Perceived Harm of Online Drug-Encouraging Messages

Abstract

This study examines third-person perceptions (TPP) of two types of online messages—antisocial messages that encourage drug abuse and prosocial messages in the youth anti-drug campaign—and their relationship with support for three types of rectifying measures: restrictive, corrective, and promotional. A survey of 778 secondary school students (approximately equivalent to students in Grades 7-13) in Hong Kong found that the self–other perceptual gap in effects of prosocial messages was significantly smaller than that of antisocial messages. Regression analysis showed that the perceptual gap of antisocial messages significantly predicted greater support for all three types of rectifying measures, but the perceptual gap of prosocial messages did not. More emphasis should be laid on educating students about the possible harmful effects of online drug-encouraging messages among their peers, which might help promote support for restrictive, corrective, and promotional anti-drug measures.

Keywords

third-person effect rectifying measures drug-encouraging messages anti-drug campaign Internet

Introduction

The third-person effect (TPE) states that people tend to perceive greater effects of a message on others than on the self. This study tests the TPE in the context of youth drug abuse, investigating how perception of effects leads to behavioral consequences about youth anti-drug policies, such as supporting more legal actions or more anti-drug promotions. Two types of messages on the Internet are examined. The first concerns online messages that encourage drug use, such as positive experiences shared by drug users; as these are related to antisocial behaviors and are socially undesirable, they are labeled as “antisocial messages.” The youth anti-drug-abuse campaign, on the contrary, is socially desirable in nature and related to prosocial behaviors, so we name the second type as “prosocial messages.”

Data were collected in Hong Kong, where the trend in youth drug abuse has been drawing greater attention. In the 2008/09 Drug Use Among Students report (Narcotics Division of the Security Bureau HKSAR, 2009), the percentage of lifetime drug-taking students in secondary schools increased from 3.3% in 2005-2006 to 4.3% in 2007-2008. Although the reported drug-taking percentages of students in Hong Kong were much lower than that in the Western countries (above 30% in the United States; Narcotics Division of the Security Bureau HKSAR, 2009), the hidden nature of youth drug abuse in Hong Kong may lead to a lower estimation of the number (Educational Bureau, 2014). In the Chinese society, drug abuse is highly stigmatized—77.6% of drug abuse students had never sought help from others (Narcotics Division of the Security Bureau HKSAR, 2013b). Among people who first reported drug abuse in 2013, more than half have been using drugs for 4.6 years or above, more than double the number (2.1 years) 5 years ago (Hong Kong’s Information Service Department, 2014). Findings from this study can provide implications for other places with an increasing trend of youth drug abuse and a culture of hidden drug abuse.

From 2008 to 2010, the Narcotics Division launched a 2-year, territory-wide anti-youth-drug-abuse campaign with the theme “No Drug, No Regrets; Not Now, Not Ever.” Anti-drug ads were broadcast on mass media including TV and radio, and many promotional activities such as training camps and talks were organized for the youth. Among these, the Narcotics Division started to make use of the Internet as a tool for anti-drug promotions. For example, the Narcotics Division collaborated with the Hong Kong Education City to set up a website with teaching resources and materials, and where teachers and social workers can share anti-drug resources and experiences (Narcotics Division of the Security Bureau HKSAR, 2011).

In spite of the government’s educational messages in the anti-drug-abuse campaign, youths still easily receive harmful messages that encourage them to abuse drugs, especially from the Internet. A survey by the Boys’ and Girls’ Clubs Association of Hong Kong (2010) showed that youths who were active in online discussion boards or forums were more likely to have had contact with information of drug abuse shared by current drug abusers; 10.5% of the respondents who frequented Internet cafes had used psychotropic drugs before, which was 10 times the percentage of those who did not go to Internet bars. In addition, 35% of the respondents reported to have read messages encouraging drug abuse on the Internet. Using the TPE as a theoretical framework with literature support, this study provides better-supported implications for attacking the problem of youth drug abuse.

This study is theoretically significant in several ways. First, as many past TPE studies have only studied Internet pornography, this study adds onto the literature by exploring the TPE of another type of harmful online message, drug-encouraging messages. Besides, supporting evidence for the linkage between third-person perception (TPP) and behavioral intentions has been largely mixed (Xu & Gonzenbach, 2008), because some studies used the perceptual gap as the predictor of behaviors whereas others used perceived effects on self and on others as predictors (Sun, Shen, & Pan, 2008). This study is significant in providing clear evidence on the TPP-behaviors linkage by examining the predictive powers of TPP and perceived effects on self and on others altogether. More importantly, research on the behavioral consequences of TPP of socially desirable messages has not been sufficient (Golan & Day, 2008). Therefore, this study contributes to the existing literature by investigating possible behavioral consequences of prosocial anti-drug messages. Whereas Sun, Shen, and Pan (2008) examined restrictive actions for harmful messages, corrective actions for messages with ambiguous influence, and promotional behaviors for PSAs, our study further broadens the scope of TPE research to all three types of rectifying behavioral actions for both antisocial and prosocial messages.

Literature Review and Theoretical Framework

TPE and TPP

The TPE was first proposed by Davison (1983). He suggested the third-person hypothesis, which predict that people tend to overestimate the influence that mass communications have on the attitudes and behavior of others. The TPP has been supported by evidence in a number of studies (Paul, Salwen, & Dupagne, 2000; Perloff, 1993, 1999). Sun, Pan, and Shen’s (2008) meta-analysis examined 60 articles to show that the TPP is robust and is not affected by different research procedures. Among different proposed explanations of the TPP, the self-enhancement explanation (Gunther & Mundy, 1993; Gunther & Thorson, 1992) is widely examined and supported. It assumes that people evaluate media influence based on a desire to maintain a positive image of themselves (Zhang, 2010). To protect their self-image, people perceive themselves less easily to be affected by socially undesirable messages than others—leading to TPP.

In view of the rise of new media, a commonly examined type of Internet content for the TPE in previous research has been Internet pornography; studies have affirmed the TPE of Internet pornography, that is, respondents perceived others to be more easily affected by Internet pornographic content than themselves (e.g., Lo & Wei, 2002; Wu & Koo, 2001; Zhao & Cai, 2008).

This study examines online antisocial messages related to youth drug abuse, including online messages on forums, blogs, instant messages, and so on that talk about ways to purchase illegal drugs, drug-taking experience and suggestions, methods of taking drugs and ways to escape from arrest. As the self-enhancement explanation suggests, people protect their self-image by perceiving others more easily affected by the antisocial Internet messages. Past studies have provided strong evidence on the third-person perceptual gap. The following hypothesis is proposed:

Hypothesis 1 (H1): Respondents will perceive antisocial messages on the Internet which encourage drug use to have a greater effect on others than on the self.

The self-enhancement explanation is based on the assumption that message desirability affects influence judgment (Perloff, 2002; Sun, Shen, & Pan, 2008). When it comes to estimating the effects of socially desirable messages, people tend to perceive themselves more easily affected as a form of self-promotions (Zhang, 2010). The third-person perceptual gap may thus become smaller (e.g., Brosius & Engel, 1996; Eveland & McLeod, 1999), nullified (e.g., Gunther & Mundy, 1993; Gunther & Thorson, 1992), and even reversed (e.g., David, Liu, & Myser, 2004; Duck, Terry, & Hogg, 1995; Henriksen & Flora, 1999).

There have been extensive studies on the TPE of public service announcements (PSAs), such as AIDS prevention (e.g., Chapin, 2000; Duck et al., 1995), traffic safety (e.g., Hoorens & Ruiter, 1996), drunk driving (e.g., Duck & Mullin, 1995), and antismoking campaigns (e.g., Henriksen & Flora, 1999). Regarding online messages, Zhong (2009) examined the self–other perceptual gap of both prosocial and antisocial online games, detecting significant TPP in both types of games, but TPP of prosocial game was significantly smaller than that of antisocial game. With the support of the self-enhancement explanation and past studies about TPP of prosocial messages, it is hypothesized as follows:

Hypothesis 2 (H2): The perceptual gap in effects on the self and on others of prosocial messages in the youth anti-drug-abuse campaign is smaller than that of antisocial messages that encourage youth drug abuse.

Behavioral Component of TPE

Mutz (1989) was the first to determine that TPP, which assumes that others are more influenced by what they have read, heard or viewed, might alter perceptions of the distributions of opinion in the society, thus changing public opinion in the society as a whole. Gunther’s (1995) study showed that people’s support for pornography restrictions was consistent with the self–other perceptual gap of pornography’s effect. Rojas, Shah, and Faber (1996) indicated that individuals with greater TPP were more likely to manifest pro-censorship attitudes on media. Support for censorship has been shown in many studies as a unique and main behavioral effect of the TPE (Tewksbury, Moy, & Weis, 2004).

Gunther and Storey (2003) extended the behavioral component of TPE to “the influence of presumed influence,” proposing the indirect effects model—people’s perceived influence of a communicative message on others can change their own attitudes or behaviors. Tal-Or, Cohen, Tsfati, and Gunther (2010) methodologically conducted an experiment to show that the linkage between TPP and behavioral intentions is causal—perceived effects cause change in behaviors instead of the opposite.

Furthermore, Sun, Shen, and Pan (2008) found that TPP predicted rectification behaviors: restrictive, corrective, and promotional behaviors. People recognize the problematic situation caused by media messages and define it by perceived media effects—“in anticipation of an inadequacy of others in dealing with media effects, in either resisting harmful influence or benefiting from positive influence, individuals may be inclined to redress the situation” (Sun, Shen, & Pan, 2008, p. 259). Rojas (2010) provided evidence that TPP of powerful media effects on public opinion was associated with off-line and online political participation to “correct” potential biases caused by media.

In Zhao and Cai’s (2008) study, self-enhancement, measured with a multi-item optimistic bias scale, predicted TPP of Internet pornography, which in turn predicted support for censorship. People who have higher self-image are more likely to perceive greater effects of antisocial messages on others and smaller effects on the self, that is, a greater self–other perceptual gap. They also tend to enhance self-image by supporting measures to protect others from such bad effects. As a result, the perceptual gap in effects can predict behaviors positively.

The antisocial nature of online drug-encouraging messages is similar to Internet pornography. TPP of drug-encouraging messages can lead to censorship as a rectifying measure of the messages’ harmful effects, as evidence of research in Internet pornography has shown. Sun, Shen, and Pan (2008) suggested, “Three types of rectifying behaviors in accordance with presumed severity and valence of potential messages influences” (p. 261), and conceptualized restrictive, corrective, and promotional behaviors specifically as regulating harmful messages, adding more educational elements in messages with mixed social implications, and more promotions on socially beneficial messages, respectively. They examined restrictive behaviors for Internet pornography, corrective behaviors for reality TV shows, and promotional behaviors for PSAs. However, we suggest restrictive actions are not the only possible rectifying measures for antisocial messages. Because the effects of drug-encouraging messages influence youth drug abuse as a whole, all three types of rectifying measures are plausible to deal with the drug problem. In this case, restrictive measures include legal actions against illegal drug use. Corrective measures include actions with mixed social implications like paying more attention to reports on online harmful messages about drug abuse, and supporting greater government investigation on how to prevent youth from being affected by harmful online information. Promotional measures include more promotions and education to students on drug abuse. From existing literature and the self-enhancement explanation, we propose the following:

Hypothesis 3 (H3): The perceptual gap in effects between the self and others of antisocial messages will be a significant predictor of support for restrictive measures.

Hypothesis 4 (H4): The perceptual gap in effects between the self and others of antisocial messages will be a significant predictor of support for corrective measures.

Hypothesis 5 (H5): The perceptual gap in effects between the self and others of antisocial messages will be a significant predictor of support for promotional measures.

The behavioral component of the TPE related to prosocial messages has not been explored a lot, and evidence is mixed. Meirick (2008) found that perceived effects on the self significantly predicted support for funding anti-drug ads and education, whereas perceptual gap in effects of PSAs did not predict the likelihood of engaging in promotional actions in Sun, Shen, and Pan’s (2008) study.

We will examine how TPP of prosocial messages is associated with three types of rectifying measures. The self-enhancement explanation suggests that people enhance their self-image by estimating others being more affected by media messages than themselves; nonetheless, when the media messages are socially desirable, being influenced by the messages does not hurt the self-image as much as socially undesirable messages do. Therefore, TPP of prosocial messages tend to be smaller than that of antisocial messages. We propose this reduced TPP will be less predictive to behaviors:

Hypothesis 6 (H6): The self–other perceptual gap in effects of antisocial messages will be a better predictor of support for restrictive measures than the self–other perceptual gap of prosocial messages.

Hypothesis 7 (H7): The self–other perceptual gap in effects of antisocial messages will be a better predictor of support for corrective measures than the self–other perceptual gap of prosocial messages.

Hypothesis 8 (H8): The self–other perceptual gap in effects of antisocial messages will be a better predictor of support for promotional measures than the self–other perceptual gap of prosocial messages.

Method

Data Collection

Data were collected by a survey of students in secondary schools (approximately equivalent to Grades 7-13), as they are the main targets of the anti-drug-abuse campaign. Using a multistage cluster sampling, we first drew 10 secondary schools at random from a pool of 506 secondary schools in Hong Kong. We then sent a letter to the principal of each of the 10 schools to obtain consent for data collection. Two schools agreed to take part in the survey. We then randomly selected two classes from each grade in each school. Finally, we administered the questionnaire to all students in these 28 classes between September and October 2010. Of the 804 students in the 28 classes, 778 (96.8%) completed questionnaires for analysis.

Measurements of Key Variables

Perceived effects of antisocial messages on the self and on others

Respondents estimated whether those messages made them (a) less anti-drug, (b) more identified with drug abusers, and (c) more willing to try drugs. The response categories were set with a 5-point Likert-type scale, ranging from 1 = no effect at all to 5 = a strong effect. Principal component factor analysis with varimax rotation showed that the three items were grouped into a single solution, with an Eigenvalue of 2.72 and 90.8% of variance explained. Therefore, the three “self” items were summated and divided by three to form a measure of “perceived effects of antisocial messages on self” (M = 2.29, SD = 1.10, α = .95).

As past research often relied on single-item measures of perceived effects on self and on others, or they used multiple items to measure perceived effects but did not create summated scales (Perse, Rubin, Rubin, Graham, & Seibold, 2009), this study created a summated scale with three items to measure perceived effects. The Cronbach’s alpha of the items was high at .94 (M = 2.74, SD = 0.94). Principal component factor analysis showed that the three items were grouped into a single solution, with an Eigenvalue of 2.70 and 89.91% of variance explained. Thus, the three “others” items were also averaged to create a measure of “perceived effects of antisocial messages on others” (M = 2.74, SD = 0.94, α = .94).

Perceived effects of prosocial messages on self and on others

Respondents were asked to estimate the perceived effects of online anti-drug messages in three aspects, including making them (a) more anti-drug, (b) less identified with drug abusers, and (c) possess better knowledge of harm of drug abuse. The first two items are similar with perceived effects of antisocial messages to provide a basis for comparison, and the third item was created specifically in the context of anti-drug messages. The response categories were also set with a 5-point Likert-type scale. Principal component factor analysis confirmed the three “self” items as a single solution, with an Eigenvalue of 2.67 and 89.55% of variance explained. The three items were averaged to form a measure of “perceived effects of prosocial messages on self” (M = 2.77, SD = 1.04, α = .94).

Principal factor analysis put the three items measuring perceived effects of prosocial messages into a single solution, with an Eigenvalue of 2.64 and 87.84% of variance explained. Thus, the three “others” items were averaged to constitute a measure of perceived effects of prosocial messages on others (M = 2.89, SD = 0.94, α = .93).

Perceptual gap in effects of antisocial messages and prosocial messages

Perceptual gap in effects was computed with the score of perceived effect on others minus that of perceived effects on the self. The mean perceptual gap in effects of antisocial messages was .45 (SD = 1.06). The mean score of perceptual gap in effects of prosocial messages was .12 (SD = .91).

Support for rectifying measures

Respondents were asked to indicate their agreement (1 = strongly disagree, 5 = strongly agree) with 11 items concerning their support for the rectifying measures related to consequences of online anti-drug and drug-encouraging messages. Results of a principal component factor analysis with varimax rotation showed these items were clearly grouped in three distinct factors, accounting for 76.27% of the variance. The first factor, which explained 32.69% of the variance (Eigenvalue = 3.60), contained 5 items: supporting more attention by society to harmful online messages about drug abuse, supporting more government investigation about how to prevent youth from being affected by harmful online drug-encouraging information, supporting censorship of Internet on harmful messages related to drug abuse, supporting the police to act to prevent youth from obtaining drug information from the Internet, and supporting the police to act against drug selling through the Internet. The 5 items were added and divided by five to form an index of “support for corrective measures” (M = 3.89, SD = 0.82, α = .91). The second factor consisted of 3 items: supporting the government to put more resources into the anti-youth-drug-abuse campaign, supporting schools to organize more anti-drug activities, and supporting organization by society of more anti-drug activities for youth. This factor explained 21.97% of the variance (Eigenvalue = 2.42). A measure of “support for promotional measures” was created by adding the 3 items and dividing by three (M = 3.69, SD = 0.86, α = .85). The third factor contained 3 items: supporting more legal actions from the police to fight against youth drug abuse, supporting more legal punishments against drug selling to youth, supporting more punishments in schools to fight against drug abuse on campus. The third factor accounted for 21.61% of the total variance (Eigenvalue = 2.38). The 3 items were added and then divided by three to form a measure of “support for restrictive measures” (M = 3.82, SD = 0.91, α = .84).

Control variables

Demographic variables include gender, year in school, and academic performance. As the messages of anti-youth-drug-abuse campaign were distributed through the Internet, television, and radio, respondents were asked to report their use of these three media. In addition, respondents were asked to estimate how frequently they had been in contact with the drug-encouraging messages (M = 1.82, SD = 0.91) and the anti-drug campaign messages on the Internet (M = 1.99, SD = 0.91), ranging from 1 (never) to 5 (very frequently). Although most respondents reported rather low exposure to the online messages related to drugs, when asked about their perception of others’ exposure to the messages, 76.7% perceived others to have read drug-encouraging messages and 82.5% perceived others to have read anti-drug campaign messages on the Internet. Although respondents might have not been in direct contact with these online messages, reports on mass media have allowed them to get certain ideas of the presence of the messages on the Internet. For instance, after the Hong Kong government announced the implementation of the Voluntary School Drug Testing Scheme in the Tai Po District, there were news reports about foreign websites such as Spectrum Labs selling substances that helped students avoid a positive result in drug testing by urine (Nextmedia, 2009). Thus, participants were aware of the messages, which reflect social concern about drug messages on the Internet.

Findings

Descriptive Statistics

Among the respondents, 508 (65.3%) were male and 270 (34.7%) were female. Year in school of respondents were almost evenly distributed between Secondary 1 and Secondary 7, and majority of the respondents (95.6%) aged between 12 and 18. In all, 245 (33.1%) reported their academic ranking in the first quartile, 207 (26.4%) in the second, 199 (26.9%) in the third, and 90 (12.1%) in the fourth quartile.

A large number of the respondents (n = 338, 43.2%) reported daily television use as between 1 and 3 hours. For radio, 39.6% (n = 309) reported listening less than ½ hour per day. Regarding daily Internet use, 37.8% (n = 295) reported between 1 and 3 hours of use every day; 20.4% (n = 159) reported between 3 and 5 hours and 19.8% (n = 155) between ½ and 1 hour.

Hypotheses Testing

H1 predicted the TPE of the antisocial online drug-encouraging messages, that is, perceived effects of the antisocial messages would be greater on others than on the self. A paired-sample t test was performed. The results of the paired-sample t test showed, t(777) = 11.92 at p ≤ .001, that there is a significant difference between the perceived effects of the online antisocial messages on self (M = 2.29, SD = 1.10) and on others (M = 2.74, SD = 0.94). H1 is thus supported.

H2 predicted that the magnitude of the gap in perceived effect of online prosocial anti-drug messages on others and on the self would be smaller than that of antisocial messages. Another paired-sample t test was performed. The result of the paired-sample t test (Table 1) showed, t(777) = −3.78 at p ≤ .001, that the perceptual gap of the prosocial messages (M = 0.12, SD = 0.91) was significantly smaller than that of antisocial messages (M = 0.45, SD = 1.06). H2 is also supported.

Table 1.

Mean Scores of Perceived Effect and Results of t Tests (N = 778).

	M score
	Self	Others	Perceptual gap	t value (others–self)
Antisocial messages	2.29 (1.10)	2.74 (0.94)	0.45 (1.06)	11.92***
Prosocial messages	2.77 (1.04)	2.89 (0.94)	0.12 (0.91)	3.89***

Note. Figures in parentheses are standard deviations.

p ≤ .05. **p ≤ .01. ***p ≤ .001.

Perceptual Gap in Effects of Antisocial Messages

To examine the perceptual gap of antisocial messages’ predictive power to behavioral intentions, three hierarchical regression analyses were performed. The first block of predictors was demographic variables, including gender, year in school, and academic performance. Age was not entered to avoid multicollinearity because of its high correlation with year in school. The second block of predictors was general media use, including daily television use, radio use, and Internet use. The third block involved exposure to the online antisocial and prosocial messages regarding youth drug abuse. Perceptual gaps in effects of both antisocial and prosocial messages were entered as the fourth block of predictors. For all variables, the variance inflation factor (VIF) were smaller than 2 and tolerances were above .80, so multicollinearity was not a problem for the regression equation.

H3, H4, and H5 predicted that the self–other perceptual gap of antisocial messages would be a significant predictor of support for restrictive, corrective, and promotional measures, respectively. For H3 (column 1, Table 2), the overall regression equation was significant, F(8, 770) = 4.19, p < .001, adjusted R² = .06. The self–other perceptual gap of antisocial messages was significantly and positively related to support for restrictive measures (β = .17, p < .001). H3 is supported.

Table 2.

Hierarchical Regression Analyses of Demographics, General Media Use, Exposure, Perceptual Gap of Antisocial and Prosocial Messages, and Support for Rectifying Measures (N = 778).

	1	2	3
Regressions	Restrictive measures	Corrective measures	Promotional measures
Predictors
Block 1: Demographics
Gender	−.02	−.01	−.02
Year in school	−.01	.03	−.08*
Academic performance	−.10**	−.07*	.01
Adjusted R²	.01	.01	.01
Block 2: General media use
TV use	.01	.05	.04
Radio use	.09*	.09*	.04
Internet use	−.06	−.15***	−.14***
Incremental adjusted R²	.01	.03	.01
Block 3: Exposure of self
Antisocial messages	−.03	−.07	−.07
Prosocial messages	.10*	.13**	.21***
Incremental adjusted R²	.01	.02	.05
Block 4: Perceptual gap
Antisocial messages	.17***	.22***	.10**
Prosocial messages	−.05	−.01	−.07
Incremental adjusted R²	.03	.04	.01
Total adjusted R²	.06	.10	.08

Note. Standardized beta weights from final regression equation with all blocks of variables in the model. Variables coded, or recoded, as follows: gender (0 = female, 1 = male); year in school ranged from 1 (Secondary 1) to 7 (Secondary 7); and academic performance ranged from 1 (first quarter in the year) to 4 (fourth quarter in the year); exposure to antisocial/prosocial messages (1 = never, 5 = very frequently); perceptual gap of effects between self and others (1 = no effect at all, 5 = a strong effect); support for rectifying measures (1 = strongly disagree, 5 = strongly agree).

p ≤ .05. **p ≤ .01. ***p ≤ .001.

For H4 (column 2, Table 2), the overall regression equation was significant, F(7, 770) = 4.89, p < .001, adjusted R² = .10. The perceptual gap in effects of antisocial messages on the Internet was a significant predictor of support for corrective measures (β = .22, p < .001). H4 is supported.

For H5 (column 3, Table 2), the overall regression equation was significant, F(7, 770) = 6.88, p < .001, adjusted R² = .08. Again, the perceptual gap in effect of antisocial messages was a significant predictor of support for promotional measures (β = .10, p < .001). H5 is also supported.

Perceived Effects of Prosocial Messages

From the three regression analyses, we compared the predictive power of the perceptual gap for prosocial messages and antisocial messages with the three types of rectifying measures for H6, H7, and H8.

For H6 (column 1, Table 2), although the overall regression equation was significant, F(8, 770) = 4.19, p < .001, adjusted R² = .06, the perceptual gap of prosocial messages was not a significant predictor of support for restrictive measures (β = −.05, p > .05). An additional t test for the difference between two regression coefficients (Cohen & Cohen, 1983) showed that the standardized beta size of perceptual gap of antisocial messages (β = .17, p < .001) was greater than perceptual gap of prosocial messages on support for restrictive measures (t = 4.23, p < .001). H6 is supported.

For H7 (column 2, Table 2), the overall regression equation was significant, F(7, 770) = 4.89, p < .001, adjusted R² = .10. The results of the analysis show that the perceptual gap was not significantly related to support for corrective measures (β = −.01, p > .05). The standardized beta size of perceptual gap of antisocial messages (β = .22, p < .001) was also greater than perceptual gap of prosocial messages on support for collective measures (t = 4.51, p < .001). H7 is supported.

For H8 (column 3, Table 2), the overall regression equation was significant, F(7, 770) = 6.88, p < .001, adjusted R² = .08. Once again, the perceptual gap failed to be a significant predictor of support for promotional measures (β = −.07, p > .05). Additional tests showed that the standardized beta size of perceptual gap of antisocial messages (β = .10, p < .01) was greater than perceptual gap of prosocial messages on support for restrictive measures (t = 3.31, p < .001). H8 is also supported.

To gain further insight into relationships among perceived effects on the self, perceived effects on others, and support for rectifying measures, six hierarchical regression analyses were performed. Tables 3 and 4 present the results of the six regression analyses. As shown in Table 3, the results of the first regression analysis indicated (column 1) that perceived effect of antisocial messages on others was a significant and positive predictor of support for restrictive measures (β = .17, p < .001). However, perceived effect on the self was a significant but negative predictor of support for restrictive measures (β = −.17, p < .001). The results of the second regression analysis show (column 2, Table 3) that perceived effect of antisocial messages on others was positively related to support for corrective measures (β = .22, p < .001), whereas perceived effect of the self was negatively related to the dependent variable (β = −.23, p < .001). In addition, the results of the third regression analysis showed (column 3, Table 3) that perceived effects on the others were a significant and positive predictor (β = .18, p < .001), but perceived effect of antisocial messages on self was a insignificant predictor of support for promotional measures (β = −.06, p > .05).

Table 3.

Hierarchical Regression Analyses of Demographics, General Media Use, Exposure, Perceived Effects on the Self and on Others of Antisocial Messages and Support for Rectifying Measures (N = 778).

	1	2	3
Regressions	Restrictive measures	Corrective measures	Promotional measures
Predictors
Block 1: Demographics
Gender	−.02	−.00	−.02
Year in school	−.01	.03	−.10*
Academic performance	−.10**	−.07	−.01
Adjusted R²	.01	.01	.01
Block 2: General media use
TV use	.01	.05	.04
Radio use	.10**	.11**	.07
Internet use	−.05	−.14***	−.10*
Incremental adjusted R²	.01	.03	.01
Block 3: Exposure of self
Antisocial messages	.00	−.03	−.03
Incremental adjusted R²	.00	.00	.00
Block 4: Perceived effects
Self	−.17***	−.22***	−.06
Others	.17***	.23***	.18***
Incremental adjusted R²	.03	.05	.02
Total adjusted R²	.05	.09	.04

Note. Standardized beta weights from final regression equation with all blocks of variables in the model. Variables coded, or recoded, as follows: gender (0 = female, 1 = male); year in school ranged from 1 (secondary 1) to 7 (secondary 7); and academic performance ranged from 1 (first quarter in the year) to 4 (fourth quarter in the year); exposure to antisocial messages (1 = never, 5 = very frequently); perceived effects on the self and on others (1 = no effect at all, 5 = a strong effect); support for rectifying measures (1 = strongly disagree, 5 = strongly agree).

p ≤ .05. **p ≤ .01. ***p ≤ .001.

Table 4.

Hierarchical Regression Analyses of Demographics, General Media Use, Exposure, Perceived Effects on the Self and on Others of Prosocial Messages and Support for Rectifying Measures (N = 778).

	1	2	3
Regressions	Restrictive measures	Corrective measures	Promotional measures
Predictors
Block 1 : Demographics
Gender	−.01	.00	−.02
Year in school	.04	.08*	−.03
Academic performance	−.13***	−.11***	−.02
Adjusted R²	.01	.01	.01
Block 2: General media use
TV use	.01	.05	.04
Radio use	.07*	.08*	.02
Internet use	−.05	−.13***	−.13***
Incremental adjusted R²	.01	.02	.01
Block 3: Exposure of self
Antisocial messages	−.01	−.00	.06
Incremental adjusted R²	.01	.01	.04
Block 4: Perceived effects
Self	.21***	.18***	.26***
Others	.12**	.16***	.12**
Incremental adjusted R²	.07	.07	.09
Total adjusted R²	.10	.11	.15

Note. Standardized beta weights from final regression equation with all blocks of variables in the model. Variables coded, or recoded, as follows: gender (0 = female, 1 = male); year in school ranged from 1 (secondary 1) to 7 (secondary 7); and academic performance ranged from 1 (first quarter in the year) to 4 (fourth quarter in the year); exposure to prosocial messages (1 = never, 5 = very frequently); perceived effects on the self and on others (1 = no effect at all, 5 = a strong effect); support for rectifying measures (1 = strongly disagree, 5 = strongly agree).

p ≤ .05. **p ≤ .01. ***p ≤ .001.

As shown in Table 4, the three regression analysis showed that both perceived effect of the prosocial messages on the self and on others were significantly and positively related to support for restrictive, corrective, and promotional measures.

Taken together, the results of the regression analyses explained why the perceptual gap in effects of prosocial anti-drug messages on self and on others was not a significant predictor of support for any type of rectifying measures. As both perceived effects on the self and perceived effects on the others positively predicted support for the rectifying measures, the perceptual gap of prosocial messages became less predictive than antisocial messages did. The self–other perceptual gap in effects of antisocial messages was a better predictor of support for rectifying measures than that of prosocial messages.

Discussions and Conclusions

The main objective of this study was to examine the perceived effects of antisocial and prosocial messages related to youth drug abuse on the Internet. The tendency to protect others from being affected by harmful messages not only occurred among the adults sampled in most past studies, but also the adolescents in our survey. Therefore, our findings suggest that self-enhancement is a potential cause in all ages for the tendency to protect others from antisocial messages. Practically, schools and the government should emphasis more on educating students about the possible harmful effects of online drug-encouraging messages among their peers, which might boost the students’ perception of the negative effects of drug-encouraging messages on others and help promote support for restrictive, corrective, and promotional anti-drug measures.

Regression analyses showed that perceived effects of the drug-encouraging messages on the self was a significant predictor of restrictive and corrective measures. One possible explanation is that a self-enhancement bias for maintaining self-image leads to people’s perception of others being more vulnerable than themselves, so they support rectifying measures to protect others from the harmful effects. Jang and Johnson (2011) found that association with drug-using peers was a predictor of drug use in adolescents. With the Internet, adolescents can have drug-using peer associations much more easily. Further anti-drug activities should focus more on those students who report greater effects of drug-encouraging messages on themselves, because these students tend to be less supportive of restrictive and corrective anti-drug measures.

Perceived effect of drug-encouraging messages on others was a significant positive predictor of support for the usage of all three of the rectifying measures: restrictive, corrective, and promotional, validating the model of presumed influence of others (Gunther & Storey, 2003). Following the self-enhancement explanation, people were likely to consider antisocial media messages as more effective on others, and therefore they saw a greater need to protect those exposed to these messages. This study supplements literature on the TPE by showing that censorship is not the only possible solution when people perceive the need to protect others from the effects of antisocial messages. In their responses to the survey, the respondents supported any rectifying measures which offered to reduce the harmful effects of the messages.

We found that perceived effects of the prosocial messages on others predicted restrictive, corrective, and promotional measures positively. Perception of more effective anti-drug campaign messages on others implies recognition of the campaign’s effectiveness as a whole. They may see other anti-drug measures can be as effective as the campaign messages. Perceived effects of the prosocial anti-drug messages on the self also significantly predicted greater support for restrictive, corrective, and promotional measures. In Cho and Boster’s (2008) study on anti-drug ads, perceived effects on the self were positively associated with adolescents’ anti-drug attitudes. Our results further show that perceived effects of anti-drug campaign messages on the self were associated with three types of rectifying behaviors. Findings about the importance of perceived effects on the self add to the model of presumed influence, which states that behavioral change is mainly predicted by presumed influence of media messages on others (Gunther & Storey, 2003). Based on the self-enhancement explanation, as the messages were socially desirable, people did not need to defend their self-images by estimating smaller effects of the messages on the self. We call for more investigation on the perceived effect on the self to give evidence on the underlying theoretical explanation.

The self–other perceptual gap in effects of the antisocial messages worked significantly in predicting support for the three types of rectifying measures. The greater the self–other difference in the perceived effects, the more likely one would be to support rectifying measures. This supported previous research that the third-person perceptual gap of antisocial messages predicted support for censoring and restricting such content to protect the others whom they perceived as more vulnerable (Gunther, 1995; Hoffner & Buchanan, 2002; Hoffner et al., 1999; Lo & Paddon, 2001; McLeod, Detenber, & Eveland, 2001; McLeod, Eveland, & Nathanson, 1997; Rojas et al., 1996; Wei & Lo, 2007; Wu & Koo, 2001). A greater third-person perceptual gap showed that they perceived others to be more easily affected by the antisocial drug-encouraging messages than themselves, so they supported more rectifying measures to protect them against the harmful drug-encouraging effects.

However, the third-person perceptual gap of prosocial anti-drug messages failed to significantly predict support for any of the measures. Results do not support the argument that the perceptual gap in effects of prosocial messages could provide behavioral implications (Golan & Day, 2008). Compared with antisocial messages, the reduced perceptual gap of prosocial messages was too small to be significant in predicting behaviors. Instead of the third-person perceptual gap, perceived effects on the self and on others were more important variables of the perceptual component and have shown significant relationships with the behavioral component for the TPE.

Exposure to the messages, media use, and demographics were examined as control predictors in regression analyses. Exposure to the prosocial anti-drug messages was significant in predicting rectifying measures, especially promotional measures. This implies that running more anti-drug promotions on the Internet can help raise youth support for anti-drug measures. In contrast, exposure to antisocial messages was not a significant predictor of behavioral intentions. This adds to literature findings showing that the correlation between exposure and the perceived effects of media messages did not hold (McLeod et al., 2001; Meirick, 2005). In light of this finding, educating students about the harmful effects of drug-encouraging messages is more important than preventing exposure to the Internet messages.

Daily radio use was found to be a significant predictor of support for restrictive and corrective rectifying measures. It is possible that those who listened to more radio—a traditional medium—might be more suspicious about the Internet or the new media, thus supporting more restrictive measures toward online drug-related messages. Television use was not a significant predictor of support for any type of rectifying measure. However, daily Internet use significantly predicted support for corrective and promotional measures. Heavier Internet users may be prone to more harmful messages related to drug abuse, and they tend to ignore these positive promotions, so it is necessary for practitioners to further investigate reaching heavier Internet users in future anti-drug policies. Among the three types of media, respondents reported the highest daily use of the Internet, followed by daily TV and radio. Despite the popularity of the Internet among young people, the majority of the respondents only reported infrequent exposure to anti-drug messages on the Internet. More anti-drug resources should go to Internet promotions instead of traditional media.

Among demographic variables, only academic performance in school constantly appeared as a significant predictor of support for rectifying measures. More efforts and anti-drug actions need to target groups with lower academic performance. Anti-drug education should be provided across all gender and age groups in school.

This study has several limitations. Although stratified random sampling was done within the sampling frame, the sampling process was not completely random, because only two secondary schools consented to take part in this study. (However, the two schools were representatives of two different types of schools. One was a co-ed school and the other was a boys’ school.) Moreover, the study did not measure actual behaviors. In spite of the label, “behavioral consequences,” what we really have measured is attitudes about behaviors. Although attitudes are a significant predictor of behaviors, future research should measure rectifying behaviors that follow from individual or public policy initiatives.

No more than 20% of variances in support for restrictive, corrective, and promotional measures are explained in the regression analyses. In particular, the perceptual gaps and perceived effects of antisocial messages predicted less than 10% of the dependent variables. This indicates that there were other important predictor(s) accounting for respondents’ support for rectifying measures. One reason for the low explained variance is that the online messages examined in this study are only a small part of the drug-related messages that youths receive in their everyday lives. Respondents can receive drug-related messages from many other sources, including anti-drug campaign activities and interpersonal sources such as family and friends. Messages from these sources can play an important role in young people’s anti-drug attitudes. Therefore, taking into account the effects of off-line drug-related sources in the regression analyses might provide better prediction of support for the rectifying measures.

Although the survey was conducted in late 2010, the data still have important implications for today’s anti-drug policies. According to the World Development Indicators provided by the World Bank (2014), the percentage of Internet users in Hong Kong increased only slightly from 72.0% in 2010 to 74.2% in 2013. The Public Opinion Survey on 2013 Anti-drug Publicity Measures (Narcotics Division of the Security Bureau HKSAR, 2013a) indicates that TV ads still ranked as the most common media channel for young people to receive anti-drug messages, and the Internet was not within the top five. Our findings that more work should be done to promote anti-drug messages on the Internet remain valid today. We conducted our survey as the 2-year anti-drug-abuse campaign “No Drug, No Regrets; Not Now, Not Ever” came to an end, a suitable time to evaluate the effects of the campaign messages. The current “Stand Firm! Knock Drugs Out” anti-drug campaign was launched in 2010 and is ongoing. When this campaign ends, further studies evaluating the perceived effects of the campaign in comparison with the results of this study will be very useful for future anti-drug campaign design.

To conclude, this study adds to the existing literature about the relationship between the TPE and behavioral consequences. All three types of rectifying measures are possible behavioral consequences of the perceived effects, for both antisocial and prosocial messages. The importance of messages’ social desirability in the behavioral component of the TPE was demonstrated. In addition, government and educational institutions should be aware of the Internet’s influence on anti-drug policies.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Wan Chi Leung is a doctoral candidate and instructor at University of South Carolina. Her research interests include media effects, health communication, entertainment, fan communities, and network analysis.

Ven-Hwei Lo, PhD, University of Missouri, 1985, is a professor in the School of Journalism & Communication, The Chinese University of Hong Kong. His research interests include media effects, news media performance, political communication, and social impact of the Internet.

References

The Boys’ and Girls’ Clubs Association of Hong Kong. (2010). 「網絡毒品資訊」調查(Internet Drug Research). Retrieved from http://www.bgca.org.hk/bgca06/research/2010/poison/Int_Drug_Research_PPT_20100307.pdf

Brosius

H. B.

Engel

(1996). The causes of third-person effects: Unrealistic optimism, impersonal impact, or generalized negative attitudes towards media influence? International Journal of Public Opinion Research, 8, 142-162.

Chapin

J. R.

(2000). Third-person perception, message type, safer-sex campaigns, and sexual risk-taking among minority “at risk” youth. Mass Communication and Society, 2, 163-173.

Cho

Boster

F. J.

(2008). First and third person perceptions on anti-drug ads among adolescents. Communication Research, 35, 169-189.

Cohen

(1983). Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum.

David

Liu

Myser

(2004). Methodological artifact or persistent bias?: Testing the robustness of the third-person and reverse third-person effects for alcohol messages. Communication Research, 31, 206-233.

Davison

W. P.

(1983). The third-person effect in communication. The Public Opinion Quarterly, 47, 1-15.

Duck

J. M.

Mullin

(1995). The perceived impact of the mass media: Reconsidering the third person effect. European Journal of Social Psychology, 25, 77-93.

Duck

J. M.

Terry

D. J.

Hogg

M. A.

(1995). The perceived influence of AIDS advertising: Third-person effects in the context of positive media content. Basic and Applied Social Psychology, 17, 305-325.

10.

Educational Bureau. (2014). Health school policy. Retrieved from http://www.edb.gov.hk/en/edu-system/primary-secondary/healthy-sch-policy/background.html

11.

Eveland

W. P.

McLeod

D. M.

(1999). The effect of social desirability on perceived media impact: Implications for third-person perceptions. International Journal of Public Opinion Research, 11, 315-333.

12.

Golan

G. J.

Day

A. G.

(2008). The first-person effect and its behavioral consequences: A new trend in the twenty-five year history of third-person effect research. Mass Communication and Society, 11, 539-556.

13.

Gunther

A. C.

(1995). Overrating the X-rating: The third-person perception and support for censorship of pornography. Journal of Communication, 45(1), 27-38.

14.

Gunther

A. C.

Mundy

(1993). Biased optimism and the third-person effect. Journalism Quarterly, 70, 58-67.

15.

Gunther

A. C.

Storey

(2003). The influence of presumed influence. Journal of Communication, 53, 199-215.

16.

Gunther

A. C.

Thorson

(1992). Perceived persuasive effects of product commercials and public service announcements third-person effects in new domains. Communication Research, 19, 574-596.

17.

Henriksen

Flora

J. A.

(1999). Third-person perception and children: Perceived impact of pro-and anti-smoking ads. Communication Research, 26, 643-665.

18.

Hoffner

Buchanan

(2002). Parent’s responses to television violence: The third person perception, parental mediation and support for censorship. Media Psychology, 4, 231-252.

19.

Hoffner

Buchanan

Anderson

J. D.

Hubbs

L. A.

Kamigaki

S. K.

Kowalczyk

Pastorek

Plotkin

R. S.

Silberg

K. J.

(1999). Support for censorship of television violence: The role of the third-person effect and news exposure. Communication Research, 26, 726-742.

20.

Hong Kong’s Information Service Department. (2014). 致力打擊隱蔽吸毒(Fighting against hidden drug abuse). Retrieved from http://www.news.gov.hk/tc/record/html/2014/06/20140628_171-525.shtml

21.

Hoorens

Ruiter

(1996). The optimal impact phenomenon: Beyond the third person effect. European Journal of Social Psychology, 26, 599-610.

22.

Jang

S. J.

Johnson

B. R.

(2011). The effects of childhood exposure to drug users and religion on drug use in adolescence and young adulthood. Youth & Society, 43, 1220-1245.

23.

V.-H.

Paddon

A. R.

(2001). Third-person effect, gender differences, pornography exposure and support for restriction of pornography. Asian Journal of Communication, 11(1), 120-142.

24.

V.-H.

Wei

(2002). Third-person effect, gender, and pornography on the Internet. Journal of Broadcasting & Electronic Media, 46, 13-33.

25.

McLeod

D. M.

Detenber

B. H.

Eveland

W. P.

(2001). Behind the third-person effect: Differentiating perceptual processes for self and other. Journal of Communication, 51, 678-695.

26.

McLeod

D. M.

Eveland

W. P.

Nathanson

A. I.

(1997). Support for censorship of violent and misogynic rap lyrics: An analysis of the third-person effect. Communication Research, 24, 153-174.

27.

Meirick

P. C.

(2005). Rethinking the Target corollary the effects of social distance, perceived exposure, and perceived predispositions on first-person and third-person perceptions. Communication Research, 32(6), 822-843.

28.

Meirick

P. C.

(2008). Targeted audiences in anti-drug ads: Message cues, perceived exposure, perceived effects, and support for funding. Media Psychology, 11, 283-309.

29.

Mutz

D. C.

(1989). The influence of perceptions on media influence: Third-person effect and the public expression of opinions. International Journal of Public Opinion Research, 1, 3-23.

30.

Narcotics Division of the Security Bureau HKSAR. (2009). The 2008/09 drug use among students report. Retrieved from http://www.nd.gov.hk/pdf/survey_drug_use/2008-2009/Report.pdf

31.

Narcotics Division of the Security Bureau HKSAR. (2011). Hong Kong Narcotics and Anti-Money Laundering Report 2007-2010. Retrieved from http://www.nd.gov.hk/pdf/report2007-10/full_report.pdf

32.

Narcotics Division of the Security Bureau HKSAR. (2013a). Public Opinion Survey on 2013 Anti-drug publicity measures. Retrieved from http://www.nd.gov.hk/pdf/pos2013_major_findings.pdf

33.

Narcotics Division of the Security Bureau HKSAR. (2013b). The 2011/12 drug use among students report. Retrieved from http://www.nd.gov.hk/pdf/survey_drug_use/2011-2012/Report.pdf

34.

Nextmedia. (2009, September 3). 假尿破驗毒. Next Magazine, p. 1017. Retrieved from http://hk.next.nextmedia.com/article/1017/13165870

35.

Paul

Salwen

M. B.

Dupagne

(2000). The third-person effect: A meta-analysis of the perceptual hypothesis. Mass Communication and Society, 3, 57-85.

36.

Perloff

R. M.

(1993). Third-person effect research 1983-1992: A review and synthesis. International Journal of Public Opinion Research, 5, 167-184.

37.

Perloff

R. M.

(1999). The third-person effect: A critical review and synthesis. Media Psychology, 1, 353-378.

38.

Perloff

R. M.

(2002). The third-person effect. In Media effects: Advances in theory and research (Vol. 2, pp. 489-506). New York, NY: Routledge.

39.

Perse

E. M.

Rubin

R. B.

Rubin

Graham

E. E.

Seibold

D. R.

(2009). Communication research measures II: A sourcebook. New York, NY: Routledge.

40.

Rojas

(2010). “Corrective” actions in the public sphere: How perceptions of media effects shape political behaviors. International Journal of Public Opinion Research, 22, 343-363.

41.

Rojas

Shah

D. V.

Faber

R. F.

(1996). For the good of others: Censorship and the third-person effect. International Journal of Public Opinion Research, 8, 163-187.

42.

Sun

Pan

Shen

(2008). Understanding the third-person perception: Evidence from a meta-analysis. Journal of Communication, 58, 280-300.

43.

Sun

Shen

Pan

(2008). On the behavioral component of the third-person effect. Communication Research, 35, 257-278.

44.

Tal-Or

Cohen

Tsfati

Gunther

A. C.

(2010). Testing causal direction in the influence of presumed media influence. Communication Research, 37, 801-824.

45.

Tewksbury

Moy

Weis

D. S.

(2004). Preparations for Y2K: Revisiting the behavioral component of the third-person effect. Journal of Communication, 54, 138-155.

46.

Wei

V.-H.

(2007). Reconsidering the relationship between the third-person perception and optimistic bias. Communication Research, 34, 665-684.

47.

World Bank. (2014). Internet users (per 100 people). Retrieved from http://data.worldbank.org/indicator/IT.NET.USER.P2

48.

Koo

S. H.

(2001). Perceived effects of sexually explicit Internet content: The third-person effect in Singapore. Journalism and Mass Communication Quarterly, 78, 260-274.

49.

Gonzenbach

W. J.

(2008). Does a perceptual discrepancy lead to action? A meta-analysis of the behavioral component of the third-person effect. International Journal of Public Opinion Research, 20, 375-385.

50.

Zhang

(2010). Self-enhancement on a self-categorization leash: Evidence for a dual-process model of first-and third-person perceptions. Human Communication Research, 36, 190-215.

51.

Zhao

Cai

(2008). From self-enhancement to supporting censorship: The third-person effect process in the case of Internet pornography. Mass Communication and Society, 11, 437-462.

52.

Zhong

(2009). Third-person perceptions and online games: A comparison of perceived antisocial and prosocial game effects. Journal of Computer-Mediated Communication, 14, 286-306.