Initial Outcomes of the VERB™ Campaign: Tweens' Awareness and Understanding of Campaign Messages

Introduction 

The initial effects of public communication and social marketing campaigns are termed the “proximal” or immediate effects on the target audience's level of awareness and understanding of the campaign itself. Although some stakeholders press eagerly for evidence of behavior change, social marketers hypothesize that initial awareness and understanding precede behavioral trialing and adoption.1 These proximal outcomes are critical sources of evidence of effect for the initial communication efforts.2

Attention to the proximal outcomes of mass-media campaigns is not a new idea but details of these outcomes are seldom reported. McGuire3 proposed that campaigns influence the target audience through a series of hierarchical and sequenced steps that start with campaign awareness and an understanding of the campaign message; these are followed by intermediate variables in the cascade of change, including changes in attitudes, beliefs, and behavioral intention. The final step is performing the behavior of interest. Thus, examining the initial campaign impact (e.g., awareness and understanding of the campaign itself) may lead to a greater probability of achieving intermediate and end-point behavioral change. Recent campaign evaluators have started to explore these initial effects. A Canadian campaign to increase walking among adults reported a positive relationship between the initial outcome of message awareness and the intermediate outcome of pedometer ownership.4 Increasing pedometer ownership in combination with campaign recall was associated with increased walking behavior.5 Knowledge of how well a campaign's messages penetrated the target audience is used to segment the audience into subgroups based on campaign recognition and understanding.6 Messages can then be modified to reach each subgroup. Other interventions such as community promotions and events targeting these subgroups can also be developed.7

The VERB™ campaign was a mass-media communications campaign based on social marketing principles that aimed to inspire children aged 9–13 years (tweens) to be physically active every day.8 Campaign planners used McGuire's hierarchy-of-effects model3 to conceptualize the campaign9 and used social cognitive theory10 and theory of planned behavior11 in the development of VERB's messages. A related paper in this issue tested the hierarchy-of-effects model in these data12 and demonstrated that in the first year of the campaign, tweens who were aware and understood the message were likely to begin engaging in physical activity; hence, the intermediate variables, including attitudes toward physical activity and outcome expectations were not directly relevant to the ways that this campaign influenced children aged 9–13 years. For this reason, exploring the initial variables and characterizing the populations that initially changed on these measures was the research purpose of this paper. In this analysis, (1) the initial associations between tweens' and parents' sociodemographic data and VERB awareness and message understanding were explored, and (2) population subgroups were identified for which adjustments in campaign strategy were needed.

Results 

Descriptive frequencies of the sociodemographic characteristics are shown in Table 1, alongside their associations with VERB campaign awareness and message understanding. The first column describes these data. Most parents reported low barriers to their tweens' physical activity and high approval of physical activity, but only around one third spent 3 or more days per week active with their tween. Only one quarter of tweens would rather watch TV than be active, and 29% reported more than 2 hours daily of screen time.

Table 1. Associations between sociodemographic factors among tweens and their parents and awareness and understanding of the VERB campaign (weighted percentages)

	Correlate	Total sample		Awareness of VERB Campaign							Understanding of VERB Campaign
				No awareness		Prompted awareness		Unprompted awareness		p value	No/low understanding		High understanding		p value
		n	%	n	%	n	%	n	%		n	%	n	%
	Overall	2729	—	640	25.6	1565	57.0	524	17.3	—	1510	57.0	1219	43.0	—
	Tween's gender
	Girl	1305	48.8	288	24.3	715	55.2	302	20.5	⁎⁎⁎	678	53.5	627	46.5	⁎⁎
	Boy	1424	51.2	352	26.9	850	58.8	222	14.4		832	60.3	592	39.7
	Tween's age group (years)
	10–11	1058	41.3	264	27.6	659	60.9	135	11.6	⁎⁎⁎	586	57.4	472	42.6	Ns
	12–14	1671	58.7	376	24.3	906	54.3	389	21.4		924	56.8	747	43.2
	Tween's race
	White	1910	62.1	392	21.7	1111	58.1	407	20.2	⁎⁎⁎	1000	53.1	910	47.0	⁎⁎⁎
	Other races	819	37.9	248	32.1	454	55.3	117	12.6		510	63.5	309	36.5
	Household income ($)
	≤25,000	480	21.6	179	38.7	244	51.0	57	10.3	⁎⁎⁎	319	68.4	161	31.6	⁎⁎⁎
	25,001–50,000	714	28.3	178	26.0	422	59.0	114	15.0		417	58.4	297	41.6
	>50,000	1535	50.1	283	19.8	899	58.5	353	21.7		774	51.3	761	48.7
	Parent's level of education
	High school grad or less	888	28.5	266	31.5	504	56.8	118	11.7	⁎⁎⁎	556	63.5	332	36.5	⁎⁎⁎
	Some college	908	32.9	224	26.5	531	57.5	153	16.0		503	57.1	405	42.9
	College graduate	933	38.6	150	16.8	530	56.7	253	26.5		451	48.1	482	51.9
	Tween's PA level
	<7 sessions/week	1397	52.8	375	29.4	782	55.6	240	15.0	⁎⁎⁎	844	61.9	553	38.1	⁎⁎⁎
	≥7 sessions/week	1332	47.2	265	21.4	783	58.6	284	20.0		666	51.6	666	48.4
	Parent's perceived barriers to PA
	Low	1788	62.0	365	21.9	1045	58.3	378	19.9	⁎⁎⁎	966	55.4	822	44.6	Ns
	Medium	788	32.3	230	31.3	437	55.9	121	12.8		452	59.1	336	40.9
	High	153	6.7	45	33.7	83	50.7	25	15.6		92	62.2	61	37.9
	Days parent and tween active together
	None	764	28.8	183	25.6	417	54.7	164	19.7	⁎	432	57.1	332	42.9	Ns
	1–2	1013	36.1	212	22.9	598	59.0	203	18.1		547	56.0	466	44.1
	≥3	952	35.2	245	28.5	550	56.9	157	14.6		531	58.0	421	42.0
	Tween would rather watch TV
	Agree	646	25.2	171	29.2	363	55.4	112	15.3	Ns	382	60.8	264	39.2	Ns
	Disagree	2083	74.8	469	24.4	1202	57.6	412	18.0		1128	55.7	955	44.3
	Parent shows approval for tween PA
	Agree	2582	94.1	588	24.9	1492	57.5	502	17.6	⁎	1407	55.9	1175	44.1	⁎⁎⁎
	Disagree	147	5.9	52	37.0	73	49.0	22	14.0		103	74.6	44	25.4
	Tween's TV time day before survey
	> 2 hours	765	29.0	167	23.5	426	55.6	172	20.9	⁎	433	57.3	332	42.7	Ns
	≤2 hours	1964	71.1	473	26.5	1139	57.6	352	15.9		1077	56.9	887	43.1

Chi square p values. Ns = not significant;

⁎ <0.05; ⁎⁎ <0.01; ⁎⁎⁎ <0.001.

The awareness and understanding prevalence data are in the first row of Table 1. About 17% of tweens had unprompted awareness of VERB; 57% had prompted awareness; and a quarter had no awareness of the VERB campaign. Overall, 43% of tweens had high levels of understanding of the VERB message. Ghost awareness, reporting seeing VERB at baseline when VERB had not yet been launched, was 17% (data not shown). Television was the largest reported source of awareness at 89%; magazines were the next largest source at 23%.

Correlates of Campaign Awareness 

After adjustment for all other variables, significant correlates of prompted awareness were the following (Table 2): residing in a household with a moderate income (OR = 1.6) or high income (OR = 1.8), or having a parent who is a college graduate (OR = 1.35). Tweens who reported at least 7 sessions of physical activity had higher odds of prompted VERB awareness (OR = 1.4) than did tweens who reported <7 sessions. Tweens who reported their parents did not praise their physical activity had 32% lower odds of prompted awareness than did tweens who reported their parents praised them.

Table 2. Adjusted and weighted odd ratiosa for tween and parent factors associated with level of campaign awareness

	Correlate	Wald χ2 (p value)	Prompted awareness vs no awareness	Unprompted awareness vs no awareness
			OR (95%CI)	OR (95%CI)
	Tween's gender
	Girl	10.72	1.00 (ref)	1.00 (ref)
	Boy	(0.0001)	0.91 (0.75, 1.10)	0.55 (0.42, 0.73)⁎
	Tween's age group (years)
	10–11	20.86	1.00 (ref)	1.00 (ref)
	12–14	(<0.0001)	1.04 (0.86, 1.26)	2.26 (1.71, 3.00)⁎
	Tween's race
	White	4.34	1.00 (ref)	1.00 (ref)
	Other races	(0.0174)	0.80 (0.64, 1.00)	0.60 (0.42, 0.86)⁎
	Household income ($)
	≤25,000	4.93	1.00 (ref)	1.00 (ref)
	25,001–50,000	(0.0017)	1.57 (1.17, 2.12)⁎	1.70 (1.14, 2.52)⁎
	>50,000		1.84 (1.33, 2.56)⁎	2.23 (1.44, 3.45)⁎
	Parent's level of education
	High school grad or less	10.06	1.00 (ref)	1.00 (ref)
	Some college	(<0.0001)	0.98 (0.73, 1.31)	1.22 (0.81, 1.85)b
	College graduate		1.35 (1.01, 1.81)⁎	2.73 (1.86, 4.02)b
	Tween's PA level
	<7 sessions/week	14.76	1.00 (ref)	1.00 (ref)
	≥7 sessions/week	(<0.0001)	1.43 (1.13, 1.81)⁎	1.88 (1.49, 2.37)⁎
	Parent's perceived barriers to their tween's PA
	Low	1.38	1.00 (ref)	1.00 (ref)
	Medium	(0.2515)	0.84 (0.63, 1.13)	0.70 (0.50, 0.97)⁎
	High		0.80 (0.44, 1.46)	1.00 (0.56, 1.78)
	Days parent and child active together
	None	1.41	1.00 (ref)	1.00 (ref)
	1–2	(0.2432)	1.12 (0.85, 1.50)	0.95 (0.65, 1.38)
	≥3		0.91 (0.72, 1.17)	0.73 (0.51, 1.05)
	Tween would rather watch TV
	Agree	0.15	1.00 (ref)	1.00 (ref)
	Disagree	(0.8620)	1.07 (0.84, 1.37)	1.04 (0.76, 1.43)
	Parent approves of tween's PA
	Agree	2.24	1.00 (ref)	1.00 (ref)
	Disagree	(0.1151)	0.68 (0.46, 0.99)⁎	0.73 (0.37, 1.42)
	Tween's TV time day before survey
	>2 hours	9.74	1.00 (ref)	1.00 (ref)
	≤2 hours	(0.0002)	0.82 (0.65, 1.04)	0.54 (0.41, 0.72)⁎

a Adjusted for tweens' physical activity level, gender, age group, race, household income, and parents' level of education ⁎ Significant at p<0.05

Correlates independently associated with unprompted awareness were being a girl, being aged 12–14 years, being white, being from a moderate- or high-income household, and having a parent with a college degree. Similarly, tweens who reported at least 7 sessions of physical activity had higher odds of unprompted VERB awareness. Tweens who watched 2 hours or less of television on the day before the survey had 40% lower odds of unprompted awareness than did tweens who watched more than 2 hours of television (Table 2, far right column).

Correlates of Message Understanding 

The factors independently associated with high understanding of the VERB message are shown in Table 3. These include being a girl, being white, having high income or college educated parents, being more physically active, and having parents who show approval of the tween's physical activity.

Table 3. Adjusted odd ratiosa for level of understanding by child and parent correlates (weighted odds ratios and 95% CI)

	Correlate	Wald χ2 (p value)	High understanding vs no/low understanding
			OR (95%CI)
	Tween's gender
	Girl	15.83	1.00 (ref)
	Boy	(0.0002)	0.71 (0.59, 0.84)⁎
	Tween's age group (years)
	10–11	0.13	1.00 (ref)
	12–14	(0.7172)	1.04 (0.86, 1.25)
	Tween's race
	White	4.25	1.00 (ref)
	Other races	(0.0435)	0.79 (0.63, 0.99)⁎
	Household income ($)
	≤25,000	5.00	1.00 (ref)
	25,001–50,000	(0.0098)	1.37 (1.04, 1.79)⁎
	>50,000		1.58 (1.18, 2.12)⁎
	Parent's level of education
	High school grad or less	5.43	1.00 (ref)
	Some college	(0.0068)	1.11 (0.83, 1.48)
	College graduate		1.45 (1.10, 1.91)⁎
	Tween's PA level
	<7 sessions/week	26.70	1.00 (ref)
	≥7 sessions/week	(<0.0001)	1.53 (1.30, 1.81)⁎
	Parent's perceived barriers to their tween's PA
	Low	0.38	1.00 (ref)
	Medium	(0.6888)	1.10 (0.85, 1.43)
	High		1.12 (0.72, 1.74)
	Days parent and tween active together
	None	0.08	1.00 (ref)
	1–2	(0.9268)	0.98 (0.75, 1.28)
	≥3		0.95 (0.74, 1.23)
	Tween would rather watch TV
	Agree	0.03	1.00 (ref)
	Disagree	(0.8565)	1.02 (0.78, 1.34)
	Parent approves of tween's PA
	Agree	14.12	1.00 (ref)
	Disagree	(0.0004)	0.50 (0.35, 0.72)⁎
	Child's TV time day before survey
	>2 hours	1.14	1.00 (ref)
	≤2 hours	(0.2908)	0.90 (0.74, 1.10)

a Adjusted for tweens' physical activity level, gender, age group, race, household income, and parents' level of education ⁎ Significant at p<0.05

Discussion 

The VERB campaign was aimed at all U.S. children aged 9–13 years. Using McGuire's hierarchy-of-effects model,3 awareness of the campaign and understanding of the campaign messages were the most immediate proximal impact variables examined (i.e., the variable closest in time to the intervention). After 1 year, almost three fourths of children aged 9–13 years knew about the campaign. Almost one fifth of tweens recalled the campaign spontaneously when asked a general question about having seen or heard any advertising for getting kids active. The overall levels of prompted and unprompted awareness exceeded expectations of the campaign planners. The goal for total media message awareness in the first year was 50%, and as reported here, total awareness was 74%. To date, there have been no other evaluated campaigns that specifically promoted physical activity to young people anywhere in the world. These levels of awareness after 1 year of a campaign are similar to, or better than, mass-media campaigns targeting other behaviors of young people.16, 17

There are two reasons for examining these proximal outcomes of awareness and understanding. First, impacting on these initial outcomes leads directly to physical activity behavior among the targeted tweens.12 Intermediate outcomes, such as attitudes and outcome expectations were not related to subsequent physical activity, at least in the first year, so the factors associated with these variables were not of use to the planning of the campaign's second year. Second, a fine-grained understanding of the subpopulation groups most likely to report these immediate outcomes provided important mid-course feedback to the VERB campaign. On the basis of having identified the groups least likely to be reached by these messages, changes were made in the messaging and media mix developed and purchased for subsequent years of the campaign.

Subgroup analyses here showed that girls, older tweens, and white tweens were more likely to report unprompted awareness. In addition, family socioeconomic status, measured through parental education and income levels, showed a direct association with proximal outcome measures. The reason for this association could be that the VERB message of having fun was salient for these groups, or that the messages could be accessed and understood better by some groups than by others. For example, print media (e.g., magazines), which tend to reach more girls than boys, was an important supplement to VERB's television advertising. Older tweens may have had greater access to the media channels VERB used than did younger tweens.

In subsequent years, strategies targeting boys were enhanced. Messages were developed that encouraged sports participation, emphasized skills development, and recognized boys' interest in competition. More media time was purchased on channels that targeted boys. Partnerships with professional sports leagues were developed with teams' logos added to the website. The strategy for younger tweens was adjusted to disseminate messages through popular cartoon characters and by providing opportunities for physical activity through events, the VERB website, and promotions led by cartoon characters.

That white tweens and tweens in the higher socioeconomic groups were more likely to have unprompted awareness and be in the high message understanding group was critically important information to the campaign planners. Since these proximal variables lead directly to behavioral outcomes, one might predict that achieving behavior changes may be more challenging for tweens from disadvantaged or minority groups. In addition, other factors that affect levels of physical activity among tweens in disadvantaged groups may include lack of facilities and supportive social and physical environments for activity. Finding that these groups were less likely to even receive the initial media message led campaign planners in subsequent years to emphasize inclusiveness and diversity in all mainstream communications. Although VERB continued to deliver its messages to all groups, the television messages and delivery targeting African-American and Hispanic youth were enhanced by purchasing more media on networks that target these groups such as BET and Univision. In response to the income gradient in exposure that could be due in part to lower cable television ownership in low-income households, the campaign increased VERB community-based promotions through schools and recreation and youth centers such as YMCA's and Boys & Girls Clubs. The message that you can develop your own game, anywhere, anyhow was developed in the second year partly because it would have more relevance to tweens who did not have access or resources to pursue organized sports.

Tweens with both prompted and unprompted awareness and high message understanding were more likely than unaware tweens to report being physically active during the week before the survey. The reason for their increased physical activity could be a response to the VERB advertising or, perhaps, active tweens noticed the advertisements more than non-active tweens did, an alternative explanation called reverse causation. It may be that less-active tweens are nonresponsive to a pro-physical activity message and may need non–media-driven strategies to encourage them to consider being more active. Finally, these tweens may need more sustained exposure to VERB's messages in order to create detectable changes in these tweens after several years of campaign exposure.

Another attribute that differentiated tweens with awareness and understanding from those without these characteristics was parental praise for being physically active, a supportive factor shown in other studies to be associated with tweens' physical activity.18, 19 These findings confirm the importance of the VERB campaign's strategy to target a secondary audience of parents and other adult influencers with a message encouraging their support of tweens' physical activity.

Tweens' screen time (more than 2 hours or less than 2 hours) was not associated with understanding of the message. This suggests that the message was relevant no matter how often the child saw VERB advertising on television. In addition, the amount of screen time was not associated with prompted recall of the campaign. However, tweens who watched less than 2 hours of television per day were less likely to recall the campaign spontaneously. Other campaigns have found similar associations of exposure and unprompted awareness16 and associations of exposure and brand recall.17 Taken together, how much a child watched television affected whether they spontaneously recalled the campaign, but once they had awareness of the campaign, watching more or less television did not affect their understanding of VERB's message.

This study had limitations. The data are self-reported and subject to social desirability biases. The small sample size for the unprompted awareness group led us to combine ethnic and racial minorities into an other races category; therefore, proximal outcomes among the ethnic and racial groups could not be differentiated. While the pre-campaign awareness levels could be a concern in outcome studies, the level of 17% ghost awareness is quite low and typical for advertising generally, and was not an influencing factor in the correlates of awareness for this study.

Increasing physical activity among young people is an important public health goal,20, 21 made ever more challenging by the countervailing societal forces that move tweens toward sedentary pursuits. Initiatives such as the VERB campaign provide a vehicle for reaching all tweens with a physical activity message. Consistent with good principles of public health, subgroups that had lower levels of campaign awareness and understanding were identified early, and strategies were built into subsequent campaign years to address them. The use of this initial outcomes assessment is an essential part of evaluating multi-year integrated public health programs.22 The process of evaluation should be continuous and cyclical, and data should be used to provide feedback and improve efforts in later program delivery. Campaign planners who are tempted to emphasize long-term outcomes should reconsider the usefulness of examining initial effects, and the subgroups that achieved them, in order to better target subsequent program efforts. By analyzing and responding to information from these initial analyses, the VERB campaign could channel resources more equitably and achieve greater population coverage during subsequent years of the initiative.