation facilities. Importantly, they argued
been acknowledged inother
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
that “marketing is unlikely to be able to fullyaddress either of these
[structural issues] and laws, public policies, and other mechanisms
are required” (2007: 2439).
As Wallack (2002: 30) noted, “everything flows from how the
problem is defined.” Therein lies an inherent limitation in social
marketing’s theoretical perspective: that the ‘problem’ to be solved
is people’s behaviour. The social marketing model focusses on
proximate, behavioural causes of ill-health rather than the wider
structural contexts in which behaviours are (or are not) produced.
This focus is consistent with the attention given to the cognitive
drivers of behaviour change and a search for the best ‘leverage
points’ on individual-level behaviour. Whilst this approach can be
highly effective in identifying and framing a compelling message, it
can only benefit those who are actually in a position to implement
such change. This type of approach is necessarily based on an im-
plicit assumption of high levels of individual agency (Blankenship
et al., 2006), with expectations of innovation diffusion having
a lasting impact on public health. Yet our data identified that some
of the most important barriers to hygiene practices are largely the
result of structural issues wholly beyond the scope of a marketing
intervention predicated on individual agency. The result was an
inadvertent push towards health inequity, a prevention paradox
where those who stand to gain the most from changing health-
related behaviours were the least able to achieve this. Those who
could change their behaviour did so; those whose behaviour was
largely constrained by wider structural factors, did not.
This negative impact on health equity is, of course, not unique to
social marketing alone. Many well-intentioned interventions have
appeared successful in improving in health behaviours at the aggre-
groups (Macintyre, 2003). It is increasingly recognised that ‘down-
to increaseinequitythan‘upstream’ interventions tackling structural
issues (Lorenc, Petticrew, Welch, & Tugwell, 2012). With its explicit
such criticisms. Yet, little attention has been paid to the impact of
social marketing on health equity (Knerr, 2011). The few studies that
have explored this issue have shown mixed results. Agha, Van
Rossem, Stallworthy, and Kusanthan (2007), for example, found
asocialmarketinginterventioninZambiahadsome positive impacts
on equity with regard to ownership of insecticide-treated bednets,
but these were driven largely by increased ownership among me-
dium income groups. The intervention did little to increase bednet
ownership among the very poorest families where the price of even
subsided nets was still too great a barrier. Conversely, in another
bednet intervention in Tanzania, Nathan et al. (2004) found social
marketing was associated with an increase in equity; the ratio of net
three years of intervention. The authors note, however, that the
extremely high and this level of effect should not be assumed else-
where in the absence of this important ‘enabling’ factor.
economic groups. However, there is also pressing need for an
assessment of other harms and unintended consequences that may
arise as a result of such programmes (Kleinman, 2010; Knerr, 2011;
Pfeiffer, 2004), particularly those targeting sensitive or morally-
loaded topics such as hygiene or sexual behaviours. Our results sug-
gest that social marketing campaigns may potentially create a dual
disadvantage for the ultra-poor: they suffer from the inability to
suffer from the inability to comply with changes in social norms
engendered by successful social marketing interventions. Our inter-
hand-washing protected one’s family, made one feel ‘right’ and was
socially responsible e but the implicit assumption was that mothers
who did not hand-wash-with-soap were less than virtuous. That
ultra-poor women faced socio-economic adversity was well-
acknowledged within slum communities, but they also faced social
censure and accusations of laziness. By linking hygiene to social re-
sponsibility to drive a culturally-compelling intervention, our mes-
sage inadvertently further marginalised the poorest households
trapped by powerlessness, competing priorities, and despondency.
While the imperative to avoid stigmatising messages is well-known
in social marketing, it is difficult to see how this need is met, given
that social marketing programmes often specifically and explicitly
the idea of some kind of censure for those who fail to live up to it.
For many of the most intractable public health issues, a behav-
ioural approach wholly focussed on individual agency is insuffi-
cient (Stokols,1996) and potentially harmful, leaving the poor more
vulnerable and marginalised than before. An over-reliance on what
have been termed ‘neo-liberal’ approaches to health promotion
such as social marketing runs the risk of “embedding a behavioural
turn away from a focus on wider social and environmental de-
terminants of well-being” (Crawshaw, 2012: 207). Evaluating
whether structure or agency is the predominant driver of behav-
iour is a critical debate within public health (Abel & Frohlich, 2012;
Cockerham, 2005; Rütten & Gelius, 2011). As Cockerham noted,
‘there are situations in which structure can be so overwhelming
that agency is rendered ineffective’ (2005: 54). Interventions can-
not focus solely on changing behaviour; they must also be mindful
of the environment in which these behaviours take place. Complex
and complementary strategies that focus on both structure and
personal agency are thus required (Frohlich & Potvin, 2009).
The utility of the social marketing model for public health
therefore depends largely on the way in which it is deployed and
embedded within other, more structural, approaches. Most public
health issues require a combination of “both individual level and
environment/policy level interventions to achieve substantial
changes in health behaviours” (Sallis, Owe, & Fisher, 2008: 467).
Social marketing’s considerable strengths are best harnessed when
used as one of a suite of activities operating at different levels of
a social ecological model (McLeroy et al.,1988; Stokols,1996). Thus,
an intervention could harness the considerable strengths of social
marketing at the more proximate level, whilst simultaneously
intervening at the more distal, structural levels to help facilitate
behaviour change. With regard to hand-washing, for example,
structural interventions that address the wider problems of the
slums e poverty, pollution, crowding, inadequate basic services,
and social exclusion e are needed, in addition to more proximate
level interventions that focus on individual hygiene behaviours.
An in-depth understanding of one’s audience is, per se, not
enough to achieve behavioural change; we must also have
a detailed understanding of the environment and structures in
which human action takes place. The inclusion of ethnographic
data into the evaluation of this programme, for example, provided
in-depth understanding of everyday social experience and re-
sponses to the intervention. This allowed for a fuller, more accurate
understanding of the intervention’s impact. In the formative phase,
qualitative data focussedprimarily
regarding how best to ‘sell’ this behaviour (Fig. 1), while in the
evaluation phase, it helped identify the more upstream de-
terminants of health equity (Fig. 2). Had this study relied purely on
‘hard’ morbidity outcomes, this important equity dimension would
have been largely overlooked. As emphasized elsewhere in the
social sciences literature, the power of ethnographic evaluation lies
in exploring the “dynamics of context, practice, agency and power”
that shape human behaviour (Evans & Lambert, 2008: 467).
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141
Our ethnographic data have raised problematic equity issues, Download full-text
indicating that social marketing techniques would benefit from
careful and critical scrutiny. We need a clearer understanding of
how social marketing messages are interpreted by local commu-
nities over time. Equally, we need to pay attention to issues of
health equity, and to the unintended consequences of social action
in terms of social marginalization. We are mindful of a critical point
made in thehealth literature thatthe ‘social determinants of health’
are not identical to the ‘social determinants of health equity’ (Jones,
Jones, Perry, & Barclay, 2009). Furthermore, efforts to design
culturally-compelling interventions do not necessarily translate
into sustainable health impacts (Panter-Brick, Clarke, Lomas,
Pinder, & Lindsay, 2006) nor, as in the present study, into
economically-compelling interventions for the ultra-poor. Indeed,
a major shift in the field of development economics began with the
realization that the ultra-poor fail to take-up a whole range of
simple interventions that would significantly improve health and
productivity (Banerjee & Duflo, 2011; Karlan & Appel, 2011). It re-
mains a challenge to design interventions that will promote
effective and sustainable changes to well-being and will also make
sound cultural and economic sense for the most under-served.
This study was supported by a UK ESRC/MRC doctoral fellow-
ship, with additional funds from the Biosocial Society and Parkes
Foundation. The writing of this paper was undertakenwith support
from The Centre for the Development and Evaluation of Complex
Interventions for Public Health Improvement (DECIPHer), a UKCRC
Public Health Research Centre of Excellence. Funding from the
British Heart Foundation, Cancer Research UK, Economic and Social
Research Council (RES-590-28-0005), Medical Research Council,
the Welsh Assembly Government and the Wellcome Trust
(WT087640MA), under the auspices of the UK Clinical Research
Collaboration, is gratefully acknowledged.
Abel, T., & Frohlich, K. (2012). Capitals and capabilities: linking structure and agency
to reduce health inequalities. Social Science & Medicine, 74(2), 236e244.
Aboud, F., & Singla, D. (2012). Challenges to changing health behaviours in devel-
oping countries: a critical overview. Social Science & Medicine, 75(4), 589e594.
Agha, S., Van Rossem, R., Stallworthy, G., & Kusanthan, T. (2007). The impact of
a hybrid social marketing intervention in inequities in access, ownership and
use of insecticide-treated nets. Malaria Journal, 6, 13.
Aunger, R., Schmidt, W., et al. (2010). Three kinds of psychological determinants for
hand-washing behaviour in Kenya. Social Science & Medicine, 70, 383e391.
Banerjee, A., & Duflo, E. (2011). Poor economics: A radical rethinking of the way to
fight global poverty. New York: Public Affairs.
Biran, A., Tabyshalieva, A., et al. (2005). Formative research for hygiene promotion
in Kyrgyzstan. Health Policy and Planning, 20(4), 213e221.
Blankenship, K., Friedman, S., et al. (2006). Structural interventions: concepts,
challenges and opportunities for research. Journal of Urban Health, 83(1), 59e72.
Briscoe, C., & Aboud, F. (2012). Behaviour change communication targeting four
health behaviours in developing countries: a review of change techniques.
Social Science & Medicine, 75(4), 612e621.
Cockerham, W. C. (2005). Health lifestyle theory and the convergence of agency and
structure. Journal of Health and Social Behavior, 46(1), 51e67.
Crawshaw, P. (2012). Governing at a distance: social marketing and the (bio) politics
of responsibility. Social Science & Medicine, 75(1), 200e207.
Curtis, V., & Cairncross, S. (2003). Effect of washing hands with soap on diarrhoea
risk in the community: a systematic review. The Lancet Infectious Diseases, 3(5),
Curtis, V., Kanki, B., et al. (1997). Dirt and diarrhoea: formative research in hygiene
promotion programmes. Health Policy and Planning, 12(2), 122e131.
Curtis, V., Kanki, B., et al. (2001). Evidence of behaviour change following a hygiene
promotion programme in Burkina Faso. Bulletin of the World Health Organiza-
tion, 79, 518e527.
Donovan, J., Mills, N., et al. (2002). Improving design and conduct of randomised
trials by embedding them in qualitative research: (ProtecT) study. British
Medical Journal, 325, 766e769.
Evans, C., & Lambert, H. (2008). Implementing community interventions for HIV
prevention: insights from project ethnography. Social Science & Medicine, 66,
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behaviour: An intro-
duction to theory and research. Reading, MA: Addison-Wesley.
Frohlich, K. L., & Potvin, L. (2009). Structure or agency? The importance of both for
addressing social inequalities in health. International Journal of Epidemiology, 39,
Grier, S., & Bryant, C. (2005). Social marketing in public health. Annual Review of
Public Health, 26, 319e339.
Higginbotham, N., Briceno-Leon, R., & Johnson, N. (2001). Applying health social
science: Best practice in the developing world. London: Zed Books.
Jenkins, M., & Scott, B. (2007). Behavioral indicators of household decision-making
and demand for sanitation and potential gains from social marketing in Ghana.
Social Science & Medicine, 64, 2427e2442.
Jones, C. P., Jones, C. Y., Perry, G., & Barclay, G. (2009). Addressing the social de-
terminants of children’s health: a cliff analogy. Journal of Health Care for the Poor
and Underserved, 20(4 Suppl.), 1e12, Number.
Karlan, D., & Appel, J. (2011). More than good intentions: How a new economics is
helping to solve global poverty. New York: Dutton.
Kleinman, A. (2010). Four social theories for global health. The Lancet, 375(9725),
Knerr, W. (2011). Does condom social marketing improve health outcomes and
increase usage and equitable access? Reproductive Health Matters, 19(37),
Kolter, P., & Roberto, E. (1989). Social marketing strategies for changing public
behavior. New York: The Free Press.
Langford, R., Lunn, P., & Panter-Brick, C. (2011). Hand-washing, subclinical in-
fections, and growth: a longitudinal evaluation of an intervention in Nepali
slums. American Journal of Human Biology, 23(5), 621e629.
Loevinsohn, B. (1990). Health education interventions in developing countries:
a methodological review of published articles. International Journal of Epi-
demiology, 19(4), 788e794.
Lorenc, T., Petticrew, M., Welch, V., & Tugwell, T. (2012). What types of interventions
generate inequalities? Evidence from systematic reviews. Journal of Epidemiol-
ogy and Community Health, 67(2), 190e193.
Macintyre, S. (2003). Evaluating the evidence on measures to reduce inequalities in
health. In A. Oliver, & M. Exworthy (Eds.), Health inequalities: Evidence, policy
and implementation. Proceedings from a meeting of the health equity network (pp.
23e28). London: The Nuffield Trust.
McLeroy, K., Bibeau, D., et al. (1988). An ecological perspective on health promotion
programs. Health Education Quarterly, 15, 351e378.
Moffat, S., White, M., Mackintosh, J., & Howell, D. (2006). Using quantitative and
qualitative data in health services research e what happens when mixed
methods findings conflict. BMC Health Services Research, 6, 28.
Nathan, R., Masanja, H., Mshinda, H., Shellenberg, J. A., de Savigny, D., Lengeler, C.,
et al. (2004). Mosquito nets and the poor: can social marketing redress in-
equities in access? Tropical Medicine and International Health, 9(10), 1121e1226.
Panter-Brick, C., Clarke, S. E., Lomas, H., Pinder, M., & Lindsay, S. W. (2006). Cul-
turally compelling strategies for behaviour change: a social ecology model and
case study in malaria prevention. Social Science & Medicine, 62(11), 2810e2825.
Pawson, R., & Tilley, N. (2006). Realistic evaluation. In DPRN thematic meeting 2006
report on evaluation, Available from http://www.dprn.nl/thematic-meeting-
Pfeiffer, J. (2004). Condom social marketing, Pentecostalism, and structural
adjustment in Mozambique: a clash of AIDS prevention messages. Medical
Anthropology Quarterly, 18(1), 77e103.
Price, N. (2001). The performance of social marketing in reaching the poor and
vulnerable in AIDS control programmes. Health Policy and Planning, 16(3),
Roen, K., Arai, L., Roberts, H., & Popay, J. (2006). Extending systematic reviews to
include evidence on implementation: methodological work on a review of
community-based initiatives to prevent injuries. Social Science & Medicine,
Rütten, A., & Gelius, P. (2011). The interplay of structure and agency in health
promotion: integrating a concept of structural change and the policy dimension
into a multi-level model and applying it to health promotion principles and
practice. Social Science & Medicine, 73(7), 953e959.
Sallis, J., Owe, N., & Fisher, E. (2008). Ecological models of health behavior. In
K. Glanz, B. Rimer, & K. Viswanath (Eds.), Health behavior and health edu-
cation: Theory, research, and practice (4th ed.) (pp. 465e485). San Francisco,
Shrestha, B., & Shrestha, K. (2005). Study on water accessibility of slums and squatters
of Kathmandu Valley. Kathmandu: NGO Forum.
Stokols, D. (1996). Translating social ecological theory into guidelines for commu-
nity health promotion. American Journal of Health Promotion, 10(4), 282e298.
Wallack, L. (2002). Public health, social change, and media advocacy. Social Mar-
keting Quarterly, 8(2), 25e31.
R. Langford, C. Panter-Brick / Social Science & Medicine 83 (2013) 133e141