Study design and setting
This study is a part of a more comprehensive cross-sectional survey of different types of abuse among schoolchildren of basic education in Aden Governorate in Yemen that was conducted during the school year 2009–2010.
Pupils of grades 7, 8 and 9 were targeted in the survey. Children, at the age of 12–17 years, usually are capable of perceiving what is and is not abuse within the school context, they also could provide reliable information, and are capable of answering the questionnaire [10, 21].
The sample size was calculated using the assumed proportion of 0.5 in order to obtain the maximum possible sample size, with a level of confidence 95%, and 0.03, as maximum allowable error. Accordingly, the calculated sample size was 1066 pupils who were proportionally distributed according to pupils' gender (667 males and 399 females).
A multi-stage stratified random sampling was performed. In the first stage, four districts were randomly selected. In the second stage, two schools from each district also were randomly selected. In the third stage, the sample size (1066) was proportionally distributed according to the proportion of students in the selected grades for each school by gender. The sample frame for this study was the pupils' list in the selected grades of the schools incorporated in this study. Systematic random sampling was applied to select the number of pupils assigned for each grade in the selected schools.
An anonymous self-administrated questionnaire adapted from the Arabic version of Child Abuse Screening Tool Children's Institutional Version , was used for data collection.
The first part of the instrument covers questions about pupil's variables: gender, age, school grade, residence, in addition to parent's socio-demographic variables: family type, parent's education and parent's marital status.
The second part includes 10 items including acts such as humiliating, shouting, calling names, embarrassing them of being orphan, poor, having health problems, threatening with giving them bad marks, or expelling from school, isolating them from other children and destroying their belongings.
Experts in child abuse from Yemeni Universities judged the questionnaire, to find out if it is socially acceptable. Accordingly, some items were rephrased, and other items were dropped out. The final modified version of the questionnaire used in this study included the following questions:
Have you ever been exposed to any of these acts at school?
Humiliating, shouting, calling names, embarrassing of being orphan, embarrassing of being poor, threatening with bad marks, threatening to expelling from school, embarrassing for having health problems, isolating from children, destroying belongings. Pupils who answered with "Yes" were asked to report the frequency of the abuse acts by indicating 1-2 times, 3-4 times, ≥ 5 times. In addition they were asked to mention who did it: a teacher or a school administrator? .
A pilot study was conducted among 60 pupils (30 males and 30 females) from two schools not included in the main study, to ensure that the questionnaire items were clear, understandable and culturally acceptable. Chronbach alpha was used to test the internal consistency reliability, which was found to be 0.78.
We explained the questionnaire in detail to the pupils, and asked them to answer 'yes' or 'no' to each item and how many times they have experienced abusive acts during scholastic years. Those who responded affirmatively were asked to identify the perpetrators: teacher or school administrative staff. The rate of physical abuse was calculated by recoding the acts into dichotomous categories, with 0 = never and 1 = once or more . A 4- point scale (0 = none; 1 = 1-2 times; 2 = 3–4 times; 3 = 5 times and more) was used to indicate how often they had experienced each abuse act .
Emotional abuse in this study refers to pupils' reports of any undesirable or unpleasant emotional act inflicted on them by teachers or other school administrators (school principal, vice principal or other workers) which could potentially make them feel embarrassed while in school. Emotionally abused pupils were defined as those who answered positively to one or more of the emotional abuse acts.
The research protocol was approved by the Committee of Research and Postgraduate Studies in the Faculty of Medicine and Health Sciences, Aden University. Several levels of permission were granted before the study could proceed, including official approval from the authority of Aden Education Office. In addition, permission was sought from the Districts Directors of Education, as well as from school principals.
A written informed consent was sent to the pupils' parents describing the nature of the study, its importance, and its objectives. The consent also stated that the data's confidentiality would be assured, that the participation in the study was voluntary and those who refused participation will not lose any rights or privilege. Parents were asked to put their signature if they agree to have their child participate in the survey. The response rate was 85% of the targeted parents. For those who declined to consent, equal number was substituted following the same methodology.
The pupils' informed assent was obtained orally during which detailed explanation of the objectives and the importance of the research was provided. Potential participants were assured that all information obtained will be handled confidentially. Pupils were informed that they have the right to decline answering any question and/or to withdraw from the study at any time. All pupils whose parents gave consent to their participation have assented to participate in the study.
The Statistical Package for Social Science -SPSS – version 15 was used for data analysis. Quantitative variables were normally distributed after testing for normality using Kolmogorove-Smirnov test. Percentage was calculated as summary measure for the qualitative variables. Arithmetic mean, standard deviation and chi square test were used for the descriptive statistics. The statistical significant level was set at p-value < 0.05.
Binary logistic regressions were conducted to describe the bivariate association between emotional abuse and the pupils and parents variables. Multivariate analysis was performed to identify risk factors associated with emotional abuse. The results were discussed in terms of the odds ratio (OR) with 95% Confidence Interval (CI). The OR of the reference category is equal to "1". If an OR is greater than "1" this indicates an increase likelihood of the event (emotional abuse in this research) occurrence, while an OR less than "1" indicates a decreased likelihood of its occurrence.