Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 May 1.
Published in final edited form as: Health Psychol. 2015 Nov 2;35(5):454–464. doi: 10.1037/hea0000273

A Randomized Trial Evaluating Child Dog-Bite Prevention in Rural China through Video-Based Testimonials

Jiabin Shen 1, Shulan Pang 2, David C Schwebel 1
PMCID: PMC4833565  NIHMSID: NIHMS713059  PMID: 26523353

Abstract

Objective

Dog-bite injuries pose significant threat to children globally. School-aged children are especially at risk because of their cognitively immature tendency toward low perceived vulnerability to bites, and this risk is elevated further for school-aged children living in rural China due to the large number of stray dogs, all potential rabies carriers, present in their communities.

Methods

This randomized controlled trial evaluated whether viewing an educational video of testimonials would change safety knowledge, perceived vulnerability, and simulated behaviors with dogs among a sample of 280 third and fourth graders living in rural China. Participants were randomly assigned to view either an educational video of testimonials on dog-bite prevention (treatment) or an educational video of testimonials on drowning prevention (comparison). Safety knowledge, perceived vulnerability to dog bites, and simulated behavior with dogs using a dollhouse model were assessed both before and after exposure to the video of testimonials.

Results

Children who watched the educational video of testimonials on dog-bite prevention had increased safety knowledge, higher perceived vulnerability, and less risky simulated behaviors with dogs compared to the comparison group. Mediation analysis revealed that the intervention successfully changed children's simulated behaviors with dogs through greater safety knowledge and increased perceived vulnerability.

Conclusions

Results suggest the incorporation of testimonials into injury prevention programs has potential for broad global dissemination. The fact that both increased knowledge and heightened perceived vulnerability mediated changes in simulated behavior suggests the dual roles of knowledge and appraisal on children's injury-risk behavior.

Keywords: dog-bite, injury, prevention, testimonial, child

Over 100 million school-aged children ages 7-14 live in rural China. They have substantial risk for many adverse health effects (Wang et al., 2008), one of which is pediatric dog bite injuries. Researchers report that 30% of rural Chinese elementary school children ages 9-13 have experienced at least one dog-bite injury in their lives (Shen et al., 2013a). Among those children, 54% of bites occur on public streets away from home, and 84% of bites occur without adults nearby (Shen et al., 2014). Dog bites result in both adverse physical and psychological consequences. Physical consequences can include acute and long-term medical treatment, painful and expensive rabies treatment, and physical restriction from routine activities and schoolwork (Eppley & Schleich, 2013; Shen et al., 2014). Psychologically, consequences can include fear, withdrawal, post-traumatic symptoms, and coping with permanent scarring, including on the face (Hon et al., 2007; Peters, Sottiaux, Appelboom, & Kahn, 2004). One recent study of 101 pediatric dog-bite cases in rural China found that pediatric dog-bite victims often developed behavioral withdrawal and psychological fear of dogs (Shen et al., 2014).

Children have increased risk for dog bites compared to adults due to both smaller physical size and underdeveloped cognitive skills. Smaller size increases both risk and severity of bites, and children are especially vulnerable to bites on the neck or head (Hon et al., 2007; Mcheik, Vergnes, & Bondonny, 2000). Cognitively, children are more likely to engage in risky interactions with dogs and to behave impulsively and without accurate perception of their vulnerability to injury by dogs (Avner & Baker, 1991; Chouinard, 2007).

The unique environment of rural China further heightens risk of child dog-bite injury. Many dogs run loose in rural Chinese villages, and children are exposed to dogs during long daily walks to and from school (Deng et al., 2004; Duan, 2008; Shen, Li, Xiang, Lu, & Schwebel, 2014). Risk from bites is amplified by the presence of rabies, which increased by 2000% among dogs in China from 1996 to 2006 and has not waned since (Si et al., 2008; Zhang, Jin, Sun, Sun, & Ruan, 2012). Furthermore, parental supervision of children in rural China is frequently insufficient because many young parents migrate to urban cities for higher-paid work (the China National Bureau of Statistics estimated 245 million migrants from rural to urban China in 2014), leaving children home with aging grandparents whose more limited mobility diminishes proper supervision of children, especially on public streets.

Efforts to prevent pediatric dog-bite injuries are few globally, and almost non-existent in rural China. In fact, a 2009 systematic review of pediatric dog-bite injury prevention programs uncovered just two studies that met inclusion criteria of randomized controlled trials or controlled before-after studies and evaluated the effectiveness of educational interventions for preventing dog bites with behavioral outcomes in populations under 20 years old (Dupperrex, Blackhall, Burri, Jeannot, 2009). Intervention programs that are reported in the literature focus primarily on teaching children basic knowledge and skills via one-way (adult to child) educational programs (Beaver, 2001; Chapman, et al., 2000; Dixon, Pomerantz, Hart, Lindsell, & Mahabee-Gittens, 2013; Spiegel, 2000), interactive classroom lesson plans (Szecsi, Barbero, Del Campo, & Toledo, 2010; Wilson, Dwyer, & Bennett, 2003) or animated computerized tasks (Meints & De Keuster, 2006, 2009; Schwebel et al., 2012). These studies yielded mixed results concerning transfer of knowledge, but very few have tested and successfully demonstrated a change in children's risky behaviors with dogs (Schwebel, 2012).

An alternative strategy to improve children's safety with dogs is to increase children's perceived vulnerability toward risk from child-dog interactions. The role of perceived vulnerability is emphasized in multiple prominent theoretical models to explain health-related behavioral change, including the Health Belief Model (Janz & Becker, 1984), and research with school-age children confirms the role of low perceived vulnerability in children's injury risk behaviors (Morrongiello & Rennie, 1998). Previous research examining dog-bite risk among school-aged children in rural China also reported that children's low perceived vulnerability was associated with both self-reported risky interactions with dogs and history of dog-bite injury (Shen et al., 2013a).

One effective strategy to alter individuals’ perceived vulnerability toward health-related behaviors is through the use of first-person testimonials that detail the antecedents, processes, and consequences of an adverse event (Slater & Rouner, 1996). Testimonial-based interventions might be especially persuasive in the context of training children to be safe with dogs, as testimonials concerning dog bite incidents can elicit strong emotional arousal among children, a strategy that may improve the potential for behavior change (De Wit, Das, & Vet, 2008; Loewenstein, Weber, Hsee, & Welch, 2001).

The current study used a repeated-measures randomized controlled trial design to evaluate the efficacy of a video-based testimonial intervention on reducing children's risks of dog-bite injury. We hypothesized that children who watched the dog-bite testimonials would demonstrate more safety knowledge, higher perceived vulnerability and less risky simulated behaviors with dogs than children in the comparison group (Hypothesis 1). Further, we hypothesized that both safety knowledge and perceived vulnerability would mediate the effect of intervention on children's simulated behaviors with dogs, with watching dog-bite testimonials leading to increased knowledge and increased perceived vulnerability, which would in turn lead to less risky simulated behaviors with dogs (Hypothesis 2).

METHODS

Participants

Power analysis was performed a priori and minimum sample size calculated based on the following two procedures. First, an estimated effect size of d = 0.33 for the intervention effect was determined based on results from a similar video-based intervention study of playground safety in Canada (Morrongiello & Matheis, 2007a). It was determined that 76 participants would be required to detect the expected effect size for Hypotheses 1 with a power of .99 (2 tails, alpha=.01; Cohen, 1988) using G*Power 3.1; our anticipated sample from any single school in rural China was much larger. Second, using the bias-corrected bootstrap method, it was determined that 120 participants were required to achieve power of .80 for the mediation analysis of Hypothesis 4 (Fritz & MacKinnon, 2007). Again, the anticipated sample from any single school in rural China was larger than this.

In the present study, 280 children in Grades 3 and 4 at an elementary school in rural Zhejiang Province, China, participated (mean age = 10.03 years, median age = 10.00, SD = 0.83; Interquartile range = 1.13; 48.9% boys; 92% Han Chinese ethnicity). Inclusion criteria were all students enrolled in either Grade 3 or 4 at the participating school at the time of study, and no exclusion criteria applied. The sample reported annual household income of less than US $8,000 (~Chinese RMB 48,000) and 85% of primary caregivers had fewer than 10 years of formal education.

Measures

Demographics

Demographic information was collected from both parents and children pre-intervention. Children reported their age, gender, ethnicity (Han vs. other), dog ownership, dog exposure frequency, and daily risky practice with dogs. Dog ownership was assessed with a single item, ‘Does your family keep dogs now?’, which was answered ‘Yes’ or ‘No’. Dog exposure frequency was also assessed with a single item, ‘How often do you usually interact with dogs?’, which children responded to using a range from 1 (I've never interacted with dogs) to 6 (at least once a day). Daily risky practice with dogs was assessed using an 8-item instrument. Items were answered on a 5-point scale and both internal reliability (Cronbach alpha = .73) and factorial validity (Confirmatory factor analysis supported unidimensionality of the measure, χ2/df = 3.81, p < .05, Standardized Root Mean Square Residual (SRMS) = .06, CFI = .85) were adequate. The final score was an average across items (possible range = 1-5).

Primary caregivers reported child dog-bite history (‘How many times has your child been bitten (scratched) by a dog?’; due to skewness, answers were dichotomized to bite history vs. none for analysis), household income and educational background. A socioeconomic status (SES) composite was computed as the mean of standardized scores from household income and the primary caregiver's education. Table 1 shows descriptive data.

Table 1.

Descriptive Data

Variable Overall (N=280) Intervention (N=143) Control (N=137) p
Age (years)a 10.03 (.83) 10.03 (.82) 10.02 (.84) .87
Gender (Male)b 135 (48.90) 67 (47.20) 68 (50.70) .55
Ethnicity (Han Chinese)b 256 (91.80) 133 (93.00) 123 (90.40) .44
Caregiver's Educationb
    No Formal School Education 19 (6.90) 10 (7.00) 9 (6.70)
    =<Grade 6 97 (35.00) 44 (31.00) 53 (39.30)
    =< Grade 9 120 (43.30) 66 (46.50) 54 (40.00) .63
    =< Grade 12 36 (13.00) 20 (14.10) 16 (11.90)
    >= College 5 (1.80) 2 (1.40) 3 (2.20)
Annual Household Incomeb
    <=48,000RMB (~$8,000) 149 (54.40) 75 (53.20) 74 (55.60) .44
Having Dog-bite Historyb 55 (20.00) 28 (19.90) 27 (20.10) .95
Dog Exposureb
    Never 61 (21.80) 33 (23.10) 28 (20.40)
    Once every few years 37 (13.20) 25 (17.50) 12 (8.80)
    >=once a year, < once a month 17 (6.10) 14 (9.80) 3 (2.20) .01*
    >=once a month, <once a week 23 (8.20) 10 (7.00) 13 (9.50)
    >=once a week, < once a day 46 (16.40) 17 (11.90) 29 (21.20)
    At least once a day 96 (34.30) 44 (30.80) 52 (38.00)
Owning dog(s) at homeb 116 (41.40) 57 (39.90) 59 (43.10) .59
Daily Risky Practicea (1-5 scale) 1.80 (.60) 1.77 (.58) 1.83 (.61) .38
Pre-Knowledgea (% correct, 0-100) 48.01 (14.21) 47.76 (13.86) 48.26 (14.60) .77
Pre-Perceived Vulnerabilitya (1-5 scale) 3.79 (.61) 3.81 (.60) 3.77 (.62) .62
Pre-Behavior in Dollhouse Taska (0-1 scale) .33 (.19) .32 (.19) .34 (.19) .33
Post-Knowledgea (% correct, 0-100) 60.89 (16.09) 66.78 (14.81) 54.74 (15.09) <.001
Post-Perceived Vulnerabilitya (1-5 scale) 4.38 (.57) 4.57 (.45) 4.19 (.62) <.001
Post-Behavior in Dollhouse Taska (0-1 scale) .27 (.18) .20 (.16) .33 (.19) <.001
Open in a new tab
a

Mean (Standard Deviation) is reported in the first three columns, and p-value for independent samples t test in the fourth.

b

N (Valid Percent) is reported in the first three columns, and p-value for chi-square test in the fourth.

Child knowledge about dog safety

The child-dog interaction safety knowledge questionnaire is a 24-item measure of children's safety knowledge around dogs. We developed this questionnaire for use in pilot research with similar age groups of children in rural China and it demonstrated adequate external validity (significant correlation with risk perception, r = −.43, p < .01; and risky interactions with dogs, r = −.23, p < .01) (Shen et al., 2013a). In the present study, the measure had inadequate psychometric properties, so WLSMV-based factor analysis was conducted and revealed 6 items that did not load well onto the primary construct of interest. Thus, a modified 18-item instrument was analyzed. It demonstrated adequate test-retest reliability (r = .58, p < .001) and factorial validity (confirmatory factor analysis supported unidimensionality of the measure (χ2/df = 1.65, p < .05; SRMR=0.12; CFI=0.81). Given the dichotomous nature of the scale items, ordinal coefficient theta was computed as a measure of internal reliability (Fabrigar, Wegener, MacCallum, & Strahan, 1999; Zumbo, Gadermann, & Zeisser, 2007) and it demonstrated adequate internal consistency, theta = .63. The questionnaire assesses children's knowledge about ‘dos’ and ‘don'ts’ when encountering dogs in typical daily scenarios in rural China. An example item is, “If you see stray dogs displaying unusual behavior like wandering around here and there aimlessly and easily getting irritated and angry, you should: A. Ignore it; B. Explore/Play with the dog to understand better what is going on; C. Yell at the dog, or else tease it; D. Walk away from the dog quickly; E. Immediately report to an adult” The correct answer is D. Participants were scored correct (1) or incorrect (0) for each item and the final composite score was the percentage of correct responses (0-100%), with higher scores indicating greater knowledge.

Perceived vulnerability to dog bites

The child-dog interaction perceived vulnerability questionnaire is a 12-item measure of children's perceived vulnerability toward child-dog interactions. Also developed in pilot research (Shen et al., 2013a), it has adequate internal reliability (Cronbach's α = .77) and construct validity (confirmatory factor analysis supported unidimensionality of the measure, χ2/df = 3.04, p < .05; SRMS = .07; CFI = .76). An example item is, “I can run faster than unknown dogs so as to avoid their biting me.” Items are scored on a 5-point scale, with higher scores indicating greater perceived vulnerability. Final scores were averaged across all items (possible range = 1-5).

Simulated behaviors with dogs

The dollhouse task assessed children's risky behaviors with dogs by asking children to act out what to do in simulated real-life situations involving dogs. The measure was based on a model used previously in the US that had adequate psychometric properties (Schwebel, Morrongiello, Stewart, & Bell, 2012), but was adapted to the rural Chinese context. It included eight situations that create the greatest risk for dog bites to children, as identified in pilot research (Shen et al., 2014). An example scenario is as follows: “Researcher acts out child doll [gender-matched to participant] walking on the street in front of the dollhouse while a stray dog approaches from behind the child. Say: ‘(Child's Name) is playing outside his/her house. A stray dog comes along the street toward (Child's Name). What will (Child's Name) do now?’” Children's verbal and behavioral responses to each scenario were recorded verbatim and then coded as safe (0), moderate (0.5) or unsafe (1) by the researcher using an objective written coding system. Inter-rater reliability was obtained by a second researcher who did not communicate with the first and was excellent (kappa>.95) in pilot testing with 60 participants.

Quantifying risk in simulated situations is necessarily inexact but we treated the dollhouse task measure analytically as a ratio scale, such that risk intervals were considered equal across measures and across scores. Thus, the leap from a score of 0 to a score of 0.5 was conceptualized to be equal to the leap in risk from 0.5 to 1. Similarly, risk of a bite for a score of 0.5 was conceptualized to be equal across all scenarios. Given this assumption, children's final scores were computed as the average safety across the eight scenarios and ranged from 0 to 1, with higher scores indicating riskier simulated behaviors with dogs. Internal reliability was borderline (Cronbach's alpha = .65).

Intervention

The intervention comprised a 36-minute video displaying scripted testimonials of actual dog-bite experiences. Four adult actors, all native Mandarin speakers unfamiliar to the participants, presented the testimonials. Unlike what might be done with research in Western cultures, we chose adults to present the testimonials rather than peers because as part of a collectivist culture, Chinese children respect and admire adult authorities (Chen, Greenberger, Farruggia, Bush, & Dong, 2003; Ma, Shek, Cheung, & Lam, 2000). Pilot research in rural China confirmed that children learned dog safety lessons more effectively from adults than from peers or siblings (Shen, et al., 2013b). The four presenters were representative of the interpersonal environment with which children in rural China interact in terms of age (younger and older adults), gender ratio (3 females and 1 male, as mothers are more likely to care for children and school teachers are mostly female), and societal roles (school teacher, parents, grandparents).

Each of the four testimonials was nine minutes long and was scripted to accomplish standardized learning objectives: (1) introduction of who the speaker is (teacher, parent, grandparent); (2) introduction about the child victim's background and circumstances; (3) antecedents/circumstances of the dog-bite incident (i.e., how the child got bit, what the child was doing, what the dog was doing); (4) consequences of the incident (i.e., gravity of the injury, how the child was treated, what happened to the dog after the bite); (5) mistakes, decisions, and actions the child victim made, including perception of invulnerability the child may have had, plus alternative cognitions/behaviors that would have been safer; and (6) a summary of safety lessons learned from this incident. To enhance the educational effect and hold children's attention during the video session, the testimonials were edited to incorporate background melodies fitting the story and illustrative pictures, texts and video clips to create a vivid and emotion-arousing lesson. Both visual and audio cues were provided between each of the four testimonial stories also. Anecdotally, child attention was excellent while viewing the videos. The testimonials were scripted but all described actual bites described in the literature (Shen et al., 2014) among children of the same age as the target population in rural Chinese culture.

The comparison group watched scripted testimonials of identical length and structure on drowning risk, another significant public health risk among rural Chinese children (Guo, 2010; Zhang, Chen, Deng, Xu, & Hu, 2003).

Procedure

IRB approval was obtained from University of Alabama at Birmingham and written consent from participating school officials. Signed informed consent was obtained from students’ caregivers and signed informed assent from students. All children in grades 3-4 (100%) at the participating school participated in the study. As shown in Figure 1, there was no attrition throughout the study, although only 276 (99%) children completed the pre-intervention dollhouse task due to scheduling conflicts at the time of the study.

Figure 1.

Figure 1

Open in a new tab

The CONSORT Flowchart Diagram of Study Enrollment

The study lasted three weeks. In week 1, all participating children completed the child-dog interaction safety knowledge and perceived vulnerability questionnaires in their classroom, and then were taken individually in randomized order to a quiet and private room to complete the simulated behavior assessment using the dollhouse task. Intervention group membership was assigned subsequently, so both researchers and participants were masked to group membership during pre-tests.

Implementation of the intervention commenced one week later. Testimonial videos were delivered to children at individual desktop computers with headphones. Each child was randomly assigned to watch one of the two testimonial videos using a 1:1 randomization scheme. Random assignment occurred within the same classrooms to reduce clustering influences in data analysis. In the final week of the study, the post-test was administered following the same procedure as the pre-test (knowledge and perceived vulnerability questionnaires completed in classroom, and individualized assessment of simulated behavior in the dollhouse task in a separate room). Researchers administering post-tests were masked to group membership.

Data Analysis Plan

Initial analyses were performed to examine descriptive data and intercorrelations between and among outcome variables at both pre- and post-test, along with demographic variables. Independent samples t-test and chi-square tests were performed next to examine group differences on demographic and outcome measures. Following this, the effects of the intervention on children's safety knowledge, perceived vulnerability and simulated behaviors were tested respectively using stepwise linear regression models, controlling for pre-intervention test scores and demographic variables. All of those analyses were conducted using SPSS 21.0. The final step of the analysis was to evaluate the mediational roles of safety knowledge and perceived vulnerability in the effect of intervention on children's behavioral changes using bias-corrected bootstrapping methods. This analysis was conducted using Mplus 6.0.

RESULTS

Thirty-one (11%) of the 280 participants had missing data on at least one variable. Together, the missing data constituted 0.3% of total data points. Four participants (1%) missed the pre-dollhouse task due to scheduling conflicts. Five participants (2%) were missing the Parent Questionnaire due to unavailability of literate caregivers. Other missing data occurred in the Child Questionnaire, most likely because children overlooked items at the bottom of a page while turning to the next page. Relations between missingness and other variables were examined using bivariate correlation analysis, and missingness was not significantly correlated with any of the variables of interest in this study. Since far less than 5% of data points were missing, and missing data points were distributed across variables and across participants, pairwise deletion and estimations (e.g. imputation) are likely to produce similar results (Tabachnick & Fidell, 2013). Therefore, pairwise deletion was applied to all missing data.

Descriptive statistics for all variables of interest appear in Table 1, both for the full sample and for each group. As expected given the randomization scheme, no statistical difference was found between the groups on most demographic variables. The only difference occurred on the exposure frequency to dogs measure, in which children in the intervention group reported somewhat less frequent exposure to dogs than those in the comparison group (χ2 (5) = 16.16, p < .05). Given this result, we controlled for exposure to dogs in all subsequent analyses. The safety knowledge, perceived vulnerability and simulated behavior in the dollhouse task measures also did not differ statistically prior to intervention.

Table 2 presents a correlation matrix including the major outcome variables plus demographic variables. No evidence of multicollinearity between predictor variables emerged. Intervention condition was significantly correlated with children's safety knowledge (r = .38, p < .01), perceived vulnerability (r = .33, p < .01), and simulated risky behaviors with dogs in the dollhouse task (r = −.34, p < .01) in post-intervention tests, but not in pre-intervention tests. Dog-bite history was not correlated significantly with any of the demographic or outcome measures, so we analyzed children with and without dog-bite history together in the main analyses.

Table 2.

Correlation Matrix for Major Outcome and Demographic Variables

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
1. Age --
2. Gender .01 --
3. Ethnicity −.02 −.02 --
4. Socioeconomic Status (SES) −.09 .00 .11 --
5. Dog-Bite History −.01 .06 .02 −.06 --
6. Dog Exposure −.07 .08 .09 −.10 .10 --
7. Dog Ownership .04 .03 −.09 −.02 .06 .36** --
8. Daily Risky Practice −.02 .19** −.01 −.02 .11 .48 .33** --
9. Condition .01 −.04 .05 .07 −.01 −.11 −.03 −.05 --
10. Pre-Knowledge .10 −.07 .01 .05 .02 −.02 .03 −.19** −.02 --
11. Pre-Perceived Vulnerability .06 −.07 −.02 −.02 .09 −.15* .05 −.33** .03 .48** --
12. Pre-Behavior −.07 .12 −.01 −.04 −.03 .13* −.06 .21** −.06 −.30** −.31** --
13. Post-Knowledge .08 −.10 .08 .04 −.02 −.10 −.09 −.19** .38** .43** .40** −.26** --
14. Post-Perceived Vulnerability .10 −.07 −.10 .07 −.02 −.19* −.02 −.26** .33** .30** .56** −.28** .62** --
15. Post-Behavior −.01 .02 −.04 −.07 .03 .18** .02 .15* −.34** −.33** −.26** .60** −.46** −.44** --
Open in a new tab
*

p<.05

**

p<.01

Our first primary hypothesis was that children in the intervention group would have greater knowledge, higher perceived vulnerability and safer simulated behavior with dogs following the intervention compared to children in the comparison group. Descriptive statistics shown in the shaded section of Table 1 support this hypothesis. One way to interpret the results is that the intervention group benefited from a 13% reduction in risk (−0.13 point difference from pre-behavior to post-behavior in the dollhouse task) compared to the comparison group, without adjusting for baseline functioning or other factors. To reinforce descriptive findings on the simulated behavior task, Figure 2 illustrates the change in responses among both groups from pre-intervention to post-intervention.

Figure 2.

Figure 2

Open in a new tab

Distribution of Children's Responses in the Dollhouse Simulated Behavior Task by Visit and Group

Three stepwise linear regression analyses were then performed to further evaluate the intervention effect on the three outcomes with respective post-intervention scores (knowledge, perceived vulnerability, simulated behavior) as the dependent variables, training condition as an independent variable, and respective pre-test scores and demographic factors (age, gender, ethnicity, SES, dog-bite history, exposure frequency to child-dog interactions, dog ownership, and daily risky practice with dogs) as control variables entered in the first step of analysis. For each outcome, we computed the adjusted effect size (adjusted Cohen's d) of the intervention by subtracting the pre-intervention effect size from the post-intervention effect size (Durlak, 2009).

Table 3 shows the effect of the intervention on safety knowledge. Covariates in the first step accounted for 22% of the variance in the outcome, R2 = .22, F (9, 254) = 7.75, p < .001. No demographic variables were predictive of post-intervention safety knowledge, but children with more pre-intervention safety knowledge had more post-intervention safety knowledge (B = .45, SE = .07, 95% CI = [.32, .58], β = .39, t = 6.81, p < .001). In the second step, training condition accounted for a significant proportion of the variance, R2 change = .14, F (1, 253) = 56.84, p < .001. Children in the intervention group had significantly more safety knowledge about dogs compared to the comparison group (B = 12.46, SE = 1.65, 95% CI = [9.20, 15.71], β = .39, t = 7.54, p < .001). The adjusted Cohen's d for the intervention effect on safety knowledge was 0.85 (pre-intervention Cohen's d = −0.04, post-intervention Cohen's d = 0.81).

Table 3.

Stepwise Linear Regression Analysis Predicting Children's Post-Intervention Safety Knowledge on Child-Dog Interactions

Step 1 Step 2
Variable B SE B β B SE B β
Age (Years) .67 1.11 .03 .50 1.00 .03
Gender (1=Male, 0=Female) −2.35 1.84 −.07 −1.78 1.66 −.06
Ethnicity (1=Han, 0=Other) 5.51 3.27 .10 6.82 2.96 .12*
Socioeconomic Status (SES) .79 1.26 .04 .40 1.14 .02
Dog Bite History (1=Yes, 0=No) −.36 2.28 −.01 −.35 2.06 −.01
Dog Exposure (1-6) .04 .54 .01 .47 .49 .06
Dog Ownership (1=Yes, 0=No) −2.60 2.02 −.08 −2.78 1.83 −.09
Daily Risky Practice (1-5) −2.73 1.80 −.10 −2.60 1.63 −.10
Pre-Knowledge (0-100%) .45 .07 .39*** .46 .06 .41***
Condition (1=Intervention, 0=Control) 12.46 1.65 .39***
R2 change .22 .14
df for R2 change 9/254 1/253
F for R2 change 7.75*** 56.84***
Open in a new tab
*

p<.05

***

p<.001.

Table 4 shows the effect of the intervention on perceived vulnerability. Covariates in the first step accounted for 34% of the variance in the outcome, R2 = .34, F (9, 262) = 15.29, p < .001. No demographic variables were predictive of post-intervention perceived vulnerability, but children with higher pre-intervention perceived vulnerability had higher post-intervention perceived vulnerability (B = .50, SE = .05, 95% CI = [.40, .60], β = .53, t = 9.76, p < .001). In the second step, training condition accounted for a significant proportion of the variance, R2 change = .10, F (1, 261) = 41.97, p < .001. Children in the intervention group had significantly higher perceived vulnerability about child-dog interactions compared to the comparison group (B = .35, SE = .05, 95% CI = [.24, .45], β = .31, t = 6.48, p < .001). The adjusted Cohen's d for the intervention effect on perceived vulnerability was 0.63 (pre-intervention Cohen's d = 0.07, post-intervention Cohen's d = 0.70).

Table 4.

Stepwise Linear Regression Analysis Predicting Children's Post-Intervention Perceived Vulnerability on Child-Dog Interactions

Step 1 Step 2
Variable B SE B β B SE B β
Age (years) .05 .04 .08 .05 .03 .07
Gender (1=Male, 0=Female) −.02 .06 −.02 −.01 .05 −.01
Ethnicity (1=Han, 0=Other) .19 .11 .09 .22 .10 .11
Socioeconomic Status (SES) .06 .04 .08 .05 .04 .06
Dog Bite History (1=Yes, 0=No) −.07 .07 −.05 −.07 .07 −.05
Dog Exposure (1-6) −.02 .02 −.08 −.01 .02 −.03
Dog Ownership (1=Yes, 0=No) −.02 .07 −.02 −.03 .06 −.03
Daily Risky Practice (1-5) −.03 .06 −.04 −.04 .06 −.04
Pre-Perceived Vulnerability (1-5) .50 .05 .53*** .49 .05 .53***
Condition (1=Intervention, 0=Control) .35 .05 .31***
R2 change .34 .10
df for R2 change 9/262 1/261
F for R2 change 15.29*** 41.97***
Open in a new tab
***

p < .001.

Table 5 shows the effect of the intervention on children's simulated behaviors with dogs in the dollhouse task. Covariates in the first step accounted for 42% of the variance in the outcome, R2 = .38, F (9, 257) = 17.22, p < .001. No demographic variables were predictive of post-intervention simulated behaviors with dogs except for dog exposure (B = .01, SE = .01, 95% CI = [.00, .02], β = .12, t = 2.00, p = .047). Children who had riskier pre-intervention behaviors with dogs demonstrated riskier post-intervention simulated behaviors with dogs (B = .59, SE = .05, 95% CI = [.49, .69], β = .60, t = 11.57, p < .001). In the second step, training condition accounted for a significant proportion of the variance, R2 change = .08, F (1, 256) = 37.31, p < .001. Children in the intervention group demonstrated significantly less risky simulated behaviors with dogs compared to the comparison group (B = −.11, SE = .02, 95% CI = [−.14, −.07], β = −.29, t = −6.11, p < .001). The adjusted Cohen's d for the intervention effect on reducing risky simulated behaviors was 0.63 (pre-intervention Cohen's d = 0.11, post-intervention Cohen's d = 0.74).

Table 5.

Stepwise Linear Regression Analysis Predicting Children's Post-Intervention Behavior in the Dollhouse Task

Step 1 Step 2
Variable B SE B β B SE B β
Age (years) .01 .01 .03 .01 .01 .03
Gender (1=Male, 0=Female) −.02 .02 −.04 −.02 .02 −.05
Ethnicity (1=Han, 0=Other) −.02 .03 −.03 −.03 .03 −.04
Socioeconomic Status (SES) −.01 .01 −.01 −.01 .01 −.01
Dog Bite History (1=Yes, 0=No) .01 .02 .01 .01 .02 .02
Dog Exposure (1-6) .01 .01 .12* .01 .01 .08
Dog Ownership (1=Yes, 0=No) .01 .02 .03 .01 .02 .03
Daily Risky Practice (1-5) −.01 .02 −.03 −.01 .02 −.02
Pre-Behavior (0-1) .59 .05 .60*** .58 .05 .58***
Condition (1=Intervention, 0=Control) −.11 .02 −.29***
R2 change .38 .08
df for R2 change 9/257 1/256
F for R2 change 17.22*** 37.31***
Open in a new tab
*

p<.05

***

p<.001

The final step of the analysis was to test whether perceived vulnerability mediated the effect of the intervention on children's simulated behaviors with dogs. The bootstrapping method with 95% bias-corrected confidence estimates and 10,000 bootstrap resamples (MacKinnon, Lockwood, & Williams, 2004; Preacher & Hayes, 2004, 2008) was applied to test a two-mediator model with training condition as the independent variable, post-test scores on simulated behaviors with dogs as the dependent variable, and post-test scores on safety knowledge and perceived vulnerability as two parallel mediators, controlling for pre-test scores on knowledge, perceived vulnerability and simulated behaviors, and demographic variables. Results are presented in Figure 3. The mediation role of perceived vulnerability was supported by the significant indirect effect from intervention condition to simulated behavior via perceived vulnerability (B = −.02, SE = .01, 95% CI = [−.04, −.007], β = −.06). Children who watched the testimonial video on dog-bite prevention had significantly higher perceived vulnerability (B = .36, 95% CI = [.25, .47], β = .33, t = 6.51, p < .001), which led to less riskier simulated behaviors with dogs (B = −.06, 95% CI = [−.10, −.02], β = −.18, t = −2.82, p = .01). The mediation role of safety knowledge was also supported by the significant indirect effect from intervention condition to simulated behavior via safety knowledge (B = −.02, SE = .01, 95% CI= [−.04, −.003], β = −.06). Children who watched the testimonial video on dog-bite prevention had significantly more safety knowledge (B = 12.62, 95% CI = [9.40, 15.76], β = .39, t = 7.74, p < .001), which led to less riskier simulated behaviors with dogs (B = −.002, 95% CI = [−.003, <−.001], β = −.15, t = −2.23, p = .03). The direct effect between condition and children's simulated behaviors with dogs was still significant after taking into account these two mediators (B = −.07, SE = .02, 95% CI = [−.10, −.03], β = .18, t = 3.39, p = .001), suggesting partial mediation.

Figure 3.

Figure 3

Open in a new tab

Mediation Model: Safety Knowledge and Perceived Vulnerability Mediating the Relation between Intervention Condition and Children's Simulated Behavior in the Dollhouse Task

DISCUSSION

The primary purpose of this study was to evaluate the efficacy of a video-based testimonial intervention to improve children's safety with dogs in rural China. We also studied the underlying mechanism of this behavior change. We hypothesized that the video-based testimonials would improve children's safety knowledge, increase perceived vulnerability, and reduce risky behaviors with dogs in a simulated dollhouse task. The results offered support for these hypotheses: Children randomly assigned to the intervention group, who watched the video-based testimonials on dog-bite prevention, demonstrated a significantly higher level of safety knowledge and perceived vulnerability about child-dog interactions, and showed significantly less risky simulated behaviors with dogs in the dollhouse task compared to those assigned to the comparison group, who watched similar video-based testimonials on drowning prevention. In interpreting the results, we caution that that our assessment of simulated behavior in the dollhouse task was based on a ratio scale measuring risk rather than a count of injury events. Thus, we approximated the likelihood a child might be bit by a dog in the future rather than measuring or estimating actual events.

To understand not only if an intervention works, but also how an intervention works, we further hypothesized that both increased knowledge and increased perceived vulnerability would mediate the effect of the intervention on children's simulated behaviors with dogs. This hypothesis was also supported. Training with video-based testimonials improved children's safety knowledge and perceived vulnerability toward child-dog interactions, which in turn led to less risky simulated behaviors with dogs in the dollhouse task. In addition, it was found that perceived vulnerability and safety knowledge only partially mediated the effect of the intervention on simulated behaviors, suggesting there are other mechanisms that also underlie the relationship and were unaccounted for in this study.

Historically, dog safety knowledge has been among the most common targets for pediatric dog-bite interventions (Dixon et al., 2013; Mello, Getz, Lapidus, Moss, & Soulos, 2007; Meints & De Keuster, 2006, 2009; Spiegel, 2000; Szecsi et al., 2010; Wilson et al., 2003). Increasing knowledge is a logical intervention objective: we would assume increased knowledge plays a foundational role for behavioral change. However, learning rules and facts is insufficient in many cases to impart health-related behavioral change (Deal et al., 2000). One mechanism to promote translation of increased knowledge about safety into behavioral change is to increase children's perceived vulnerability to injury through the intervention (Jackson & Aiken, 2000; Janz & Becker, 1984; Morrongiello & Matheis, 2007a). The testimonials in this study were designed to teach safety rules and knowledge in the context of vividly presented first-person accounts that facilitate richer learning and processing of cognitive information using emotional stimuli rather than simple repetitive memorization, thus improving translation to behavioral change (De Wit et al., 2008; Loewenstein et al., 2001; Morrongiello & Matheis, 2007b).

This study offered several methodological strengths: the use of first-person testimonial accounts, inclusion of an assessment of simulated behavior, and recruitment of a large sample in an understudied area, rural China. The study also had limitations. First, some of our measures showed suboptimal psychometric properties. Like much research in relatively unexplored areas, results should be interpreted cautiously until psychometrically sound measures of relevant constructs are created. Second, outcomes were evaluated just one week after implementation of the intervention (thus the long-term proportion of children bitten by a dog in each condition were unknown), and children's simulated behaviors with dogs were assessed rather than evaluating actual behaviors with live dogs. Future work should consider long-term retention of intervention effects, including outcomes such as the proportion of children bitten by a dog in training versus comparison groups, and should incorporate behavioral outcome measures with live dogs rather than only simulations. Third, this study focused only on children in a single rural Chinese school. Future research should evaluate translation across rural China and to other cultures and settings with similar risks (e.g., Latin America, South Asia, Africa). Translation to other developmental stages (younger and older children) also would be valuable, which may be achieved through a computerized story-telling method which is both attention-drawing (especially for younger children) and cost-effective for mass delivery.

Finally, although this study found that the testimonial-based intervention significantly changed all outcomes, the intervention effect was larger for safety knowledge (adjusted Cohen's d = 0.85) than perceived vulnerability (d = 0.63) or simulated behavior (d = 0.63). This is a common result in health behavior change programs (Meyer & Mark, 1995; Wilson, Lipsey, & Derzon, 2003) and does not necessarily imply that the program is ineffective or should be abandoned. In particular, even a modest effect size can have substantial impact when the population is large; given the size of our target population (>100 million children in rural China), even small behavior changes can result in large reduction of the public health burden. One future direction that might improve the intervention's efficacy would be to consider incorporation of live testimonials presented by actual people rather than video-taped stories. Such an experience might influence children's emotions and thus behaviors more dramatically, although it also would increase the economic and labor cost of the intervention.

Implications of this study are many; we highlight three. First, in terms of public health implications, over 100 million school-aged children live in rural China, all at significant risk of dog-bite injuries and rabies infection. This short-term efficacy study offers preliminary evidence of an intervention program that has potential to reduce dog-bite injury risk, including for children who have a history of previous dog bite. Second, the study has implications to behavioral science and injury prevention. Grounded in health behavior change theory, the positive results remind us how grounded theory-based interventions can be effective to change long-standing health-related habits and ultimately to reduce risk. Finally the adoption of mediation analysis to consider not just if an intervention works but also how it works is critical to health behavior intervention development. We discovered behavioral change was mediated both by increased knowledge and increased perceived vulnerability among the sample. Another way to frame this mediational effect is that testimonials influenced both cognition and perception-induced emotion (Morrongiello & Matheis, 2007b). Future work evaluating health-related interventions should target both knowledge and perceived vulnerability to improve the efficacy of intervention programs, and should also try to parse out the relative contributions of the two elements on children's behavioral change at different ages/developmental stages.

In conclusion, this study developed and evaluated a video-based testimonial intervention to reduce pediatric dog-bite injury risk in rural China. The positive findings have implications for future research and consideration of how interventions such as this one may be used throughout China and also translated to other cultures and regions with high rates of pediatric dog-bite injury.

Acknowledgments

Thanks to my dissertation committee members Dr. Samuel Cartner, Prof. Michele Forman, Dr. Sylvie Mrug, and Dr. Despina Stavrinos for insightful guidance on this project. Thanks to Prof. Forman and her team at the UAB Media Commons for support of the development of the testimonials. Thanks to Jinhong Guo, Hui Xu, and Manhong Yao for assisting with data collection. The research was partially supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under Award Number R21HD075960, American Psychological Association Division 38 (Health Psychology), Clinton Global Initiative University, and UAB Sparkman Center for Global Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the above-mentioned agencies and institutions.

REFERENCES

  1. Avner JR, Baker MD. Dog bites in urban children. Pediatrics. 1991;88(1):55–57. [PubMed] [Google Scholar]
  2. Beaver BV. A community approach to dog bite prevention. Journal of the American Veterinary Medical Association. 2001;218:1732–1749. doi: 10.2460/javma.2001.218.1732. [DOI] [PubMed] [Google Scholar]
  3. Chapman S, Cornwall J, Righetti J, Sung L. Preventing dog bites in children: Randomised controlled trial of an educational intervention. BMJ: British Medical Journal. 2000;320(7248):1512–1513. doi: 10.1136/bmj.320.7248.1512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chen C, Greenberger E, Farruggia S, Bush K, Dong QI. Beyond parents and peers: The role of important non-parental adults (VIPs) in adolescent development in China and the United States. Psychology in the Schools. 2003;40(1):35–50. [Google Scholar]
  5. Chouinard MM. Children's questions about animals. Monographs of the Society for Research in Child Development. 2007;72(1):58–82. doi: 10.1111/j.1540-5834.2007.00412.x. [DOI] [PubMed] [Google Scholar]
  6. De Wit JB, Das E, Vet R. What works best: Objective statistics or a personal testimonial? An assessment of the persuasive effects of different types of message evidence on risk perception. Health Psychology. 2008;27(1):110–115. doi: 10.1037/0278-6133.27.1.110. [DOI] [PubMed] [Google Scholar]
  7. Deal LW, Gomby DS, Zippiroli L, Behrman RE. Unintentional injuries in childhood: Analysis and recommendations. The Future of Children. 2000;10:4–22. [PubMed] [Google Scholar]
  8. Deng K, Tan J, Zhang X. Shenzhen shi futian qu 2000-2002 nian bei dong wu yao shang ren qun de liu xing bing xue fen xi [An epidemiological analysis of animal-bite cases in Futian District, Shenzhen City between 2000 and 2002] (in Chinese). Chinese Journal of Public Health Management. 2004;1:52–53. [Google Scholar]
  9. Dixon CA, Pomerantz WJ, Hart KW, Lindsell CJ, Mahabee-Gittens EM. An evaluation of a dog bite prevention intervention in the pediatric emergency department. Journal of Trauma and Acute Care Surgery. 2013;75(4):S308–S312. doi: 10.1097/TA.0b013e31829be2bc. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Duan B. Mo hu shi nong cun gou huan [Do not neglect the danger of dogs in rural areas] (in Chinese). Henan Science & Technology. 2008;4:44. [Google Scholar]
  11. Duperrex O, Blackhall K, Burri M, Jeannot E. Education of children and adolescents for the prevention of dog bite injuries. Cochrane Database Syst Rev. 2009;2:1–22. doi: 10.1002/14651858.CD004726.pub2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Durlak JA. How to select, calculate, and interpret effect sizes. Journal of Pediatric Psychology. 2009;34(9):917–928. doi: 10.1093/jpepsy/jsp004. [DOI] [PubMed] [Google Scholar]
  13. Eppley BL, Schleich AR. Facial dog bite injuries in children: treatment and outcome assessment. Journal of Craniofacial Surgery. 2013;24(2):384–386. doi: 10.1097/SCS.0b013e31827fee33. [DOI] [PubMed] [Google Scholar]
  14. Fritz MS, MacKinnon DP. Required sample size to detect the mediated effect. Psychological Science. 2007;18(3):233–239. doi: 10.1111/j.1467-9280.2007.01882.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Guo Q. Guang dong lian ping xian xiao xue sheng ni shui jian kang jiao yu gan yu jin qi xiao guo ping jia [The short-term evaluation of health education on drowning prevention among elementary school students in Lianping County, Guangdong Province](in Chinese) Jinan University; China: 2010. (Unpublished master thesis) [Google Scholar]
  16. Hon KE, Fu CA, Chor C, Tang PH, Leung T, Man CY, Ng P. Issues associated with dog bite injuries in children and adolescents assessed at the emergency department. Pediatric Emergency Care. 2007;23(7):445–449. doi: 10.1097/01.pec.0000280509.67795.a9. [DOI] [PubMed] [Google Scholar]
  17. Jackson KM, Aiken LS. A psychosocial model of sun protection and sunbathing in young women: the impact of health beliefs, attitudes, norms, and self-efficacy for sun protection. Health Psychology. 2000;19(5):469–478. doi: 10.1037//0278-6133.19.5.469. [DOI] [PubMed] [Google Scholar]
  18. Janz NK, Becker MH. The health belief model: A decade later. Health Education & Behavior. 1984;11(1):1–47. doi: 10.1177/109019818401100101. [DOI] [PubMed] [Google Scholar]
  19. Loewenstein GF, Weber EU, Hsee CK, Welch N. Risk as feelings. Psychological Bulletin. 2001;127(2):267–286. doi: 10.1037/0033-2909.127.2.267. [DOI] [PubMed] [Google Scholar]
  20. Ma HK, Shek DT, Cheung PC, Lam COB. Parental, peer, and teacher influences on the social behavior of Hong Kong Chinese adolescents. The Journal of Genetic Psychology. 2000;161(1):65–78. doi: 10.1080/00221320009596695. [DOI] [PubMed] [Google Scholar]
  21. MacKinnon DP, Lockwood CM, Williams J. Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research. 2004;39(1):99–128. doi: 10.1207/s15327906mbr3901_4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Meyer TJ, Mark MM. Effects of psychosocial interventions with adult cancer patients: a meta-analysis of randomized experiments. Health Psychology. 1995;14(2):101–108. doi: 10.1037//0278-6133.14.2.101. [DOI] [PubMed] [Google Scholar]
  23. Mcheik JN, Vergnes P, Bondonny JM. Treatment of facial dog bite injuries in children: A retrospective study. Journal of Pediatric Surgery. 2000;35(4):580–583. doi: 10.1053/jpsu.2000.0350580. [DOI] [PubMed] [Google Scholar]
  24. Meints K, De Keuster T. Test Yourself--A First Assessment of the Dog Bite Prevention Project” Blue Dog”. Part I: Education--The” Blue Dog” is Born; Part II: The” Blue Dog”--Do Children Really Get the Message? Proceedings WSAVA Congress. 2006:11–14. [Google Scholar]
  25. Meints K, De Keuster T. Don't kiss a sleeping dog: The first assessment of “the blue dog” bite prevention program. Journal of Pediatric Psychology. 2009;34(10):1084–1090. doi: 10.1093/jpepsy/jsp053. [DOI] [PubMed] [Google Scholar]
  26. Mello MJ, Getz MA, Lapidus G, Moss J, Soulos P. Innovations in injury prevention education. Journal of Trauma - Injury, Infection, and Critical Care. 2007;63(3):S7–S9. doi: 10.1097/TA.0b013e31812f5ecd. [DOI] [PubMed] [Google Scholar]
  27. Morrongiello BA, Matheis S. Addressing the issue of falls off playground equipment: An empirically-based intervention to reduce fall-risk behaviors on playgrounds. Journal of Pediatric Psychology. 2007a;32(7):819–830. doi: 10.1093/jpepsy/jsm029. [DOI] [PubMed] [Google Scholar]
  28. Morrongiello BA, Matheis S. Understanding children's injury-risk behaviors: The independent contributions of cognitions and emotions. Journal of Pediatric Psychology. 2007b;32(8):926–937. doi: 10.1093/jpepsy/jsm027. [DOI] [PubMed] [Google Scholar]
  29. Morrongiello BA, Rennie H. Why do boys engage in more risk taking than girls? The role of attributions, beliefs, and risk appraisals. Journal of Pediatric Psychology. 1998;23(1):33–43. doi: 10.1093/jpepsy/23.1.33. [DOI] [PubMed] [Google Scholar]
  30. Peters V, Sottiaux M, Appelboom J, Kahn A. Posttraumatic stress disorder after dog bites in children. The Journal of Pediatrics. 2004;144(1):121–122. doi: 10.1016/j.jpeds.2003.10.024. [DOI] [PubMed] [Google Scholar]
  31. Preacher KJ, Hayes AF. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers. 2004;36(4):717–731. doi: 10.3758/bf03206553. [DOI] [PubMed] [Google Scholar]
  32. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods. 2008;40(3):879–891. doi: 10.3758/brm.40.3.879. [DOI] [PubMed] [Google Scholar]
  33. Schwebel DC. Reducing child injury: evaluation of interventions targeting children and the adults who supervise them. Injury Prevention. 2012;18(Suppl. 1):A61. [Google Scholar]
  34. Schwebel DC, Morrongiello BA, Davis AL, Stewart J, Bell M. The Blue Dog: Evaluation of an interactive software program to teach young children how to interact safely with dogs. Journal of Pediatric Psychology. 2012;37(3):272–281. doi: 10.1093/jpepsy/jsr102. [DOI] [PubMed] [Google Scholar]
  35. Shen J, Li S, Xiang H, Pang S, Xu G, Schwebel DC. A multi-site study on knowledge, attitudes, beliefs and practice of child-dog interactions in rural China. International Journal of Environmental Research and Public Health. 2013a;10(3):950–962. doi: 10.3390/ijerph10030950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Shen J, Li S, Xiang H, Pang S, Xu G, Yu G, Schwebel DC. Dog safety in rural China: Children's sources of safety information and effect on knowledge, attitudes, and practices. Accident Analysis & Prevention. 2013b;59:164–169. doi: 10.1016/j.aap.2013.05.014. [DOI] [PubMed] [Google Scholar]
  37. Shen J, Li S, Xiang H, Lu S, Schwebel DC. Antecedents and consequences of pediatric dog-bite injuries and their developmental trends: 101 cases in rural China. Accident Analysis & Prevention. 2014;63:22–29. doi: 10.1016/j.aap.2013.10.025. [DOI] [PubMed] [Google Scholar]
  38. Si H, Guo Z, Hao Y, Liu Y, Zhang D, Rao S, Lu J. Rabies trend in China (1990–2007) and post-exposure prophylaxis in the Guangdong province. BMC Infectious Diseases. 2008;8(1):113–122. doi: 10.1186/1471-2334-8-113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Slater MD, Rouner D. Value-affirmative and value-protective processing of alcohol education messages that include statistical evidence or anecdotes. Communication Research. 1996;23(2):210–235. [Google Scholar]
  40. Spiegel IB. A pilot study to evaluate an elementary school-based dog bite prevention program. Anthrozoos: A Multidisciplinary Journal of the Interactions of People & Animals. 2000;13(3):164–173. [Google Scholar]
  41. Szecsi T, Barbero M, Del Campo A, Toledo M. Teaching respect, responsibility, and kindness through dog safety lessons. Childhood Education. 2010;87(2):125–128. [Google Scholar]
  42. Tabachnick BG, Fidell LS. Using multivariate statistics. 6th Ed. Pearson; Boston: 2013. [Google Scholar]
  43. Wang SY, Li YH, Chi GB, Xiao SY, Ozanne-Smith J, Stevenson M, Phillips MR. Injury-related fatalities in China: an under-recognised public-health problem. The Lancet. 2008;372(9651):1765–1773. doi: 10.1016/S0140-6736(08)61367-7. [DOI] [PubMed] [Google Scholar]
  44. Wilson F, Dwyer F, Bennett PC. Prevention of dog bites: Evaluation of a brief educational intervention program for preschool children. Journal of Community Psychology. 2003;31(1):75–86. [Google Scholar]
  45. Wilson SJ, Lipsey MW, Derzon JH. The effects of school-based intervention programs on aggressive behavior: a meta-analysis. Journal of Consulting and Clinical Psychology. 2003;71(1):136–149. [PubMed] [Google Scholar]
  46. Zhang P, Chen R, Deng J, Xu B, Hu Y. Jian kang jiao yu dui nong cun 0-4 sui er tong yi wai zhi xi yu ni shui gan yu xiao guo de ping jia [Evaluation of the intervening efficacy of health education on accidental suffocation and drowning of children aged 0-4 in countryside] (in Chinese). Chinese Journal of Pediatrics. 2003;41(7):497–500. [PubMed] [Google Scholar]
  47. Zhang J, Jin Z, Sun GQ, Sun XD, Ruan S. Modeling seasonal rabies epidemics in China. Bulletin of Mathematical Biology. 2012;74(5):1226–1251. doi: 10.1007/s11538-012-9720-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES