Parental Attachment, Gender and Deviant Behaviour Among Secondary School Students in Homabay County- Kenya

I. INTRODUCTION

Despite all the educative information and preventive measures availed to the adolescents they still engage in deviant behaviour like drug use, risky sexual behaviour and violence. Such behaviours have irreversible negative effects to the society at large and the individual student. Some of these effects include destruction of self esteem, early pregnancies and early marriages, destruction of learning facilities, school dropouts and sexually transmitted diseases. The schools in isolation cannot successfully fight deviance and hence the need for all the stakeholders to join hands in fighting deviance. Parents as the primary caregivers contribute a lot to the behaviour of the secondary school students. What the parents give to their children and expose their children to determines what the child becomes in future. As the saying goes 'garbage in garbage out'. For deviance to be reduced in the schools, parents have to be fully involved right from the home environment where the child grows up.

If deviance in schools is not addressed in good time it may escalate to more aggressive forms of behaviour like terrorism and other criminal acts. The gender roles as defined by the societal norms also dictate the behaviour of an individual depending on whether one is a boy or a girl. Even priorities and opportunities in the society are availed based on gender. Some of these gender differences predispose one sex to deviant behaviour more than the other.

1.1 Theoretical Framework

This study was guided by parental attachment theory by John Bowlby and Mary Ainsworth. Parental attachment theory elaborates the specific types of parental attachment and their consequences on the development of the child. Guided by this theory, the study assumes that if a child experiences a strong positive attachment with the primary caregivers, then as an adolescent and later as an adult, the individual will exhibit less deviant behaviour and will coexist peacefully with others in the community. According to this theory a child is either securely attached to the parent or insecurely attached to the parent. Those who are securely attached are self reliant and have a positive self image. They relate well with others and are less likely to involve themselves in deviance. On the other hand, those who experience insecure attachment with their parents do not trust their parents, they do not care about close relationships, they are also likely to engage in casual sex and use sex as proof of their attractiveness and status. Those who experience the anxious insecure attachment style are highly dependent on relationships and live in constant fear of rejection and criticism. Beneath their conscious thoughts, the people with anxious attachment style ache and yearn for parental love or the love of a protector who can play a parental role (Brogaard, 2016). Such children feel that the people around them have not done enough for them and have in a way let them down. Parents behave inconsistently; showing love at times and at times disengaged.

1.2 Area of the Study

This study was carried out in Homabay County. This is an area situated along Lake Victoria and covers an area of approximately 3,154.7 km ${}^2$. In Nyanza region, Homabay County is the poorest ranking 15 in the whole country with a poverty index of 44.1% (Kenya County Fact Sheet, 2011). Apart from the high poverty index in Homabay, there are also other problems experienced by the residents for example, according to the National HIV and AIDs Estimates reports released by the National Aids Control Council, in Nyanza region, Homa bay County is the leading in HIV/AIDS infection at 25.7%, Siaya 23.7%, Kisumu 19.3% Migori 14.7%, Kisii 8% and Busia 6.8%. Recently Homabay County was also ranked second in the whole country at 33% in teenage pregnancy. This study therefore is timely as the findings would be useful for policy makers in formulating policies to address issues of deviant behaviour among secondary school students in the county. The study sought to ascertain the relationship between parental attachment and deviant behaviour and which gender is more susceptible to deviance between males and females.

1.3 Related Literature

Literature was reviewed in terms of the objectives of the study. The study reviewed literature from the world, Africa and finally literature from Kenya.

Imtiaz and Naqvi (2012) investigated the relationship between parental attachment and identity styles among adolescents in Pakistan. Using convenient sampling technique they selected a sample of 252 adolescents from four colleges/ universities. The findings showed that parental attachment was positively related with informational, normative and commitment identity styles. While Imtiaz and Naqvi (2012) employed convenient sampling technique and college students, the current study used simple random sampling technique and the participants were secondary school students, teachers and parents.

Ngai (2015) investigated the effects of parental care and parental control on adolescents' authenticity, bravery, perseverance, kindness, love, social intelligence, fairness and self regulation. They used qualitative paradigm and survey data collected from a convenient sample of 2010 Chinese adolescents recruited from secondary schools in Hong Kong. They revealed a significant relationship between parental care and authenticity, bravery, perseverance, kindness, love, social intelligence, fairness and self regulation. Ngai (2015) adopted qualitative paradigm as opposed to the current study which adopted the mixed methodology.

Tadesse, Mitikie, Yemane, Amenu and Tesfaye (2016) did a research in an attempt to unearth the possible determinants of risky sexual behaviour among preparatory school Students in Gurage Zone, SNNPR, Ethiopia. The study adopted a cross sectional quantitative study design. Using the multistage sampling technique, a sample of 418 participants was selected. The study revealed that parental attachment was significantly correlated to risky sexual behavior of students. Tadesse et al (2016) adopted a cross sectional design and a quantitative paradigm while the current study adopted a correlational design and mixed method of research.

A study conducted by Edobor and Ekechukwu (2015) in Nigeria which focused on the factors influencing lesbianism among senior secondary school students also revealed that personality traits depend on the parenting styles. Although Edobor and Ekechukwu used the simple random sampling technique, their study adopted the ex-post facto research design and their data was collected using questionnaires only. This differs from the current study which adopted a correlational study design and the data was collected through triangulation. The study by Edobor and Echukwu (2015) was purely quantitative while the current study adopted the mixed methodology.

Njagi (2012) carried out investigation on intimate partner violence among students of the University of Nairobi. The target population consisted of students who were in relationships. The sample was selected using snowball sampling. The study adopted a descriptive cross sectional design and data was collected using structured questionnaires and in depth interviews. The study discovered that main contributors to intimate partner violence include lack of trust, infidelity, alcohol and drug abuse. In the study by Njagi (2012), the participants were university students while the current study used a sample of secondary school students. Njagi (2012) sampled the participants purposively and used a cross sectional design as opposed to the current study which selected its sample using simple random sampling and purposive sampling techniques and adopted a correlational study design.

Chebukaka (2014) investigated drug abuse among students in public secondary schools in Vihiga County-Kenya. The study employed descriptive survey design and ex-post facto approach. The sample was selected randomly and data collected through questionnaires. The data collected was purely quantitative implying that the study only employed the quantitative paradigm. The study found that the students in Vihiga County averagely involved themselves in drugs. Chebukaka (2014) adopted ex-post facto design which only allows for purposive sampling and a quantitative paradigm while the current study adopted a correlational survey design and a mixed method of research. The sample in the current study was also randomly selected as opposed to the sample in the above reviewed study which was purposively selected.

Gender difference in behaviour is exhibited all over the world. Moitra and Mukherjee (2010) examined the relationship between parents' behavior and delinquency in male adolescents in India. They adopted a comparative study design. Their findings showed that parents' behavior is associated with delinquency. While Moitra and Mukherjee (2010) examined delinquency in male adolescents and utilized a comparative design, the current study involved both male and female adolescents and also adopted a correlational survey design.

Subhash (2011) carried out research on the socioeconomic and demographic impact on child labour in India. The study utilized a sample of 114,216 children of ages 10-14 and analyzed a secondary data from family health survey. The findings revealed that child labour is higher among the poor than the rich and that the number of female children subjected to child labour is higher than that of the male children. The sample in the study by Subhash (2011) consisted of children of age 10-14 while in the current study the participants consisted of adolescents, aged 14-17. While the study analyzed a secondary data, the present study used primary data and very recent secondary data.

Tsvetkova and Antonovna (2013) investigated the prevalence of drug use among university students in Russia. The study used a sample of 1,477 university students aged between 19 -24 years. Data analysis was done through simple distributions and central tendency measures. It was revealed that the prevalence of drug use among male students is higher than female students. The study also revealed that there is a higher social acceptability of drug use among males. Tsvetkova and Antonovna (2013) conducted research among university students while the current study involved secondary school students. While their data was analyzed descriptively the current study analyzed data using both descriptive and inferential statistics.

Gobopamang, (2011) studied the influence of gender role attitudes on risky sexual behavior in Botswana. The study used a secondary data and adopted a cross sectional survey design. The participants included individuals aged 10-64. The findings of this study revealed that women are more susceptible to risky sexual behaviour compared to the men. While Gobopamang (2011) analyzed a secondary data collected in 2008, in the current study the researcher collected primary data at the time of study and also analyzed recent secondary data. As opposed to the current study which adopted a correlational survey design, Gobopamang (2011) adopted a cross sectional survey design. Mehra (2013) examined sexual behavior among Ugandan university students using a sample size of 1,954 and found that there is a relationship between poor academic performance among females and inconsistent condom use. The study adopted a cross sectional study design and used questionnaires to collect data. While Mehra (2013) collected data using questionnaires, the current study collected data using triangulation. The study adopted a cross sectional design as opposed to the current study which employed a correlational survey design. In addition, Shraboni (2016) explored the social and cultural factors and risky sexual behavior associated with HIV/AIDS and STIs in Uganda in terms of gender. Data from the Uganda AIDS Indicator Survey (UAIS) 2011 was used. With a sample size of 16,607 (7,122 men 9485 women) aged 15-49 years, the study revealed that women are more infected with HIV/AIDS more than men. The study further concluded that women are significantly more vulnerable to any STIs and HIV/AIDS than men. While the above study used data that was collected in 2011 the current study collected primary data at the time of the research and also used recent secondary data.

Kabiru, Elung'ata, Mojola and Donatien (2014) investigated adversity in life and delinquency among Kenyan adolescents. They employed a cross sectional study design and revealed a significant positive association between adversity in both gender and delinquency. Kabiru et al (2014) utilized cross sectional design while the current study utilized a correlational survey design.

Wepukhulu, Mauyo, Poipoi, Achoka, Kafu and Walaba (2012) carried out research on the influence of socioeconomic status on premarital sex among secondary school students in western Kenya. Using a sample of 284 students who were selected using the systematic random sampling method and data collected through questionnaires, the cross sectional study established that girls are more sexually experienced than boys of the same age and academic level. The study further revealed that a significant association between gender and the attitude of the youth towards premarital sex. Albeit Wepukhulu et al (2012) conducted their study among secondary school students, their study adopted a cross sectional design wither ward to the present study which adopted a correlational survey design.

A research carried out in Kericho County by Sang, Chepchieng and Kariuki (2015) on the relationship between students' family socioeconomic status, school category and academic achievement failed to establish a relationship between students' gender and their academic achievement. However, they reported that girls in girls' only schools scored higher academic grades compared to the girls in mixed schools. While Sang et al (2015) conducted an ex-post facto research which employed a correlational research design and used a population of form four students in public secondary schools and data collected through questionnaires only, the present study adopted a correlational study but used form two students and the data was collected through triangulation.

Juma, Simatwa and Ayodo (2012) carried out a research on the impact of family socioeconomic status on the academic achievement of girls in secondary schools in Kisumu East. Their study revealed as the family SES improves, the academic performance of girls also improved. Although the study adopted a correlational study design and collected data through triangulation, the target population consisted of form four girls contrary to the current study whose participants consisted of form two students.

II. RESEARCH METHODOLOGY

The study adopted a correlational survey design and the data was analyzed both qualitatively and quantitatively. A correlational research design is the measurement of two or more variables to determine the relationship/ association among the variables or how the interaction of the variables influences their behavior pattern (Privitera & Wallace, 2011). The target population consisted of form two students from the 315 public and private secondary schools within Homabay County with a total population of 20,160 (11,752 boys and 8,408 girls). This population constituted 27 girls' schools, 35 boys' schools and 253 mixed day and boarding secondary schools (Homabay County Education Officer, 2019). Teachers and parents also took part in the study.

2.1 Sample Size

Conducting a survey of an entire population lacks practicability in terms of time and budget constraints and therefore a representative sub set of the population is required. Homa bay County has a total of eight sub counties with a population of 315 secondary schools and a target population of 20,160 form two students. 30% of the sub counties were randomly sampled. This translated to two sub counties with a population of 80 secondary schools and 5,120 form two students. According to Mertens (2010), a sample size of between 10% and 30% adequately represents the target population. Further still, Research Advisors (2006) opine that the recommended sample size for a population of 5,000, a confidence level of 95% and a margin of error (degree of accuracy) of 5% would be 357. Therefore, the representative sample of the students was 512 students from eight schools being 10% of the 5,120 form two students and 80 secondary schools respectively.

2.2 Research Instruments

Data was collected from students, teachers and parents through questionnaires, oral interviews and focus group discussions analysis of documents. In this study data was collected using Parental attachment questionnaires to measure the various parental attachment styles, deviant behavior variety scale for measuring violence and drug use, Interest, Emotions and Relationships Scale for measuring irresponsible sexual behaviour, interview schedules, focus group discussions and document analysis. These surveys did not request for sensitive personal specific information from participants and therefore rights to confidentiality were taken care of. All the adapted instruments were piloted to ascertain the clarity of the questions and phrases used before being used on the real sample of the study.

2.3 Pilot Study

The tools were piloted in order to detect any ambiguity and correct them. The researcher administered a set of structured questionnaires and interview guides to students and teachers not included in the sample schools. This helped to examine the appropriateness of the instruments and to get a rough estimate of the time required for the study. This further helped the researcher to make any necessary changes to the instruments with the help of experts from MMUST. For example, specific words which were misinterpreted by the pilot participants were replaced by more familiar words. The interview schedules were flexible enough and ensured that no relevant information was left out. To evaluate the worth of a study, its trustworthiness must be established. Lincoln and Guba (2005) assert that trustworthiness is established when findings of the study reflect the true situation described by the respondents. They propose a variety of strategies to ensure trustworthiness. These include credibility- trust in the findings, transferability showing that the findings can be relevant in other contexts, dependability – this also refers to the consistency of the findings, confirmability- ensuring that the findings reflect the responses of the participants and not the wishes of the researcher (Shenton, 2004).

2.4 Reliability of the instruments

Reliability is a measure of the consistency of the results produced by a research instrument after repeated measurements are taken of the same subjects under similar conditions (Ngwiri et al,

2016). Reliability of the questionnaires was determined by the internal consistency method which demands that the instrument or test to be run once only through the split half method contrary to the test retest method and the equivalent forms reliability which require the test or the instrument to be administered twice. The split half method eliminates chances of error due to differing conditions. In this research, the test items were divided into two with each half matched in terms of items or item difficulty and content. The halves were marked separately. The marks obtained in each half was correlated with the other. Any student's mark on one half should match his or her marks in the other half. This was calculated using the Spearman's- Brown formula.

This calculation required a correlation coefficient to be calculated for example, Spearman's rank order or a Pearson's product moment correlation. Kothari (2004) argues that a reliability coefficient of 0.6 will be considered appropriate similar to Mugenda and Mugenda (2003) who also argue that a reliability coefficient of 0.80 or more means there is a high degree of reliability. The calculation of the split half coefficient was done with all the questionnaires and the results indicated that parental attachment questionnaire 0.763, deviant behavior variety scale 0.764, and finally IERS 0.689. The measurement procedure was considered to demonstrate split half reliability since the two sets of scores were highly correlated.

2.5 Data Collection Procedures

The researcher obtained a letter authorizing her to conduct the field study from the faculty of education MMUST. The letter also introduced the researcher and the study to be undertaken. A permit to carry out the study was also obtained from the National Council of Science Technology and Innovation (NACOSTI). The researcher developed a work plan and piloted the instruments in schools not included in the study. These steps enabled the researcher to effectively carry out the research and the desired sequencing of the steps Kothari, (2004). The permit was presented to the head teachers of the selected schools. It was also used to brief the respondents in order to promote trust with them. The researcher visited respective schools to establish rapport with respondents in preparation for data collection. This helped the researcher to determine a suitable time to administer the questionnaires. The head teachers were the entry point to the selected schools and they introduced the study to the teacher counselors, deputy head teachers, selected members of PA and students who were informed of the study and given research consent forms. The researcher then administered the questionnaires in person. This ensured 100% return rate of the questionnaires.

2.6 Collection of qualitative data

Qualitative data was collected through individual interviews, focus group discussions and analysis of documents. The interviews were audio taped so that information about participants' lived experiences would be accurate (Creswell, 2012). Field notes such as observational notes were used in order to minimize loss of data and to have detailed descriptions of the researcher's observations, reflections and experiences during the research process. Eight focus group discussions each composed of ten students from the schools in the sample were conducted. 32 individual interviews were also conducted as follows; 8 parents, 8 principals, 8 deputies and 8 HODs. The students in the focus groups reported their perceptions pertaining to their parents' attachment styles and their behavior. The students also gave their views on how their gender contributed to their behavior. The parents were asked about their attachment to their children and how this contributed to the behavior of their children. Teachers also gave reports on how parental attachment and gender contributed to the behavior of the students.

III. RESULTS AND DISCUSSION

The findings were presented on the basis of the research objectives, which were to evaluate the relationship between parental attachment and deviant behaviour and to assess gender differences in deviant behaviour among secondary school students. Data was analyzed using descriptive statistics and inferential statistics such as Pearson Correlation and independent t-test. Qualitative data was analyzed using thematic framework.

3.1 Questionnaire Return Rate

The study achieved 100% response return rate in all the questionnaires. This was achieved because the researcher in person issued the instruments to each respondent to ensure that each and every respondent took part in the study.

3.2 Student Respondents' Demographic Characteristics

In terms of gender, majority of the students were boys at 57.42% while the girls were represented at 42.58%. This is captured in Table 1 below.

This implies that the gender distribution was even. There was approximately 3:2 gender split for males and females.

The first objective of the study was to evaluate the relationship between parental attachment and deviant behaviour among secondary school students in Homabay County. This was achieved by testing the null hypothesis that there is no statistically significant relationship between parental attachment and deviant behaviour among secondary school students in Homabay

County. A two-tailed bivariate Pearson correlation analysis coefficient was computed to evaluate the relationship between parental attachment and deviant behaviour and the results were as shown in Table 2 below.

The results in Table 2 indicate a significant negative correlation between parental attachment and deviant behavior among secondary school students in Homabay County, with (r=-.244, p<0.05). This implies that as parental attachment improves, deviant behavior reduces.

The second objective of the study was to assess gender differences in deviant behavior among secondary school students in Homabay County. This was achieved by testing the null hypothesis that there is no statistically significant gender difference in deviant behavior among secondary school students in Homabay County. An independent sample t-test was run to compare the means between the boys and the girls and the results were as shown in Tables 3 and 4 below.

The findings in table 2 revealed that the male learners had a higher mean of deviant behaviour score of 1.93, with a standard deviation of 0.685 and standard error of 0.076 than the female learners who had a mean of 1.30, with a standard deviation of 0.524 and a standard error of 0.027.

The results indicate that there was a significant difference in the scores for males (M=1.93, SD=0.685) and females (M=1.30, SD=0.524); t (512)=9.194, p<.0.05. These results suggest that gender contributes to the deviant behavior of the form two students. Specifically, the results indicate that boys are more deviant than the girls. From the results of the study, it was credible to conclude that there is statistically significant gender difference in deviant behavior of form two students in Homabay County. These findings corroborate to those of Moitra and Mukherjee (2010) who revealed that adolescent boys who do not experience any kind of love from their mothers are likely to exhibit problematic behavior. In addition, Yifru and Asres (2015) also found a significant association between male youths' SES and the likelihood that they practice risky sex.

Others who also concur with the findings of this study include Borsuk and Juhnke (2015) who reported that women are more vulnerable to prescription Opioids misuse than men. Similarly,

Jules et al (2015) in a purely qualitative study which employed a cross sectional design and was conducted among university students revealed that males and females differ significantly in terms of illicit drug use. Tsvetkova and Antonovna (2013) reported higher social acceptability of drug use among males than females. Their study was carried out among university students and data only analyzed descriptively.

In addition, Onukwufor (2013) noted that physical and verbal aggression was higher among males than females. Finally, Odimegwu and Somefun (2017) who analyzed secondary data collected from youths out of school reported that males have more sexual partners than females. Similarly Kabiru (2014) also revealed that gender determines delinquency.

3.3 Qualitative Data Analysis

Qualitative data revealed varied views concerning the study objectives which led to the emergence of various themes

During the focus group discussion sessions high involvement in deviant behavior among the male and female students was reported. It emerged that some aspects of violence, drug use and irresponsible sexual behavior were driven by gender differences as suggested by the comments below.

Whenever there is a disco matanga boys are allowed to go but girls will only go if they sneak.

Boys who sleep in their brothers' houses sneak in girlfriends for sex from the discos. Girls who sleep at the grandmothers' houses go for night dances where a lot of sex talk and activities go on. I don't think they use condoms [Girl, FGD].

The statement above implies that many boys are deviant as compared to the girls. This is because they have a lot of freedom, and when they go to the disco matanga they watch a lot of violence and initiate sexual relationships with the girls whom they find there. From the above statement there are more boys than girls at the discos. Another participant also reported that he was told by a friend that at the disco matanga they sell locally brewed alcohol and cigarettes.

Some girls also confessed that at times they do not intend to have sexual intercourse but they are coerced by their boyfriends who threaten to abandon them if they don't cooperate. Whereas female students engaged in sexual intercourse through coercion, for monetary benefits and sense of belonging, the male students sought to prove their masculinity. Almost 90% of the boys believed that real love can only be expressed through sexual intercourse. This is suggested by the following statements; "I cannot refuse him because I love him and he gives me financial support and so I do not want to lose him"[Girl, FGD] "Without that (sex), then there is no need for the relationship" [Boy FGD]

When asked whether they have fought in school and the reasons behind the fight, most of the respondents agreed with the majority of the culprits being boys. Boys either fought over a girl or when they felt they have been ridiculed by a colleague while girls engaged mostly in verbal aggression over petty issues which most of them could not define.

"I hate her and I will still fight her if I find the opportunity. She thinks she is prettier than everybody."[Girl FGD].

"You know our school is mixed and so when someone demeans you before the girls, the girls will laugh at you. I can't take that." [Boy FGD].

Both parents and teachers also agreed that although both boys and girls exhibit deviant behavior, more boys exhibit deviant behavior as compared to the girls. This was reported by six parents and five teachers. For example, Grace (pseudonym) has little or no communication with her parents since she lives with her grandmother because the parents are not at home. She attends almost all the discos in the area. She has a child at home. She is very rude to teachers and fellow students and sleeps a lot in class.

IV. SUMMARY

The purpose of this study was to assess gender differences in deviant behaviour among secondary school students in Homabay County. The independent variable was gender while the dependent variable was deviant behaviour. The objective to guide the study was developed which stated; to assess gender differences in deviant behavior among secondary school students. There was an elaborate review of related literature from the world, Africa and Kenya. From the reviewed literatures a lot of gaps were identified which the study has elaborated how they were filled in the current study. The study adopted a correlational survey design and data analysis was done both qualitatively and quantitatively. Target population consisted of 315 schools with a population of 20,160 form two students from within Homabay County. Out of 8 sub counties in Homabay County, two sub counties with a population of 80 schools and 5, 120 form two students were sampled using the simple random sampling technique being 30% of the sub counties. From the schools within the two sub counties, 8 schools were sampled randomly being 10% of the schools and 512 students being 10% of 5,120 students. Principals, deputies, HODs and parents were purposively sampled. The schools in the sample were stratified into boys' boarding, girls' boarding and mixed day and boarding.

4.1 Conclusion

Quantitative data was collected using questionnaires and analyzed using descriptive statistical analysis as well as inferential statistics such as Pearson correlation analysis and independent t-test analysis. Qualitative data was collected using structured and unstructured interview guides, FGD and analysis of documents. The documents analyzed included class registers, students' files and house registers incase of boarding schools. Qualitative data was analyzed using thematic framework. The findings proved that gender plays a significant role in deviant behavior as the t-test analysis revealed a statistically significant difference in deviant behavior between the males (M=1.93, SD=0.685) and females (M=1.30, SD=0.524); t(512)=9.194, p<.0.05). Analysis of qualitative data proved the presence of association between the independent variables and the dependent variable. The participants confirmed that the attachment between the parents and their children influence the behavior of the children. Similarly, participants confirmed that there is a difference in deviant behavior which is determined by gender. It emerged that the boys in this region are given excess autonomy and freedom which most of them do not use responsibly. For example boys engaging in irresponsible sexual behaviour in order to satisfy their ego. This could be as a result of the belief that the boys are strong/protectors of the home and therefore are expected to take care of themselves even while still young. This gives them a lot of freedom which most of them misuse to their detriment.

4.2 Recommendations

Since the findings of the study revealed a significant gender difference in the behaviour of the secondary school students, the study therefore recommends that the residents of Homabay County should be made aware of the need for a collaborative effort by all the stakeholders to help eradicate deviance among our students and to shun the retrogressive cultures which give more preference and autonomy to the boy child and take full responsibility in ensuring that the boy child is properly disciplined and also give the girl child equal preference.

The study also recommends a public awareness conducted in homabay county to sensitize the parents on the need to cultivate an appropriate attachment with their children as this would help reduce deviance among the secondary school students.