Study design and period
A mixed-method concurrent triangulation study design was conducted from November to April 2021.
All female private college students in the Dessie city administration were taken as the source population. Female students from private universities in Dessie City administration during the data collection period were considered as the study population.
Sampling method and sample size determination
The sample size was determined using a single population proportion formula taking the following assumptions: The lifetime prevalence of gender-based violence (physical and/or sexual violence) was 59.9% with a confidence interval of 95%, marginal error of 5%, marginal error of 10%. for non-response rate in study of Hawassa City college students a confidence interval (CI) of 95%, marginal error of 5%.
n = (Z α/2)2 P (1 − P)/W2, n = (1.96)2*0.599 (1 − 0.599)/(0.05)2, n = 3.8416 * 0.599 * 0.401 /0.0025 = 369. Finally, by adding 10% of non-responses, the total sample size was 406.
The sample size for the qualitative method was determined by the degree of saturation.
Simple random sampling using a computer random number generator was used to select study participants at each university using the student registry list to select 406 students. For the Qualitative part, the sample size was taken when the saturation level was reached. We say that the sample is Saturated when adding more participants to the study did not result in additional insights or information. Therefore, our sample size was 12 people conducted for an in-depth interview.
Violence: is defined as: “The intentional use of physical force or power, actual or threatened, against oneself, another person, or a group or community, that causes or has a high probability of causing injury, death, psychological harm, or deprivation”. .
gender violence “Gender violence” For this study, it is defined as physical and/or sexual violence directed at girls based on their gender. .
GV for life: defined as physical and/or sexual always faced that is directed at girls based on their gender in their university life .
physical violence: is any form of violent act that can result in physical harm, including the mild form (slapping and punching) or the severe form (kicking/drugs, beating/hitting with any object, burning/choking, and threatening with a knife or gun, etc) against women or girls .
sexual violence: is defined as acts done to a girl through the intentional use of physical force or power, intimidation, or the threat (make fear) of having sex or engaging in sexual acts without the girl’s consent. Includes completed rape, attempted rape, and sexual harassment .
completed rape: is defined as any non-consensual penetration of the vagina, penetration obtained by physical bodily harm, by threat or deceit, or when the victim is unable to give consent .
attempted rape: defined as an attempt to engage in sexual intercourse without consent through physical bodily harm, threats, or deception, or when the victim is incapable of consent but without actual penetration of the vagina .
sexual harassment– These are unwanted sexual behaviors including physical contact, verbal comments, jocks, questions, and suggestions that are made intentionally on women or girls. .
family control: different types of measuring actions taken by parents against the behavior of children .
strict family control parents always know close friends, what they are doing away from home and their plan for the next day .
average family control parents sometimes meet close friends, what he does away from home and his plan for the next day .
Poor or loose family control parents never know her close friends, what she does outside the home, or her plan for the next day .
Family income: the amount of money incurred per month at the household level in Ethiopia 
Data collection procedure and quality assurance
The questionnaire was prepared in English and translated into the Amharic version. Finally, someone else translated it back into English to ensure consistency. A pre-test was held at Rada College Kombolcha Town. The data collection tool included 8 family histories, 8 sexual experiences, 8 physical violence states, 11 sexual violence states, 12 substance use, 6 school conditions, 6 GBV perceptions, and 11 sociodemographic questions. Qualitative data was collected using in-depth interviews and an observation checklist.
Rigor and reliability of the study.
The investigator and assessors discussed the objectives of the study, the confidentiality of the information, the content of the open-ended semi-structured questionnaire, and data quality management before data collection began. Few participants signed up once the data was transcribed to review ideas for transcription verification, so they thought the researcher was going to present them with a true picture from their perspective.
To ensure authenticity, all data was included in the final report to allow readers to see the basis on which conclusions were drawn. Rigor was achieved through strict attention to detail, adherence to procedures, and consistency and precision throughout the research process. A peer review was carried out to ensure credibility. Data, analyses, interpretations, and conclusions were continually reviewed by peers undergoing training in qualitative research. A session with peer researchers from the same background was used to check consistency between data analysis and thematic development.
An audit trail was created throughout the data collection, transcripts, a clear description of the analysis process, and interpretation of the data. Audio-recorded interviews were not destroyed until member review and transcript verification. Furthermore, reflexivity implied self-awareness, and the bracket was used to set aside any preconceived ideas regarding lived experience to ensure confirmation of the ability of the results that can reduce investigator bias.
After data collection, the principal investigators transcribed the audio-recorded data in the participant’s local language into written form and then translated into English, then the data was coded, categorized, and analyzed to facilitate interpretation. Therefore, the use of a tape recorder, careful probing, verbatim transcription, interviewing to saturation point, peer review, audit trail, and consideration of disparity were used to maintain the reliability of the data. study data.
Data processing and analysis
Data was checked for completeness and consistency after it was coded and entered into EPI-data v 126.96.36.199, and will then be exported to SPSS 23 statistical software for processing and analysis. Different frequency tables, graphs, and descriptive summaries were used to describe the study variables. A binary logistic regression analysis was used to see the importance of the association between the dependent and independent variables. Bivariate and multivariate variables were calculated to see the association between gender violence and the selected independent variables. In variable bivariate logistic regression analysis with a p A value of 0.20 or less was entered for multivariable logistic regression analysis to adjust for potential confounders, then a p value less than 0.05 was considered a predictor of VG, and an adjusted odds ratio was considered. with a 95% CI for the strength and significance of the association. The adequacy of the final model was verified using the Hosmer and Lemeshow goodness-of-fit test with a p the value was 0.9. The multiple collinearity test was verified using variance inflation factors (VIF) with 4.12.
Qualitative data were analyzed using thematic analysis. Prior to analysis, all collected data was transcribed into English. The interview was transcribed and conducted by the researcher. The data transcribed by the principal investigator were read several times to critically evaluate to obtain the concept and to group into themes according to the concept they contain. Therefore, the responses were categorized into each topic and subtopic. A researcher analyzed the data to answer the objective of the study and write a report based on categorized themes. Quotes were used to highlight each category and show association with each theme.