Published on 27.08.12 in Vol 14, No 4 (2012): Jul-Aug
Preprints (earlier versions) of this paper are available at http://preprints.jmir.org/preprint/2162, first published May 25, 2012.
Online Advertising to Reach and Recruit Latino Smokers to an Internet Cessation Program: Impact and Costs
Background: Tobacco cessation among Latinos is a public health priority in the United States, particularly given the relatively high growth of this population segment. Although a substantial percentage of American Latinos use the Internet, they have not engaged in Web-based cessation programs as readily as other racial/ethnic subgroups. A lack of culturally specific advertising efforts may partly explain this disparity.
Objective: Phase I of this study focused on the development of four Spanish-language online banner advertisements to promote a free Spanish-language smoking cessation website (es.BecomeAnEX.org). Phase II examined the relative effectiveness of the four banner ads in reaching and recruiting Latino smokers to the cessation website.
Methods: In Phase I, 200 Spanish-speaking Latino smokers completed an online survey to indicate their preference for Spanish-language banner ads that incorporated either the cultural value of family (familismo) or fatalism (fatalismo). Ads included variations on message framing (gain vs loss) and depth of cultural targeting (surface vs deep). In Phase II, a Latin square design evaluated the effectiveness of the four preferred ads from Phase I. Ads were systematically rotated across four popular Latino websites (MySpace Latino, MSN Latino, MiGente, and Yahoo! en Español) over four months from August to November 2009. Tracking software recorded ad clicks and registrants on the cessation website. Negative binomial regression and general linear modeling examined the main and interacting effects of message framing and depth of cultural targeting for four outcomes: number of clicks, click-through rate, number of registrants, and cost per registrant.
Results: In Phase I, smokers preferred the four ads featuring familismo. In Phase II, 24,829,007 impressions were placed, yielding 24,822 clicks, an overall click-through rate of 0.10%, and 500 registrants (2.77% conversion rate). Advertising costs totaled US $104,669.49, resulting in an overall cost per click of US $4.22 and cost per registrant of US $209.34. Website placement predicted all four outcomes (all P values < .01). Yahoo! en Español yielded the highest click-through rate (0.167%) and number of registrants (n = 267). The message framing and cultural targeting interaction was not significant. Contrary to hypotheses, loss-framed ads yielded a higher click-through rate than gain-framed ads (point estimate = 1.08, 95% CI 1.03 1.14, P = 0.004), and surface-targeted ads outperformed deep-targeted ads for clicks (point estimate = 1.20, 95% CI 1.13 1.28, P < .001), click-through rate (point estimate = 1.22, 95% CI 1.16 1.29, P < .001), and number of registrants (point estimate = 2.73, 95% CI 2.14 3.48, P < .001).
Conclusions: Online advertising can be an effective and cost-efficient strategy to reach and engage Spanish-speaking Latino smokers in an evidence-based Internet cessation program. Cultural targeting and smoking-relevant images may be important factors for banner ad design. Online advertising holds potential for Web-based cessation program implementation and research.
J Med Internet Res 2012;14(4):e116
Cigarette smoking is a major cause of disease and death in Latinos . Approximately 12.5% of Latinos currently smoke [ ] translating into roughly 6.3 million Latino smokers. With a projected growth of the Latino population to 133 million by 2050 (30.2% of the US population) [ , ] and the potential for smoking prevalence to increase as Latinos become more acculturated [ , ], there may be more than 16 million Latino smokers by 2050 if smoking rates remain unchecked. Tobacco cessation among Latinos is a public health priority [ , ].
Web-based cessation programs are a promising approach to reach Latino smokers with evidence-based cessation treatment. More than 60% of Latinos (32 million) are online—projected to increase to 70% by 2014 —and almost half of online Latinos (45%) have used the Internet to search for health information [ ]. A growing body of evidence supports the reach, efficacy, and cost-efficiency of Web-based cessation programs [ - ], but additional research is needed to understand their impact among subgroups of online smokers, including racial/ethnic minorities. In general, there are few studies of smoking cessation interventions developed specifically for Latinos [ , ], and only two studies of the effectiveness of Web-based interventions for Latinos [ , ]. One of the challenges in conducting Web-based cessation research with Latinos in the United States is that they have not engaged in Internet cessation programs as readily as other racial/ethnic subgroups [ ] and recruitment of Latinos into Internet cessation trials has been challenging [ , , , ].
Online advertising is widely acknowledged as an increasingly important method to reach Latinos who use the Internet [- ]. Also known as “display ads,” banner advertisements appear as graphical ads embedded into a webpage, typically including a combination of static/animated images, text, and/or video designed to convey a marketing message and/or cause the user to take an action [ ]. Unlike “offline” ads (eg, billboards, newspapers, and flyers), banner ads can immediately link smokers to Web-based cessation programs, thus capitalizing on the motivation to quit when it occurs which can be critical to engaging consumers with cessation treatment [ , ]. Moreover, banner ads can target consumers by strategic placement on selected websites with synergistic content, known demographic profiles, or past online behavior. Therefore, online advertising may represent a “participant-friendly” and cost-efficient solution to reach and recruit online Latino smokers to Web-based cessation programs [ - ].
Reaching online Latino smokers requires not only efficient and effective advertising channels, but also an understanding of the target audiences’ preferred content and context of messages [, , ]. Approximately half of online Latinos use the Internet either primarily in Spanish (19%) or in both English and Spanish (28%), with the growth of both segments outpacing English language usage as more Spanish content becomes available online [ ]. Therefore, we tested ads in Spanish to reach this large and growing segment of online Latinos. With regard to message content, Prospect Theory [ , ] suggests that people respond differently to factually equivalent messages depending on whether they are framed to emphasize benefits through gain-framed messages (eg, “if you quit smoking, you will live longer”) or costs through loss-framed messages (eg, “if you do not quit smoking, you will die sooner”). Applied to health behaviors, research suggests that when behaviors have a relatively certain outcome, gain-framed messages are more persuasive; if behaviors result in an uncertain outcome, loss-framed messages are more effective [ ]. Because quitting smoking will almost certainly prevent disease, Prospect Theory predicts that gain-framed messages will be more persuasive than loss-framed messages. Recent evidence suggests that highlighting the benefits of quitting (through gain-framed messages) is more effective to encourage preventive behaviors such as smoking cessation [ - ]. However, there have been few studies that have examined the impact of message framing in health interventions designed specifically for racial/ethnic minorities [ , ].
With regard to message context, research with Latinos indicates that addressing cultural elements is critical in developing smoking cessation interventions [, - ]. The process of designing messages around group-level characteristics has been referred to as both cultural tailoring and cultural targeting [ - ]. We use the term “cultural targeting” in this study to refer to the delineation of a particular population segment. Cultural targeting can be done at a surface level (eg, attending to the visual characteristics or language of intervention materials) or at a deep level (eg, incorporating specific Latino cultural values) [ ]. In general, few studies have investigated how surface- and deep-targeted messaging functions in specific subgroups of racial/ethnic minorities [ ].
The purpose of this study was: (1) to investigate whether online advertising is an effective strategy to promote engagement with a Spanish-language cessation website among online Latino smokers, and (2) to identify the optimal message characteristics for this audience. To address these questions, we conducted a two-phased, mixed-methods study. In Phase I, formative research with online Latino smokers guided the development of four Spanish-language banner ad prototypes (deep-targeted/gain-framed, deep-targeted/loss-framed, surface-targeted/gain-framed, and surface-targeted/loss-framed). In Phase II, we tested the effectiveness of these four ads in reaching and recruiting Latino smokers to a free Spanish-language smoking cessation website. The primary outcomes in Phase II were: (1) the absolute number of clicks on an ad, (2) the click-through rate to the smoking cessation website, (3) the number of ad responders who went on to register on the cessation website (“registrants”), and (4) the cost per registrant. We hypothesized there would be an interaction between message framing and cultural targeting, such that the deep-targeted/gain-framed ad would outperform all others on all metrics. Based on Prospect Theory, we also hypothesized that gain-framed messages would outperform loss-framed messages on all metrics holding constant cultural targeting. Finally, we hypothesized that deep-targeted ads would outperform surface-targeted ads on all metrics holding constant the message frame.
Phase I: Advertising Development and Pre-testing
A multicultural, multidisciplinary, and bilingual expert panel with expertise in behavioral science, marketing, communication technology, and advertising identified two dominant cultural values to consider for cultural targeting of messages to Latino smokers: familismo and fatalismo. Familismo is central to the Latino health experience; family and close friends are often consulted for general health issues before a medical professional, and family is often a strong motivator of health behavior. Fatalismo refers to the belief “si Dios quiere” (“what God wants”) which often guides Latino perceptions about their control over health and wellness. Our formative Phase I focused on developing banner ad prototypes for each cultural value that crossed message frame (gain vs loss) with level of cultural targeting (surface vs deep) resulting in a total of eight Spanish-language ads (see). All ads were written in Spanish, developed in Adobe Flash with animation and motion by a professional creative developer, and included a call to action (ie, “click here for more information”).
To examine preferences for ad characteristics, 200 American Spanish-speaking smokers were recruited from two Spanish-language websites: MSN Latino  and MySpace Latino [ ]. A study invitation banner ad transferred potential participants to an online screening form where eligible individuals (ie, current smokers, aged 18 or older, and Hispanic/Latino ethnicity of any race) provided online consent and completed a survey. Survey items included demographics, smoking history, Internet use, general receptivity to online ads, and the language subscale of the Short Acculturation Scale for Hispanics (SASH) [ ]. Next, each participant sequentially viewed two of the eight ads. The two ads differed on cultural value (familismo vs fatalismo), but were identical in message frame and depth of cultural targeting (eg, gain-framed/deep-targeted) for direct comparison of message characteristics. Ad presentation was counterbalanced so that an equal number of participants saw each of the 8 ads first. Participants were asked to indicate their receptivity to the ads using a 10-point Likert scale (1 = not at all, 10 = very much). Participants also responded to 12 statements that addressed the strength of cultural targeting (eg, this ad is for someone like me), ad credibility (eg, this ad is believable), perceived efficacy of the ad (eg, this ad has me thinking about quitting smoking), and intention to respond (eg, I would click on this ad if I saw it) using a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). Finally, both ads were presented in tandem, and participants were asked which one they liked the most and why. An equal number of participants viewed each ad pair. The survey was administered in Spanish.
Descriptive statistics were used to examine the demographic, smoking history, Internet use, online advertising receptivity, acculturation measures, and ratings of individual ad characteristics. Analysis of variance (ANOVA) procedures, Chi-square tests (χ2), and t tests were used to evaluate differences between the two ad types (familismo vs fatalismo).
Phase II: Evaluation of Online Banner Ad Effectiveness
Phase II examined the effectiveness of the four final ads (see Multimedia Appendix) in promoting the free Spanish-language version of BecomeAnEX (www.becomeanex.org) . All ads were tested in Spanish (the English language version of each ad is shown in the Appendix for reference only). BecomeAnEX is a free, branded smoking cessation website designed to engage smokers through videos, interactive content, a personalized quit plan, and an online community of current and former smokers. Using a Latin square design ( ), the four ads (deep-targeted/gain-framed, deep-targeted/loss-framed, surface-targeted/gain-framed, and surface-targeted/loss-framed) were systematically rotated across four Spanish-language websites over four time periods of approximately one month each between August and November 2009. This design controlled for variability in ad performance associated with website placement and time.
The four websites chosen for this study were MySpace Latino , MSN Latino [ ], MiGente [ ], and Yahoo! en Español [ ]. Our media partner, PHD Media LLC, examined various metrics of reach and performance [ ] using data from comScore [ ]—a widely used Internet analytics company—to identify sites with the best ability to reach our target audience of American adult Latinos. We selected a mix of portals with vast foreign-language segments and broad reach to Latino users as well as social network sites that specifically targeted Latinos. Banner ads ran on the home page and in various sections of each site (eg, email, weather, news, women’s health, and men’s health) to optimize their reach.
Banner ads were assigned a click tag that linked an individual’s use of the BecomeAnEX website with the specific banner ad that prompted the visit. Each time an individual visited a website where one of the banner ads was placed, an impression (a unit of analysis analogous to a page view) was recorded by a tracking software program. If the user clicked on the banner ad, a click-through was also recorded. Click-through rate is calculated by dividing the total number of clicks by the total number of impressions, expressed as a percentage. This is a standardized metric that allows for comparison of ad performance across websites and varying costs by essentially controlling for the number of impressions . All banner ads linked directly to the Spanish BecomeAnEX home page. When a user arrived at the BecomeAnEX website from a banner ad, a “home page visit” was recorded and the click tag associated with the banner ad was recorded along with standard website utilization metrics (eg, whether the individual registered to become a member, number of page views, and number of minutes online). Advertising media contracts with the four websites were designed to ensure a specific number of impressions of each banner ad over each of the four time blocks. Advertising costs by website for each time block are shown in . We also report the cost per click for each ad—calculated as the total cost divided by the number of clicks in a given ad campaign—to enable comparison to other published studies. Cost per registrant is calculated as the total cost divided by total number of registrants in a given ad campaign.
|Website||Metric||Month 1||Month 2||Month 3||Month 4|
|MySpace Latino||Ad type||Surface gain||Deep loss||Deep gain||Surface loss|
|Cost per clickb (US$)||$4.45||$5.16||$4.72||$3.99|
|Cost per registrantc (US$)||$268.42||$1275.00||$1275.00||$728.57|
|Yahoo||Ad type||Surface loss||Surface gain||Deep loss||Deep gain|
|Cost per click (US$)||$2.94||$3.36||$3.90||$4.58|
|Cost per registrant (US$)||$73.76||$79.37||$235.22||$346.63|
|MSN Latino||Ad type||Deep gain||Surface loss||Surface gain||Deep loss|
|Cost per click (US$)||$3.84||$2.97||$2.86||$3.10|
|Cost per registrant (US$)||$184.51||$95.66||$92.61||$615.56|
|Mi Gente||Ad type||Deep loss||Deep gain||Surface loss||Surface gain|
|Cost per click (US$)||$9.57||$9.29||$8.89||$8.48|
|Cost per registrant (US$)||N/Ad||$6250.00||$2083.33||$3125.00|
a the number clicks divided by the number of impressions
b cost divided by the number of clicks in a given ad campaign (not an outcome metric in this study, but included in the table for comparison to other published studies)
c cost divided by total number of registrants in a given ad campaign
d cost per registrant cannot be calculated since there were 0 registrants during this segment of the study
First, descriptive statistics were calculated for each of the four outcome variables (ie, absolute number of clicks, click-through rate, number of registrants, and cost per registrant) by website placement, time, and ad type. Next, a series of regression models explored the interaction of message framing and cultural targeting, and the main effects of website placement, time, message framing, and cultural targeting. Three of the outcome variables are count data (eg, clicks and registrants) or are based on count data (eg, click-through rate) and typically would be analyzed using Poisson regression. However, since the variance of these outcomes demonstrated over-dispersion, we determined that Poisson regression models would not be a good fit. Instead, we used negative binomial regression models to examine clicks, registrants, and click-through rate, and general linear regression models to examine cost per registrant. We evaluated the presence of an interaction between message framing and cultural targeting by including an interaction term in the model. For those models without evidence of an interaction, we evaluated separately the main effects of message framing and cultural targeting for loss-framed versus gain-framed ads and surface-targeted versus deep-targeted ads, respectively. All statistical analyses were performed using SAS Version 9.2 for Windows .
Phase I: Prototype Development and Pre-testing
A total of 5,698,776 impressions to the Phase I recruitment banner ad were generated from MSN Latino and MySpace Latino between June 1-24, 2009. A total of 8231 individuals clicked on the study invitation ad. Of those who reached the survey site, approximately half (3712/7436, 49.92%) completed eligibility screening and 25.51% (947/3712) met the eligibility criteria. Of those eligible, 67.2% (636/947) provided informed consent, and of those who consented, 32.1% (204/636) completed the survey.
A total of 204 Latino current smokers completed the survey: 56.4% (115/204) were male, average age was 36.7 years (SD 11.7, range 18-70), 53.9% (110/204) were white, 3.4% (7/204) black, 4.9% (10/204) American Indian/Alaskan native, 1.5% (3/204) native Hawaiian/other Pacific Islander, 0.5% (1/204) Asian, and 35.8% (73/204) choose the “other” category. Participants used the “other” category to describe multiracial background and country of origin. The mean score on the SASH was 1.63 (SD 0.56, range 1.0-3.5) indicating a low level of acculturation. Respondents reported smoking an average of 10.1 cigarettes per day (SD 9.7). Over half (56.9%, 116/204) reported smoking every day (43%, 88/204 smoke some days). The average number of smoking days per month was 22.9 (SD 9.7). Approximately half (48.5%, 99/204) had their first cigarette > 60 minutes after waking suggesting a low level of nicotine dependence, and 84.3% (172/204) expressed a desire to quit smoking within the next 6 months. Respondents were largely Internet-savvy and receptive to online advertising: 72.5% (148/204) reported using the Internet several times a day and about half (49.5%, 101/204) reported using the Internet for more than 5 years. Two-thirds (66.2%, 135/204) of respondents viewed Web advertising favorably (ie, like it “somewhat” or “a lot”).
All eight ads were reviewed positively, with mean ratings of “liking” ranging from 7.24 (SD 2.8) to 8.17 (SD 2.6). For the 12 items related to ad characteristics, there were no differences between ratings of familismo versus fatalismo ads: all eight ads were rated positively across all 12 dimensions both with regard to mean score and proportion of positive responses. When presented with both ads simultaneously, most of the participants (55.9%, 114/204) preferred the familismo ad, and gave reasons that related directly to its cultural relevance and the value of familismo (eg, “Both talk about quitting smoking, but this one tells me to think about my family;” “This ad definitely has a more profound message and most of all talks about family—it is more profound and I identify with it;” or “Because it speaks of future consequences that could happen; both [ads] are important but I believe the family one leaves more of a mark”). Given these results, we selected the Latino cultural value of familismo—illustrated by a young boy playing soccer with his parents—to incorporate into the deep-targeted ads evaluated in Phase II.
Phase II: Evaluation of Online Banner Ad Effectiveness
Across the Phase II study period (August 10 to November 29, 2009), a total of 24,829,007 impressions were placed across the four websites for the four Spanish-language banner ads. There were 24,822 clicks on the banner ads, yielding an overall click-through rate of 0.10%. A total of 500 unique individuals registered on BecomeAnEX, yielding a conversion rate of 2.77%. The cost of the online advertising efforts totaled US $104,669.49, resulting in an overall cost per click of US $4.22 and an overall cost per registrant of US $209.34.
As shown in, website placement was a significant predictor of clicks (point estimate = 0.51, 95% CI 0.46 0.56, P < .001), click-through rate (point estimate = 0.45, 95% CI 0.41 0.48, P < .001), number of registrants (point estimate = 0.14, 95% CI 0.10 0.20, P < .001), and cost per registrant (P = 0.01). The MiGente website yielded the lowest number of clicks (n = 2766), click-through rate (0.037%), and number of registrants (n = 6), and highest cost per registrant (US $4166.67). Yahoo! and MSN Latino were the two best performing websites, with Yahoo! yielding a higher click-through rate (0.167% vs 0.144%; χ21 = 9.8, P = .002) and higher number of registrants (267 vs 193; χ21 = 8.0, P = .005).
Our main hypothesis that there would be a significant interaction between message framing and cultural targeting was not supported by any of the four outcomes (all P values > .05). Descriptive data for the message framing and cultural targeting interaction are shown in. Our second hypothesis that gain-framed ads would outperform loss-framed ads holding constant level of cultural targeting was not supported for absolute number of clicks, the number of registrants, or the cost per registrant. There was a statistically significant difference in click-through rate, with loss-framed ads yielding a higher click-through rate than gain-framed ads (0.105% vs 0.095%, point estimate = 1.08, 95% CI 1.03 1.14, P =.004). Our third hypothesis was that deep-targeted ads would outperform surface-targeted ads holding message frame constant; however, the results showed that for all outcomes, surface-targeted ads performed better than deep-targeted ads. Surface-targeted ads outperformed deep targeted ads on clicks (13,674 vs 11,148; point estimate = 1.20, 95% CI 1.13 1.28, P < .001), click-through rate (0.111% vs 0.089%; point estimate = 1.22, 95% CI 1.16 1.29, P < .001), registrants (386 vs 114; point estimate = 2.73, 95% CI 2.14 3.48, P < .001), and cost per registrant (US $135.75 vs US $458.49, P < .001).
|Yahoo! en Español||MSN Latino||MySpace Latino||MiGente|
|Click-through rate||0.167%||0.144%||0.074%||0.037%||< .001|
|Cost per registrant (US$)||$119.95||$141.15||$600.00||$4166.67||.01|
|Message Frame||Metric||Level of cultural targeting|
|Cost per registrant (US$)||$137.80||$533.04||$218.83|
|Cost per registrant (US$)||$133.77||$402.30||$200.65|
|Cost per registrant (US$)||$135.75||$458.49||$209.34|
This is one of the first studies to examine the impact and costs of online Spanish-language banner advertising to reach and recruit Latino smokers to a Web-based cessation intervention. Our strategy was specifically designed to test the efficacy of various messaging elements and website placement to generate online smoker response and engagement. Overall, the results demonstrated that online advertising can effectively reach Latino smokers: during the 4-month study period, 24,822 individuals responded to banner advertising yielding a click-through rate of 0.10%; of these, 500 registered on the BecomeAnEX Spanish-language smoking cessation website.
The overall click-through rate in this study (0.10%) compares favorably to industry averages  and other research studies. Klausner et al [ ] used online banner ads to promote a San Francisco Department of Public Health website for syphilis with an overall click-through rate in the 2-month study of 0.1% with different ads yielding click-through rates of 0.05% to 0.14%. Bull et al [ ] evaluated the effectiveness of online banner ads for recruitment to human immunodeficiency virus (HIV) and sexually transmitted disease (STD) prevention research trials and reported a click-through rate of 0.05%. A recent study examining the effectiveness of online study recruitment via Facebook advertising yielded a click-through rate of 0.05% [ ]. To some, these rates may raise questions about the generalizability of banner advertising. However, it is important to remember that of the tens of thousands of impressions placed on any given website, only a small percentage of those who view a banner ad may be members of the target audience or meet eligibility criteria. Given the current and growing magnitude of the Latino online audience, even reaching a small percentage can translate into relatively large numbers of smokers. This recruitment yield may, in fact, be comparable to or even higher than a newspaper or radio advertisement that yields several hundred responses when thousands of individuals may have been exposed to the ad. The primary difference is that the denominator can be determined with more certainty in online advertising compared to traditional mass media [ ].
Results also indicate that online banner ads can be cost efficient, particularly when compared to other traditional program recruitment methods for racial/ethnic minority participants. While there was significant variation in cost per registrant across website placement and message effects, the most efficient and cost-effective advertising approach (surface-targeted ads placed on Yahoo! en Español) yielded 210 registrants in 2 months at a cost of US $73-$79 per registrant. These results compare favorably to other efforts to recruit minority participants using traditional approaches (ie, proactive face-to-face methods or reactive “offline” media-based methods) that reported lower and/or slower recruitment yield at higher cost [- ]. The capacity of online advertising to reach and recruit 500 Latino smokers to a cessation website during a 4-month period has important implications for cessation-related eHealth research. Registered users of cessation websites report high levels of motivation to quit [ , , - ] and may represent the strongest pool available to efficiently recruit Latino smokers into clinical trials.
Contrary to the published literature and our hypotheses, results indicated that surface targeting was a more effective approach than deep targeting. However, these results should be considered within the context of several study parameters. The use of family imagery to reflect deep-level targeting and tobacco product imagery to reflect surface-level targeting was intentional because we were interested in examining two different types of imagery. It is possible that surface-targeted ads may have yielded better results because smokers are more likely to attend and respond to images featuring tobacco products given their more obvious relevance. Thus, we cannot conclude from this study that deep-level targeting is not effective. Future research should consider an additive approach in which deep cultural targeting graphics include culturally relevant elements as well as tobacco products/images to catch the attention of smokers.
Also contrary to our hypotheses, loss-framed messages yielded a higher click-through rate than gain-framed messages. This finding is intriguing, especially in light of the non-significant results for the other outcome metrics for message framing. On the one hand, it suggests that the behavior of clicking on an ad is indeed sensitive to message frame whereas the more distal behavior of registering on a cessation website is not. However, a number of variables may explain why this may be true. For example, previous advertising research has shown that consumers’ level of involvement or interest with a product or issue exerts a strong influence on the effect of message framing . Consumers with high involvement pay more attention to the issue of losses, whereas consumers with low levels of involvement with a product or issue pay more attention to gains in messaging [ ]. Thus, response variation across websites may reflect varying levels of engagement whereby visitors who are specifically looking to obtain general news information, check email, or connect in an online social network were more likely to attend to an ad with negatively framed content. Future research will be important to examine factors such as consumer involvement, ad placement, and message processing to better understand how to engage smokers at the point of an initial recruitment message.
Data from this study also suggest that social media websites may not be as fruitful for online recruitment as more general websites. Click-through rates from the MiGente and MySpace websites were significantly lower than MSN Latino and Yahoo! en Español. Bull and colleagues  also found lower-than-average click-through rates for MiGente (0.06%). These lower click-through rates may reflect differences in engagement patterns on social networking sites as compared to websites related to news and information, particularly for Latinos. Social networking sites, which seek to maximize interpersonal connections, may elicit high levels of engagement and possibly lower the probability of attending to an advertisement or of having a positive attitude toward online advertising on the site. An exception to this hypothesis is Facebook which has been an effective context for research recruitment [ , ] perhaps due to its particular advertising guidelines. Future research is needed to understand the opportunities and challenges of using online social networks to reach and recruit smokers to cessation interventions [ ].
The results should be considered within the context of several limitations. First, campaign costs reflect a variety of factors and may be difficult to replicate. Specific information related to costs and advertising yield is provided to enable comparisons. Second, the extent to which our results are generalizable to other websites is unclear. The similarity in results obtained for Yahoo! en Español and MSN Latino lends support to the notion that other portals may perform similarly, but this remains to be tested. Relatedly, these findings are limited to Latinos using a Spanish-language website. The effectiveness of online ads in reaching English-preferring or bilingual Latinos should be addressed in future research. Third, it is possible that repeat visitors to any of the four websites where we placed banner ads could have seen more than one study-related ad, and this repeat exposure may have primed a subsequent click. Fourth, although not specifically a limitation, it is important to note that each of the four outcomes may be influenced by various factors, including but not limited to the variables examined in this study (ie, message framing and cultural targeting). For example, click-through rate is largely determined by the content of an ad, but it may also be a function of the degree to which users of certain websites attend to advertising; the number of registrants involves not only clicking on an ad but also then perusing a website and deciding to provide personal information to register. Lastly, the reach of this method is obviously limited to those Latino smokers who are online. Alternative strategies are needed to reach “offline” Latino smokers with cessation services.
Despite these limitations and considerations, this study has a number of strengths. Evaluating the use of online advertising to promote evidence-based online cessation treatment among Spanish-speaking Latino smokers is unique within the field of tobacco control. We assembled a multicultural, multidisciplinary, and bilingual team with expertise that spans behavioral science, marketing, communication technology, and advertising to examine how to optimize our ability to reach and recruit online Latino smokers to an evidence-based cessation program. The Latin square quasi-experimental design was extremely efficient in allowing us to control for two potentially confounding factors: the volume of website traffic and time. We considered other research designs and approaches including laboratory experiments to examine cognitive processing factors such as message recall, message recognition, and psychological variables that mediate or moderate response to ads. However, message testing under highly controlled conditions at the expense of external validity would not yield the kind of information needed to successfully promote registration to a Web-based cessation website that could be used to inform recruitment of Latino smokers in subsequent cessation trials. We encourage the use of quasi-experimental methods as appropriate and hope that the results presented herein provide a useful benchmark in future eHealth research. Our approach may also help other public health researchers consider how to use the Internet for disease prevention.
In summary, identifying the effectiveness of various advertising parameters to promote an evidence-based online cessation program among Spanish-speaking Latino smokers holds potential for online program implementation and intervention research. Although a large number of Latino smokers use the Internet and are interested in using it for smoking cessation, the participation rate of Latinos in Web-based cessation interventions has been relatively low. Digital metrics such as click-through rate and registration rate related to various online message characteristics provide a deeper understanding of factors related to actual attention and appeal of online programs, particularly for specific subgroups . Future research is needed to build on the methods used in this preliminary investigation to better understand how to increase consumer demand for tobacco cessation services, especially among racial/ethnic minorities [ ].
Primary funding for this work was from the National Cancer Institute at the National Institutes of Health (R21 CA133319 to ALG). This work was also supported in part by the National Cancer Institute of the National Institutes of Health (U01 CA114593). The funding agency had no involvement in the conduct of the study or preparation of this manuscript. The authors would like to acknowledge the contributions of Jeffrey Costantino (Marketing Department, Legacy) in managing the creative development work done by GSD&M and the media planning and buying done by PHD Media LLC. We also thank Raymond Niaura, PhD, for his comments and assistance in conducting statistical analyses.
Conflicts of Interest
Multimedia Appendix 1
Final banner ads in English (N=4) and Spanish (N=4).PDF File (Adobe PDF File), 310KB
- Tobacco use among U.S. racial/ethnic minority groups--African Americans, American Indians and Alaska Natives, Asian Americans and Pacific Islanders, Hispanics. A Report of the Surgeon General. Executive summary. MMWR Recomm Rep 1998 Oct 9;47(RR-18):v-xv, 1 [FREE Full text] [Medline]
- Centers for Disease Control and Prevention (CDC). Vital signs: current cigarette smoking among adults aged ≥18 years--United States, 2005-2010. MMWR Morb Mortal Wkly Rep 2011 Sep 9;60(35):1207-1212 [FREE Full text] [Medline]
- Ennis SR, Rios-Vargas M, Albert NG. The Hispanic Population: Census Briefs. 2010. 2011 URL: http://www.census.gov/prod/cen2010/briefs/c2010br-04.pdf [accessed 2012-05-01] [WebCite Cache]
- United States Census Bureau. US Department of Commerce. 2008 Aug 14. An older and more diverse nation by midcentury URL: http://www.census.gov/newsroom/releases/archives/population/cb08-123.html [accessed 2012-05-01] [WebCite Cache]
- Pérez-Stable EJ, Ramirez A, Villareal R, Talavera GA, Trapido E, Suarez L, et al. Cigarette smoking behavior among US Latino men and women from different countries of origin. Am J Public Health 2001 Sep;91(9):1424-1430. [Medline]
- Unger JB, Cruz TB, Rohrbach LA, Ribisl KM, Baezconde-Garbanati L, Chen X, et al. English language use as a risk factor for smoking initiation among Hispanic and Asian American adolescents: evidence for mediation by tobacco-related beliefs and social norms. Health Psychol 2000 Sep;19(5):403-410. [Medline]
- US Department of Health and Human Services. Tobacco Use Among US Racial/Ethnic Minority Groups - African Americans, American Indians and Alaska Natives, Asian Americans and Pacific Islanders, and Hispanics: A Report of the Surgeon General. Atlanta, Georgia: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 1998. URL: http://www.cdc.gov/tobacco/data_statistics/sgr/1998/complete_report/pdfs/complete_report.pdf [accessed 2012-05-02] [WebCite Cache]
- Fiore MC, Jaén CR, Baker TB, et al. Treating Tobacco Use and Dependence: 2008 Update. Clinical Practice Guideline. Rockville, MD: US Department of Health and Human Services. Public Health Service; May 2008.
- Interactive Advertising Bureau Hispanic Research Working Group. US Latinos Online: A driving force. 2011 May. URL: http://www.iab.net/media/file/USLatinosOnline_RevisedMay2011FINAL.pdf [accessed 2012-05-01] [WebCite Cache]
- Fox S, Jones S. The social life of health information: Americans' pursuit of health takes places within a widening network of both online and offline sources. Washington, DC: Pew Internet & American Life Project; 2009 Jun. URL: http://www.pewinternet.org/~/media//Files/Reports/2009/PIP_Health_2009.pdf [accessed 2012-05-01] [WebCite Cache]
- Graham AL, Cobb NK, Papandonatos GD, Moreno JL, Kang H, Tinkelman DG, et al. A randomized trial of Internet and telephone treatment for smoking cessation. Arch Intern Med 2011 Jan 10;171(1):46-53. [CrossRef] [Medline]
- Muñoz RF, Barrera AZ, Delucchi K, Penilla C, Torres LD, Pérez-Stable EJ. International Spanish/English Internet smoking cessation trial yields 20% abstinence rates at 1 year. Nicotine Tob Res 2009 Sep;11(9):1025-1034. [CrossRef] [Medline]
- Cobb NK, Graham AL, Bock BC, Papandonatos G, Abrams DB. Initial evaluation of a real-world Internet smoking cessation system. Nicotine Tob Res 2005 Apr;7(2):207-216. [CrossRef] [Medline]
- Myung SK, McDonnell DD, Kazinets G, Seo HG, Moskowitz JM. Effects of Web- and computer-based smoking cessation programs: meta-analysis of randomized controlled trials. Arch Intern Med 2009 May 25;169(10):929-937. [CrossRef] [Medline]
- Walters ST, Wright JA, Shegog R. A review of computer and Internet-based interventions for smoking behavior. Addict Behav 2006 Feb;31(2):264-277. [CrossRef] [Medline]
- Barrera AZ, Pérez-Stable EJ, Delucchi KL, Muñoz RF. Global reach of an Internet smoking cessation intervention among Spanish- and English-speaking smokers from 157 countries. Int J Environ Res Public Health 2009 Mar;6(3):927-940. [CrossRef] [Medline]
- Rabius V, Pike KJ, Wiatrek D, McAlister AL. Comparing internet assistance for smoking cessation: 13-month follow-up of a six-arm randomized controlled trial. J Med Internet Res 2008;10(5):e45 [FREE Full text] [CrossRef] [Medline]
- An LC, Schillo BA, Saul JE, Wendling AH, Klatt CM, Berg CJ, et al. Utilization of smoking cessation informational, interactive, and online community resources as predictors of abstinence: cohort study. J Med Internet Res 2008;10(5):e55 [FREE Full text] [CrossRef] [Medline]
- Leykin Y, Aguilera A, Torres LD, Pérez-Stable EJ, Muñoz RF. Interpreting the outcomes of automated internet-based randomized trials: example of an International Smoking Cessation Study. J Med Internet Res 2012;14(1):e5 [FREE Full text] [CrossRef] [Medline]
- Webb MS, Rodríguez-Esquivel D, Baker EA. Smoking cessation interventions among Hispanics in the United States: A systematic review and mini meta-analysis. Am J Health Promot 2010;25(2):109-118. [CrossRef] [Medline]
- Lawrence D, Graber JE, Mills SL, Meissner HI, Warnecke R. Smoking cessation interventions in U.S. racial/ethnic minority populations: an assessment of the literature. Prev Med 2003 Feb;36(2):204-216. [Medline]
- Muñoz RF, Lenert LL, Delucchi K, Stoddard J, Perez JE, Penilla C, et al. Toward evidence-based Internet interventions: A Spanish/English Web site for international smoking cessation trials. Nicotine Tob Res 2006 Feb;8(1):77-87. [CrossRef] [Medline]
- McCausland KL, Curry LE, Mushro A, Carothers S, Xiao H, Vallone DM. Promoting a Web-Based Smoking Cessation Intervention: Implications for Practice. Cases in Public Health Communication & Marketing 2011;5(5 Proc):3-26.
- Graham AL, Bock BC, Cobb NK, Niaura R, Abrams DB. Characteristics of smokers reached and recruited to an internet smoking cessation trial: a case of denominators. Nicotine Tob Res 2006 Dec;8 Suppl 1:S43-S48. [Medline]
- Cox J. Getting Connected.: Hispanic Business Magazine; 2004 Dec. URL: http://www.hispanicbusiness.com/news/newsbyid.asp?id=19387&page=1&cat=&more= [accessed 2012-05-01] [WebCite Cache]
- La Ferle C. Lee W-N. Can English Language Media Connect with Ethnic Audiences? Ethnic Minorities' Media Use and Representation Perceptions. Journal of Advertising Research 2005 Mar;45(1):140-153. [CrossRef]
- Lopez M. iMedia Connection. 2005. Reaching Hispanics online URL: http://www.imediaconnection.com/content/6693.asp [accessed 2012-05-24] [WebCite Cache]
- Interactive Advertising Bureau. 2004. Reaching Hispanics through online marketing URL: http://www.iab.net/media/file/resources_docs_Hispanic_Presentation_Final.pdf [accessed 2012-05-01] [WebCite Cache]
- Interactive Advertising Bureau. 2012 Feb 26. IAB display advertising guidelines: the new 2012 portfolio URL: http://www.iab.net/guidelines/508676/508767/displayguidelines [accessed 2012-05-24] [WebCite Cache]
- Graham AL, Milner P, Saul JE, Pfaff L. Online advertising as a public health and recruitment tool: comparison of different media campaigns to increase demand for smoking cessation interventions. J Med Internet Res 2008;10(5):e50 [FREE Full text] [CrossRef] [Medline]
- Swanson GM, Ward AJ. Recruiting minorities into clinical trials: toward a participant-friendly system. J Natl Cancer Inst 1995 Dec 6;87(23):1747-1759. [Medline]
- Bolen S, Tilburt J, Baffi C, Gary TL, Powe N, Howerton M, et al. Defining "success" in recruitment of underrepresented populations to cancer clinical trials: moving toward a more consistent approach. Cancer 2006 Mar 15;106(6):1197-1204 [FREE Full text] [CrossRef] [Medline]
- Ford JG, Howerton MW, Bolen S, Gary TL, Lai GY, Tilburt J, et al. Knowledge and access to information on recruitment of underrepresented populations to cancer clinical trials. Evid Rep Technol Assess (Summ) 2005 Jun(122):1-11 [FREE Full text] [Medline]
- Lai GY, Gary TL, Tilburt J, Bolen S, Baffi C, Wilson RF, et al. Effectiveness of strategies to recruit underrepresented populations into cancer clinical trials. Clin Trials 2006;3(2):133-141. [Medline]
- Maibach E, Parrott RL. Designing health messages: approaches from communication theory and public health practice. Thousand Oaks, Calif: Sage Publications; 1995.
- CapturaGroup. Hispanic Online Marketing. 2012 Mar 07. Are online Hispanics part of your 2012 plans? URL: http://www.hispaniconlinemarketing.com/2012/03/are-online-hispanics-part-of-your-2012-plans/ [accessed 2012-07-06] [WebCite Cache]
- Kahneman D, Tversky A. Prospect Theory: An Analysis of Decision under Risk. Econometrica 1979 Mar;47(2):263-291.
- Tversky A, Kahneman D. The framing of decisions and the psychology of choice. Science 1981 Jan 30;211(4481):453-458. [Medline]
- Rothman AJ, Salovey P. Shaping perceptions to motivate healthy behavior: the role of message framing. Psychol Bull 1997 Jan;121(1):3-19. [Medline]
- Gallagher KM, Updegraff JA. Health message framing effects on attitudes, intentions, and behavior: a meta-analytic review. Ann Behav Med 2012 Feb;43(1):101-116. [CrossRef] [Medline]
- Toll BA, Martino S, Latimer A, Salovey P, O'Malley S, Carlin-Menter S, et al. Randomized trial: Quitline specialist training in gain-framed vs standard-care messages for smoking cessation. J Natl Cancer Inst 2010 Jan 20;102(2):96-106 [FREE Full text] [CrossRef] [Medline]
- Salovey P. National Registry of Evidence-based Programs and Practices (NREPP): SAMHSA's National Registry of Evidence-based Programs and Practices - Message Framing. 2002 Dec. Legacy program summary URL: http://www.nrepp.samhsa.gov/ViewLegacy.aspx?id=82 [accessed 2012-05-24] [WebCite Cache]
- Apanovitch AM, McCarthy D, Salovey P. Using message framing to motivate HIV testing among low-income, ethnic minority women. Health Psychol 2003 Jan;22(1):60-67. [Medline]
- Schneider TR, Salovey P, Apanovitch AM, Pizarro J, McCarthy D, Zullo J, et al. The effects of message framing and ethnic targeting on mammography use among low-income women. Health Psychol 2001 Jul;20(4):256-266. [Medline]
- Marin G, Marin BV. Research with Hispanic populations. Newbury Park: Sage Publications; 1991.
- Marin BV, Perez-Stable EJ, Marin G, Sabogal F, Otero-Sabogal R. Attitudes and behaviors of Hispanic smokers: implications for cessation interventions. Health Educ Q 1990;17(3):287-297. [Medline]
- Huerta EE, Macario E. Communicating health risk to ethnic groups: reaching Hispanics as a case study. J Natl Cancer Inst Monogr 1999(25):23-26. [Medline]
- Marín G. Defining culturally appropriate community interventions: Hispanics as a case study. Journal of Community Psychology 1993;21(2):149-161. [CrossRef]
- Pasick RJ. Response to Kreuter and Skinner: consensus regarding use of the terms targeting and tailoring. Health Educ Res 2001 Aug;16(4):503-508 [FREE Full text] [Medline]
- Kreuter MW, Skinner CS. Tailoring: what's in a name? Health Educ Res 2000 Feb;15(1):1-4 [FREE Full text] [Medline]
- Resnicow K, Baranowski T, Ahluwalia JS, Braithwaite RL. Cultural sensitivity in public health: defined and demystified. Ethn Dis 1999;9(1):10-21. [Medline]
- Resnicow K, Braithwaite R, Dilorio C, Glanz K. Applying theory to culturally diverse and unique populations. In: Glanz K, Rimer BK, Lewis FM, editors. Health behavior and health education: theory, research, and practice, 3rd Edition. San Francisco: Jossey-Bass; 2002:485-509.
- MSN Latino. URL: http://latino.msn.com/ [accessed 2012-07-24] [WebCite Cache]
- MySpace Latino. URL: http://latino.myspace.com/ [accessed 2012-07-24] [WebCite Cache]
- Marín G, Sabogal F, Marín B, Otero-Sabogal F, Pérez-Stable E. Development of a short acculturation scale for Hispanics. Hispanic Journal of Behavioral Sciences 1987;9:183. [CrossRef]
- BecomeAnEX en Español. URL: http://es.becomeanex.org/ [accessed 2012-07-24] [WebCite Cache]
- MiGente. URL: http://www.migente.com/ [accessed 2012-07-24] [WebCite Cache]
- Yahoo! en Español. URL: http://espanol.yahoo.com/ [accessed 2012-08-19] [WebCite Cache]
- Hackett J. comScore. 2011. What's my comp index? URL: http://blog.comscore.com/2011/03/whats_my_comp_index.html [accessed 2012-07-16] [WebCite Cache]
- comScore. URL: http://www.comscore.com/ [accessed 2012-07-24] [WebCite Cache]
- Novak TP, Hoffman DL. New metrics for new media: toward the development of web measurement standards. World Wide Web Journal 1997;2(1):213-246.
- SAS Institute Inc. SAS/STAT Users Guide, Version 9.2. Cary, NC: SAS Institute, Inc; 2008.
- Wasserman T. Mashable Business. 2011 Jan 24. Behind the race to build a better banner ad URL: http://mashable.com/2011/01/24/better-banner-ad/ [accessed 2012-07-06] [WebCite Cache]
- Klausner JD, Levine DK, Kent CK. Internet-based site-specific interventions for syphilis prevention among gay and bisexual men. AIDS Care 2004 Nov;16(8):964-970. [CrossRef] [Medline]
- Bull SS, Lloyd L, Rietmeijer C, McFarlane M. Recruitment and retention of an online sample for an HIV prevention intervention targeting men who have sex with men: the Smart Sex Quest Project. AIDS Care 2004 Nov;16(8):931-943. [CrossRef] [Medline]
- Ramo DE, Prochaska JJ. Broad reach and targeted recruitment using Facebook for an online survey of young adult substance use. J Med Internet Res 2012;14(1):e28 [FREE Full text] [CrossRef] [Medline]
- Harris KJ, Ahluwalia JS, Catley D, Okuyemi KS, Mayo MS, Resnicow K. Successful recruitment of minorities into clinical trials: The Kick It at Swope project. Nicotine Tob Res 2003 Aug;5(4):575-584. [Medline]
- Keyzer JF, Melnikow J, Kuppermann M, Birch S, Kuenneth C, Nuovo J, et al. Recruitment strategies for minority participation: challenges and cost lessons from the POWER interview. Ethn Dis 2005;15(3):395-406. [Medline]
- Graham AL, Lopez-Class M, Mueller NT, Mota G, Mandelblatt J. Efficiency and cost-effectiveness of recruitment methods for male Latino smokers. Health Educ Behav 2011 Jun;38(3):293-300. [CrossRef] [Medline]
- Cobb NK, Graham AL. Characterizing Internet searchers of smoking cessation information. J Med Internet Res 2006;8(3):e17 [FREE Full text] [CrossRef] [Medline]
- Graham AL, Cobb NK, Raymond L, Sill S, Young J. Effectiveness of an internet-based worksite smoking cessation intervention at 12 months. J Occup Environ Med 2007 Aug;49(8):821-828. [CrossRef] [Medline]
- Etter JF. Internet-based smoking cessation programs. Int J Med Inform 2006 Jan;75(1):110-116. [CrossRef] [Medline]
- Etter JF, Perneger TV. A comparison of cigarette smokers recruited through the Internet or by mail. Int J Epidemiol 2001 Jun;30(3):521-525 [FREE Full text] [Medline]
- Stoddard JL, Augustson EM. Smokers who use internet and smokers who don't: data from the Health Information and National Trends Survey (HINTS). Nicotine Tob Res 2006 Dec;8 Suppl 1:S77-S85. [Medline]
- Jung WS, Villegas J. The effects of message framing, involvement, and nicotine dependence on anti-smoking public service announcements. Health Mark Q 2011;28(3):219-231. [CrossRef] [Medline]
- Yoo CY. Interplay of message framing, keyword insertion and levels of product involvement in click-through of keyword search ads. International Journal of Advertising 2011;30(3):399-424.
- Bull SS, Vallejos D, Levine D, Ortiz C. Improving recruitment and retention for an online randomized controlled trial: experience from the Youthnet study. AIDS Care 2008 Sep;20(8):887-893. [CrossRef] [Medline]
- Fenner Y, Garland SM, Moore EE, Jayasinghe Y, Fletcher A, Tabrizi SN, et al. Web-based recruiting for health research using a social networking site: an exploratory study. J Med Internet Res 2012;14(1):e20 [FREE Full text] [CrossRef] [Medline]
- Cobb NK, Graham AL, Byron MJ, Niaura RS, Abrams DB, Workshop Participants. Online social networks and smoking cessation: a scientific research agenda. J Med Internet Res 2011;13(4):e119 [FREE Full text] [CrossRef] [Medline]
- Viswanath K, Kreuter MW. Health disparities, communication inequalities, and eHealth. Am J Prev Med 2007 May;32(5 Suppl):S131-S133. [CrossRef] [Medline]
- National Tobacco Cessation Collaborative. Innovations in Building Consumer Demand for Tobacco Cessation Products and Services. 2007 May. URL: http://www.rwjf.org/files/publications/other/InnovationsinBuildingConsumerDemand.pdf [accessed 2012-05-24] [WebCite Cache]
|ANOVA: Analysis of variance|
|SASH: Short Acculturation Scale for Hispanics|
Edited by G Eysenbach; submitted 25.05.12; peer-reviewed by D Ramo, S Bull; comments to author 13.06.12; revised version received 17.07.12; accepted 23.07.12; published 27.08.12
©Amanda L. Graham, Ye Fang, Jose L. Moreno, Shawn L. Streiff, Jorge Villegas, Ricardo F. Muñoz, Kenneth P. Tercyak, Jeanne S. Mandelblatt, Donna M. Vallone. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 27.08.2012.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.