Objectives The aims of this study are to test the effectiveness of an online quiz competition in changing adolescents' alcohol-related attitudes and behaviour and to explore the fe¬asibility of using Internet viral marketing to reach a significant number of adolescents.

Hypotheses to be tested

1. The present online quiz competition is more effective in changing adolescents' alcohol-related attitudes and behaviours than traditional health promotion.
2. Internet viral marketing is effective in reaching a significant number of adolescents.

Design and subjects The study will constitute a cluster randomised controlled trial for 20 secondary schools (6,720 Grade 7 to 9 students). Schools will be randomised to intervention or control arm with equal likelihood. Students in intervention schools will be invited to take part in the Internet campaign whereas those in control schools will receive relevant promotional leaflets.

Study instrument An alcohol attitude and behaviour survey will be compiled with reference to standardised, validated instruments.

Intervention Participants will log into the website and answer alcohol-related multiple-choice questions. They will be encouraged to promulgate the campaign to others.

Main outcome measures Alcohol-related attitude and behaviour will be the primary outcome measures.

Data analysis Multilevel regressions will mainly be used to evaluate the difference in attitude and behaviour.

Expected results The results of the proposed study will provide evidence on the efficacy of an Internet intervention in modifying adolescents' attitudes and behaviour and guide further investigation into the prevention of and intervention in such risk behaviours as underage drinking.

Intervention Arm: Internet Quiz Competition Campaign

The intervention is designed based on the Theory of Planned Behavior. Participants will gain alcohol-related knowledge through the participation of the intervention, which would in turn alter their attitude concerning drinking and alcohol use. According to the Theory of Planned Behavior, such knowledge and attitude improvement would reduce participants' intention to drink and thus change or prevent their drinking behavior. To promote the intervention, the investigators will conduct a briefing session and a demonstration session for all eligible students in participating schools in the intervention arm. Campaign promotion brochures and posters will be distributed within those schools.

Campaign Design

Participants of the campaign will receive a referral code from those who refer them to the online quiz website; starters of the campaigns will receive the code directly from the research team. These referral codes, sets of 8-digit numbers, will be unique to each user and used to track the referral process. During registration, participants will be asked to provide basic demographic information and their referral code. After registration, their baseline attitudes and behavior will be assessed using the Alcohol Attitudes and Behavior survey.

The campaign will last for 30 days, starting from the day of the school promotion session. The aim of the competition from the participants' perspective will be to obtain the highest score possible to win a prize. They will have two ways of doing so.

The first way will be to answer the online quiz questions. The po-ints gained in this way will constitute the "answer score." Alcohol-related multiple-choice questions will be presented on the webpage one by one in random non-repeated order. Each correct answer will result in 10 points being awarded to the user's account. Giving an incorrect answer or skipping a question will not result in any points being deducted. Users will be shown the correct answer immediately after they have chosen their own. The purpose of this design is to promote correct alcohol-related knowledge and attitudes.

The second way of accruing points to obtain a high score will be to make referrals, a step designed to enhance the effectiveness of the Internet viral marketing strategy. Upon a successful referral, the referrer will receive all of the answer scores of their direct referrals (which will constitute their "referral score"). To minimize the chance of self-referrals, users will receive no points from making a successful referral. Furthermore, to ensure campaign competitiveness, the referral scores of lower-level users will not be added to starters' accounts. The answer score and referral score will be summed into a total score that will be used for prize distribution. Figure 2 illustrates the point system.

The campaign website will be based on our previous Internet viral marketing campaign, which has shown satisfactory user acceptance. The investigators will make necessary modifications such as enhancing the user interface and preventing duplicate registrations for our study.

Setting the quiz questions will be guided by the Elaboration Likelihood Model (ELM) of persuasion. The quiz questions will be designed by youth volunteers to convey accurate and interesting facts based on international and local sources such as World Health Organization and Hong Kong Department of Health reports and research articles from peer-reviewed journals. The questions will be concrete, specific, and of personal relevance to the participants so that they will be more likely to adopt a "central path" of message processing, which the ELM predicts will induce a more sustainable effect. An experienced clinician and a research assistant will examine every set of 100 questions and discuss them with the question setters to ensure language accuracy and information reliability. This process will be repeated until a total of 1000 quiz questions are generated. Finally, all of the questions will be proofread, refined, and approved by a local alcohol research team from the School of Public Health, University of Hong Kong.

Control Arm: Information Package

A comprehensive package on alcohol-related information will be distributed to all students in the control arm. The information package will contain both printed promotional leaflets from the Department of Health and an electronic guideline for understanding alcohol-related information. The electronic guideline sent through the school intranet system would serve as a unidirectional knowledge transfer medium, which would include a brief summary of the most recent research regarding alcohol and a comprehensive list of relevant website and information sources. The investigators will use these relevant websites and information sources in setting the quiz questions for the intervention arm. As a result, both intervention and control arms could have comparable accessibility to alcohol-related information and the sole contrast between the two groups would be the method of presentation (interactive online quiz competition vs unidirectional information package).

Due to the referral nature of the Internet campaign, it is possible that a small portion of students in the control group school would be referred to participate in the quiz competition, which could in turn dilute the effect size estimate. As a result, the investigators will ask the students in the control group in the postintervention survey whether they have participated in the Internet quiz competition campaign and exclude those who participate in subsequent data analysis.

Inclusion Criteria:

• Grade 7 to 9 students in participating schools

Exclusion Criteria:

• Students who could not comprehend written Chinese