Are social marketing campaigns effective in preventing child abuse and neglect?

NCPC Issues No. 32 – October 2010

Are social marketing campaigns about child maltreatment effective?

In this section, evidence is examined regarding the extent to which social marketing campaigns targeting child maltreatment had an effective impact on levels of awareness, knowledge, attitudes, behaviours and prevalence of the social problems that they were designed to address. Box 2 below describes the common measures of success applied to social marketing campaigns targeting change in individuals and the broader community. The evaluation of social marketing campaigns is important because it provides an insight into the impact and effectiveness of this approach to reducing child abuse and neglect.

Box 2: Measuring success of social marketing campaigns related to child maltreatment

  • Raising awareness - measured as awareness of either the level of recognition or recall of the key campaign (campaign awareness) or of the social problem (topic awareness) across the population exposed to the campaign.
  • Imparting knowledge - an improvement in people's knowledge about what constitutes child abuse and improvement in people's knowledge about the risk factors of child abuse.
  • Changing attitudes - a measured improvement in people's attitudes related to child maltreatment. The importance of this change is based on the theory that attitude change is a precursor to behaviour change.
  • Changing behaviour - can be indicated by people's self-reported changes in contemplating and enacting behaviour (subjective) or by objective measures. Objective measures of behaviour change include:
    • reports from an independent observer (such as a social worker or psychologist report);
    • a decrease in demand for statutory child protection services (e.g., a change in rates of notifications and substantiations of child abuse and neglect); and
    • an increase in help-seeking behaviour, such as calls to a phone helpline and growth in demand for family support services.
  • Actual prevalence of child maltreatment - a measurable reduction in the prevalence of child abuse and neglect in the population would indicate community-wide change. This measure of success relies on accurate baseline prevalence data. However, this information is not currently available in Australia.

As shown in Table 2, 15 of the 21 campaigns had some form of evaluation but only 12 had reported evidence of impact and/or outcome evaluations. Box 3 provides a list of the 12 campaigns with published impact and/or outcome evaluation. Overall, this means that less than half of the Australian campaigns (5/12) compared to the majority of international campaigns (7/9) had published evidence pertaining to their impact. This finding is significant because it indicates that social marketing campaigns are not routinely evaluated in Australia for their effectiveness. Consequently, there is only limited evidence available to assess the short-term and longer-term outcomes in the Australian context.

The lack of published evidence on impact/outcome of Australian-based campaigns could be a reflection of limited resources. As discussed previously, evaluations require resources and publication of evaluation findings in the form of a report or journal article require further resources. Limited resources may also influence the impact of a social marketing campaign. The impacts of a campaign that is relatively short in duration (e.g., 6 weeks) are likely to be of less intensity than a campaign that is longer in duration (e.g., 2 years).

Some campaigns used more than one method to collect impact evaluation data. The three most common methods used to evaluate the impact and/or outcome of the 12 campaigns were:

  • pre- and post-campaign comparison surveys (4/12);
  • post-campaign survey only (4/12); and
  • data on the number of calls to phone helplines (4/12).

Box 3: Campaigns with impact and/or outcome evaluation evidence

The 12 social marketing campaigns related to child maltreatment that had information published pertaining to impact and/or outcome evaluations were:

  • It's Not OK to Shake Babies (NAPCAN, 1995) (Australia);
  • It's Got to Stop (Crundall & Trevena, 1997) (Australia);
  • NAPCAN child abuse prevention media campaigns (Keys Young, 2000) (Australia);
  • Accentuate the Positive (Moorhead, 1998; Henley et al., 1998) (Australia);
  • Be Cool...Not Cruel (Rudd & Jacob, 2000) (Australia);
  • Alcohol Abuse, Drug Abuse, Child Abuse, One Thing Leads to Another (Andrews, McLeese, & Curran, 1995) (US);
  • Stop It Now (Chasan-Taber & Tabachnick, 1999) (US)
  • Breaking the Cycle (Hall & Stannard, 1997; Stannard et al., 1998; Young et al., 1999) (NZ);
  • Triple-P Parenting Population Trial - mass media component (Prinz & Sanders, 2007; Prinz et al., 2009) (US);*
  • It's Not OK (Ministry of Social Development and Families Commission, 2007) (NZ);
  • Some Secrets You Have to Talk About (Hoefnagels & Baartman, 1997; Hoefnagels & Mudde, 2000) (Netherlands); and
  • Domestic Abuse - There's No Excuse: wave 11 Domestic Abuse Effects Children (Scotland Government, 2008) (Scotland).

Note: * The Triple-P Positive Parenting Program is an evidence-based parenting program that is often delivered within the context of a parenting group. The Triple-P parenting population trial sought to provide quality parenting information to a broad audience.

Raising awareness

Measures indicating awareness can relate to either awareness of a campaign itself (referred to below as campaign awareness) and awareness of the social problem (referred to below as topic awareness) that is the subject of the campaign. There was mixed evidence for how effective the campaigns were at raising campaign and topic awareness.

Campaign awareness

High levels of awareness of campaign awareness were generated in five out of 12 campaigns:

  • Accentuate the Positive: 69% of participants were aware of the television advertisement.
  • Be Cool...Not Cruel: 60% of participants had heard of the key character from the campaign (Captain Harley) and 22.5% could name him without prompting.
  • Breaking the Cycle: Awareness of three different advertisements was measured and reported as follows: 56% aware of "Backwards/Forwards" ad; 47% aware of "Vicious Cycle" ad; and 48% aware of parenting radio ads.
  • It's Not OK: 95% of participants recalled something from at least one of the It's Not OK advertisements.
  • Domestic Abuse - There's No Excuse (Wave 11): 72% of participants were aware of advertising/publicity on domestic abuse.

Awareness of the six NAPCAN child abuse preventions campaigns ranged from a low of 22% (Fathering the Future, 1997) to a high of 71% (It's Not OK to Shake Babies, 1994; Keys Young, 2000). Keys Young (2000) noted that campaign slogans had a higher rate of recall than media announcements.

Two problems associated with low levels of awareness were identified. The post-campaign survey of 300 people for the It's Not OK to Shake Babies campaign found less than half of respondents recalled the campaign message (NAPCAN, 1995). This was partly attributed to the TV adverts being screened too late at night due to reliance on pro-bono airtime. The second problem was that increasing awareness was less likely to be achieved when there were already high levels of community awareness about the subject of a campaign.

Topic awareness

In relation to topic awareness there was an overall 8% increase in unprompted awareness of emotional abuse as a social problem between the pre- and post-campaign surveys for the Breaking the Cycle campaign, in addition to awareness of the campaign itself (Hall & Stannard, 1997; Stannard et al., 1998; Young et al., 1999). In particular, there was a 12% increase in the proportion of participants who acknowledged that yelling/screaming and swearing at a child is a form of emotionally abusive behaviour.

For the It's Got to Stop campaign, there were no significant improvements in awareness about the negative effect of domestic violence on children (Crundall & Trevena, 1997). This was because, as the pre-campaign survey showed, existing levels of awareness were already high.

The effectiveness of a campaign can be misunderstood if existing levels of awareness, and existing attitudes and behaviours, are not taken into account. A formative assessment may overcome this problem by determining the current awareness, attitudes and behaviours of the target audience before the campaign is delivered.

Knowledge gain

Gains in knowledge were achieved by three out of 12 child maltreatment social marketing campaigns. Interestingly, not all survey measures recorded knowledge gain even though knowledge gain was a common aim across all the campaigns.

The Breaking the Cycle campaign recorded a 10% increase in knowledge of emotional abuse in the pre- and post-campaigns surveys (Stannard et al., 1998). The post-campaign survey (n = 1,012) for the Domestic Abuse - There's No Excuse campaign found 58% of respondents identified the message about the effects of domestic violence on children (Scotland Government, 2008). The Stop It Now campaign recorded a significant improvement in knowledge of child sexual abuse, with a 20% reduction in the number of people unable to define child sexual abuse across the pre- and post-campaigncomparison surveys (Chasan-Taber & Tabachnick, 1999).

Although the Stop It Now campaign recorded an increase in knowledge of the definition of child sexual abuse, the campaign evaluation reported that there was no significant improvement in the proportion of participants who knew the characteristics or warning signs of a child sexual abuser. After 2 years of the campaign only 27% of participants thought they would take action if they suspected child sexual abuse. It may be that the hard-hitting content of the Stop It Now campaign was ineffective because the campaign did not offer a positive model from which the audience could learn about responding to child sexual abuse.

Attitude change

Impact evaluations measured the extent of change in audience attitudes for four out 12 campaigns. Three campaigns recorded significant improvements in audience attitudes about child maltreatment:

  • 10% decrease in the proportion of participants who self reported that "parents who need help in their parenting role are failures" (Accentuate the Positive)
  • 10% increase in the proportion of participants who believe it is possible for child sexual abusers to live within one's own community3 (Stop It Now); and
  • 24% of participants state that their views on family violence have changed as a result of the It's Not OK campaign.

The fourth campaign associated with attitude change was Domestic Abuse - There's No Excuse about the effect of domestic violence on children. Although changes in attitude towards the view that domestic abuse is unacceptable did not change significantly between wave 10 and wave 11 of the campaign, strong agreement that domestic violence is unacceptable was sustained (Scottish Government, 2008).4 Progress had also been recorded since the inception of the original campaign in 1998.

All four campaigns were associated with community capacity building, public health initiatives and/or social policy. It is possible that the positive impact that the four campaigns had upon attitudes was aided by their broader community-level strategies.

Behaviour change

Of the 12 campaigns with impact and/or outcome evaluations, 10 intended to have an effect on either the intermediate or end behaviour of individuals. It is important to note that almost all the results discussed below are based upon parent self-reports of behaviour change. One limitation of self-reported measures is that they introduce the potential for response bias (Nutbeam & Bauman, 2006). It is possible that some people will be less likely to admit to behaviour as a result of a social marketing campaign because they are more aware of the sensitivities surrounding the issue. "Gold standard" evaluations that utilise independent observational measures can be costly (Sanders et al., 2000; Tomison, 2000). Observational measures are also problematic when investigating "hidden" social problems such as domestic violence and child maltreatment.

Intermediate behaviours

The intermediate behaviours reported upon were primarily related to help-seeking. For three campaigns, their effectiveness in influencing intermediate behaviours was measured according to the number of phone calls to helplines:

The Stop It Now campaign recorded an increase from 100 calls in the first year to 141 in the second. As of September 1997 the helpline had received a total of 241 calls. Of those callers 23% were perpetrators of child sexual abuse and 50% were people who knew a perpetrator or victim (Chasan-Taber & Tabachnick, 1999).5

The phone helpline for the Alcohol Abuse, Drug Abuse, Child Abuse, One Thing Leads to Another campaign provided information and advice to people who were concerned about a parent struggling with a substance abuse problem (Andrews et al., 1995). The helpline recorded a growth in average monthly calls, from 33.9 per month pre-campaign to 54.7 per month during the campaign. However, a key concern raised in the evaluation of the campaign was the poor capacity of support services to respond to the influx of referrals (Andrews et al., 1995). This reflects the need to adequately prepare the service system for the growth in demand when people are motivated to seek support following exposure to a social marketing campaign related to child maltreatment.

The Some Secrets You Have to Talk About campaign (Hoefnagels & Baartman, 1997; Hoefnagels & Mudde, 2000) established a helpline designed to support children and young people experiencing child sexual abuse. The phone helpline for Some Secrets You Have to Talk About recorded approximately 52,000 calls between the start of the campaign (23 September 1991) and the end of the school year (1 July 1992). This was triple the amount of calls received during the same period prior to the campaign. After three years though, the number of phone calls from children disclosing sexual abuse declined sharply (Hoefnagels & Baartman, 1997). The campaign can be considered successful given that it was implemented prior to the commencement of mandatory reporting of child sexual abuse in the Netherlands.

For these two campaigns however, the extent that children and young people felt that calling a phone line service actually helped them it is not known. Data on the number of calls to phone helplines provides some indication of the number of people who have recognised the social problem, contemplated taking action and then were motivated to take some action. A limitation of this measure though, is that it does not provide an indication of repeated attempts to practice a new behaviour or any long-term behaviour change after making contact with a helpline. The same limitation would apply to future evaluation measures using hits to campaign websites as an indication of help-seeking behaviour.

Another campaign that reported an impact upon intermediate behaviour was Accentuate the Positive. The Accentuate the Positive campaign evaluation reported that compared with the pre-campaign survey, parents responding to the post-campaign survey indicated 40% planned to change their parenting behaviour and 25% had already changed their behaviour because of the campaign.6 There had also been a 43% increase in parents nominating they would seek help for parenting problems. Given that the campaign strategy included community out-reach through positive parenting information shop-fronts, those motivated to change their behaviour had a location in the community to directly receive support.7

End behaviour

Changes in actual child maltreatment behaviours were reported in the evaluation of the Breaking the Cycle campaign. The pre- and post-campaign comparison surveys (n = 611) showed up to 44% of parents had contemplated changing their behaviour and up to 16% had actually changed their behaviour in relation to emotional and physical abuse of children. Some of the ways in which participants stated how their behaviour had changed included trying to stop yelling at, swearing at or putting their child down and ceasing to fight or argue in front of the child.

Furthermore, a post-campaign survey (n = 2,500) conducted to evaluate the It's Not OK family violence campaign showed that 22% of respondents reported having taken action against family violence (Ministry of Social Development and Families Commission, 2007). However, it was not specified what actions participants had taken in relation to family violence.

Two other campaigns that sought to impact upon actual child maltreatment behaviours, It's Not OK to Shake Babies (NAPCAN, 1995) and Domestic Abuse - There's No Excuse (Scotland Government, 2008) did not include measures of behaviour change in their impact evaluation. This limited the extent to which a positive change in individual behaviour could be demonstrated as a result of these campaigns.

Actual prevalence of child maltreatment

Strong evidence for changes in relation to actual prevalence of child maltreatment at a population level has been reported for the US Triple-P System Population Trial. One part of this trial was a media strategy relating to positive parenting. The outcome evaluation demonstrated a significant reduction in rates of substantiated child maltreatment, out-of-home care placement and child injuries requiring hospital attention compared to communities that were not exposed to the Triple-P campaign and the parenting education programs.

It is highly likely that the potential for a campaign to impact upon child maltreatment prevalence at a population level is reliant upon the resources and intensity of the campaign. The US Triple-P System Population Trial was 2 years in duration, a high level of intensity when compared with most of the other campaigns reviewed in this paper. The intensity of the campaign is demonstrated also by the fact that it was accompanied by a range of strategies including parenting seminars, skills training for parents and consultations for individual parents.

The success of the US Triple-P System Population Trial in reducing child maltreatment prevalence highlights the benefit of a program that pairs a media campaign with support services that are tailored to meet the needs of parents. However, it also highlights the benefit of an intensive program that incorporates a range of activities including a media campaign.

3 Chasan-Taber and Tabachnick (1999) described the significance of this particular finding as follows: “The findings indicate an increased level of awareness that child sexual abuse can happen within one’s own community to children with whom one is familiar” (p. 289).

4 Respondents’ attitudes towards domestic violence was measured via a series of attitude statements.

5 This helpline was instituted especially for the “Stop It Now” campaign (Chasan-Taber & Tabachnik, 1999, p. 285).

6 The available evaluation publications did not indicate how these parents planned to change their behaviour or what their changed behaviour constituted.

7 The published articles available for this systematic review do not report whether there was increased use of these supports within the community.