International Journal of Educational Development 34 (2014) 30–41

Factors determining educational quality: Student mathematicsachievement in Nepal

Krishna Prasad Pangeni *

Graduate School of International Development, Nagoya University, Japan

A R T I C L E I N F O

Keywords:

Family characteristic

Student characteristic

Parental involvement

School characteristic

Mathematics achievement

Nepal

A B S T R A C T

This study employed the multiple regression analysis to examine the effects of family characteristics,

student characteristics, parental involvement, and school characteristics with regard to learning

achievements in mathematics of students of eighth grade in Nepal. Mathematics test was administrated

with 762 students (400 girls) of 21 secondary schools in nine districts of three ecological zones-

mountain, hill and lowland (tarai). Findings indicated that family characteristics explained the largest

amount of variance in mathematics achievement, followed by student characteristics, school

characteristics and parental involvement. Variables – parental education, number of books at home,

absenteeism, parental support for homework, school type and location, teacher training, the number of

school days in the academic year and school physical facilities were significantly related to students’

mathematics achievement. The implications of this study for improving educational quality are

discussed.

� 2013 Elsevier Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

International Journal of Educational Development

jo ur n al ho m ep ag e: ww w.els evier . c om / lo cat e/ i jed u d ev

1. Introduction

Quality of education associated with high learning achievementimproves the quality of human resources and is directly related toincreased individual earnings and productivity, economic growthand governments’ ability to alleviate poverty (UNESCO, 2005).Although educators and researchers have yet to agree upon aconsensus on the nature of educational quality and its determinants,it is typically measured by higher achievement in examinations(Chapman et al., 2005). Learning is a product of families,communities and peers as well as formal institutions (Rothstein,2000), and social, economic and cultural factors influence the extentto which family members support children’s classroom learning.Coleman et al. (1966) claimed that family characteristics were moreimportant than formal education for children’s learning in theUnited States. However, based on data from 29 developed anddeveloping countries, Heyneman and Loxley (1983) found that theoverall proportion of variance in student achievement was largelyassociated with school characteristics as compared to studentbackground characteristics particularly in less industrialisedcountries and that ‘‘the predominant influence on student learningis the quality of schools and teachers to which children are exposed’’(p. 1162). In addition to the effects of family and school

* Correspondence address: 465-0065, Nagoyashi Metoku Umemorizaka 4-101,

Umemorisou 24-707, Japan. Tel.: +81 9099482641.

E-mail address: kp_pangeni@yahoo.com

0738-0593/$ – see front matter � 2013 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.ijedudev.2013.03.001

characteristics, child characteristics such as motivation, academiceffort, involvement in household chores, homework and genderinfluence academic outcomes (Konu and Rimpela, 2002; Keith et al.,1986; Veenstra and Kuyper, 2004). Moreover, parents who providesupport and opportunities to their children to explore and expressideas make a vital contribution to their academic achievement(Fehrmann et al., 1987; Astone and McLahanan, 1991; Iverson andWalberg, 1982; Barnard, 2004).

Adopting a theoretical framework, the present study examinedhow family characteristics, student characteristics, parentalinvolvement and school characteristics influenced students’mathematics achievement in Nepal. More specifically, the researchfocused on three issues: determining the various factors thatinfluenced mathematics achievement of grade eight students inNepal; identifying the factors that primarily affected academicachievement; and determining the extent to which differences inachievement are associated with family characteristics, studentcharacteristics, parental involvement and school characteristics.This study differs from earlier studies as it examined thecontribution of students’ caste/ethnic group and preschoolexperience in addition to the effects of family characteristics,student characteristics, parental involvement and school char-acteristics on academic achievement simultaneously, which is rarein such studies and is very innovative in the context of Nepal.

A multi-stage random stratified sampling method was used.Initially, two stages – ecological zones and districts – were identified.Based on the proportion of students from three ecological zones –mountain, hill and lowland (tarai) – nine districts were selected: one

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–41 31

mountain district, four hill districts and four lowland (tarai) districts.Altogether, the sample consisted of 21 secondary schools with 762eighth grade students (400 girls). Students’ academic achievementwas measured by a mathematics test based on the nationalcurriculum. Participating students received questionnaires thatprovided information regarding personal characteristics, familycharacteristics, and the extent of parental involvement in schooling.In addition, information on school characteristics was obtained from21 head teachers and 21 mathematics teachers. To identify thefactors underlying mathematics achievement, a multiple linearregression analysis was performed that tested four increasinglycomprehensive models.

The structure of this paper is as follows. The Section 2 describesthe cultural context of education in Nepal. Section 3 discusses thetheoretical framework of the study, which primarily focuses on therelationship of learning outcomes to family characteristics, studentcharacteristics, parental involvement and school characteristics.Section 4 presents the study methodology and describes the studyinstruments, sample and dependent and independent variables.Sections 5 and 6 present data analysis and discussions. Section 7presents study conclusions and discusses the implications of studyfindings.

2. The cultural context of education in Nepal

Nepal is a small landlocked country in southern Asia with adiverse population of 26.6 million that consists of 92 differentlanguage groups, 103 caste/ethnic groups and seven religiousgroups (CBS, 2011). Geographically, the country is divided intothree distinct ecological regions. The mountain region at analtitude between 4877 and 8848 m comprises 35 percent of theland area, hills at an altitude between 610 and 4876 m comprises42 percent of the land area, and the tarai region, which is anextension of the Gangetic Plains, lies below an elevation of 610 mand comprises 23 percent of the land area. For administrativepurposes, there are five development regions, 14 zones and 75districts. Currently, poverty is pervasive in Nepal. The Nepal LivingStandards Survey (NLSS, 2011) estimated that 25.1 percent of thetotal population is below the poverty line, and the GDP per capita isapproximately US$ 480. In 2011, Nepal was classified in themedium human development category and ranked 157th (with avalue of .458) on the Human Development Index (HDI) and 68th onthe Human Poverty Index (HPI). Nearly 83 percent of thepopulation lives in rural areas and the adult literacy rate (15years and above) is only 56.64 percent (NLSS, 2011).

With the establishment of democracy in 1990, the governmentof Nepal considered education as a priority and committed toprovide quality education to all, especially girls and disadvantaged.The share of the gross domestic product (GDP) devoted to theeducation sector increased from 2.1 percent in 1990 to 3.3 percentin 2009; during the same period, overall education expendituresincreased from 13 to 17.5 percent. Despite a decade-long conflict,several projects and programs, such as the Basic and PrimaryEducation Project I (1992–1998), the Basic and Primary EducationProgram II (1999–2004) and Education for All (EFA, 2004–2009)were implemented, leading to greater participation of children andstudents in formal education at all levels. The net enrolment ratio(NER) at the primary level increased from 67 percent in 1990 to 95percent in 2011 (DOE, 2011). However, improving learningachievement continues to present daunting challenges as manychildren and young people leave school without basic skills andknowledge needed to improve their standard of living, functioneffectively in society and develop their potential. Assessmentsperformed by the Research Centre for Educational Innovation andDevelopment (CERID, 1999) and by the Educational DevelopmentService Centre (EDSC, 2008) revealed poor learning outcomes in

core subjects, including mathematics. Average achievement scorein mathematics of grade eight students was 28.8 in 1999 and 31.7in 2008. However, prior researches have not investigated thefactors that influence learning achievement of students in Nepal.

Recently, the country has experienced significant politicalchange, and the transformation from a constitutional monarchy toa democratic republic has raised people’s expectations andincreased the demand for quality public education. To respondto the public demand for educational quality and opportunity, theSchool Sector Reform Plan (SSRP, 2009–2015) is in the fourth yearof implementation. The plan aims at restructuring of schooleducation between basic (grades 1–8) and secondary (grades 9–12) education. Grade eight is the final basic education grade leveland it is supposed that grade 8 graduates achieve functionalliteracy and numeracy skills to fulfill the requirement of society.

The interim constitution of Nepal 2007 emphasizes onproviding free education up to secondary level. In this context,it is important to examine learning outcomes of eighth gradestudents in mathematics. Currently, two hundred thousandteachers teach eight million children in a network of over 34thousand schools; 11 percent of these schools are privatelymanaged. The relevant age groups are 3–4 years for preschool/early childhood education and development (ECED), 5–12 yearsfor basic education (5–9 years for primary and 10–12 years forupper primary education) and 13–16 years for secondaryeducation. The Ministry of Education (MOE) is the chief bodyresponsible for formulating education policies, while theDepartment of Education (DOE) supports central level agenciesand the District Education Offices (DEO) in implementingprogrammes. Regional Education Directorates (REDs) are respon-sible for monitoring educational activities at the regional level,and the District Education Offices (DEOs) are primarily responsi-ble for implementing educational programmes and managingschools at the district level with the support of School Supervisors(SS) and Resource Persons (RP). The Resource Centre (RC) is thelowest institutional agency at the district level that providesacademic and technical support to schools.

3. Factors associated with educational quality: the theoreticalframework

Many studies have investigated the factors that contribute tostudent learning outcomes across contexts. However, due toconflicting research findings consensus has yet to be achievedregarding the factors that determine student academic achieve-ment. The effect of family socio-economic status (SES) is one ofthe most frequently examined factors because the family plays acrucial role in shaping children’s educational experiences andacademic achievement in every society. A widely accepted viewis that SES consists of the financial, social and cultural capital thatis transmitted from one generation to next (Buchmann, 2002).Parents with higher levels of SES are more likely to have financialand social resources to invest in education (Becker, 1964), andchildren in these families are more likely to succeed in school(Buchmann, 2002) and exhibit higher levels of performance(Veenstra and Kuyper, 2004). In the United States, Coleman et al.(1966) examined national survey data and found that familycharacteristics were more important than school characteristicsin determining academic outcomes. SES as measured by parents’education, occupation and income, significantly influencesstudents’ educational attainment and achievement (Colemanet al., 1966; Cameron and Heckman, 2001; Schiller et al., 2002)because these students have access to greater economic andsocial resources (Coleman, 1988; Gregg and Machin, 1999;McNeal, 1999) to support academic success compared to familieswith lower levels of parental education (Schiller et al., 2002).

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–4132

Parents with higher levels of education have greater access to awide variety of economic and social resources (family structure,home environment, parent–child interaction) that are used tohelp their children succeed in school. Parents’ and grandfathers’social status and literacy were found strongly related to children’school attendance in Nepal (Jamison and Lockheed, 1987). Usingthe Turkish Third International Mathematics and Science Study(TIMSS) data Yayan and Berberoglu (2004) claimed that whenparental education and number of books at home increased,eighth grade student achievement in mathematics also increased.Similarly, Johnston and Thompson (2006) analyzed the Pro-gramme for International Student Assessment (PISA) results andfound that students at the highest SES levels, as measured bynumber of books in a student’s home, had an educationaladvantage over students at the lowest SES levels. Generally, indeveloping countries, fathers’ occupations and level of educationexhibit little variation because parental education is uniformlylow and occupations are predominantly agriculture. However,social class differences (e.g. land ownership, literacy, caste anddistrict of residence) were strongly associated with schoolattendance in Nepal (Stash and Hannum, 2001; Heyneman,2002). PISA results indicated that ‘‘poor performance in schooldoes not automatically follow from a disadvantaged socio-economic background, despite the fact that socio-economicbackground does appear to be a powerful influence on perfor-mance’’ (OECD, 2003:191).

Several studies have indicated that increased number ofchildren in the family is negatively related with children’s schooloutcomes because of the mechanism of resource dilution (Blake,1989; Patrinos and Psacharopoulos, 1995; Downey, 2001).Resource dilution refers to the quantity of time and material thatparents are able to invest in their children (Teachman et al., 1996);when the number of children increases, parents can offer fewerresources per child. Under such condition, all forms of familycapital-financial, human and social – are spread across the children(Coleman, 1991). Empirical evidence supports these claims:children from larger families have been found to have lessfavourable home environments and lower levels of verbal facility(Parcel and Menaghan, 1994) as well as higher rates of behaviourproblems and lower levels of educational achievement (Downey,1995).

Children’s personality and self-concept are related to schoolperformance. Many studies have investigated the relationshipbetween academic achievement and children’s motivation tostudy, perception of the school environment, gender, age,absenteeism, grade repetition, homework completion, involve-ment in scholastic activities, household chores and preschoolexperience as student characteristics. Academic motivation refersto the cognitive, psychomotor and affective psychological process-es that influence pupils’ learning (Slavin, 1997), guide and sustainstudy behaviours (Schunk, 1990), contributing to foster students’engagement in academic activities. Motivated learners are morelikely to use higher cognitive processes to learn, absorb and retaininformation and try to improve their performance despite thepossibility of failure.

Gender is another important variable to be considered inexplaining variations in school achievement. Using the data of theSouthern and Eastern Africa Consortium for Monitoring Educa-tional Quality (SACMEQ), Kulpoo (1998) in Mauritius found thatgirls outperformed boys in mathematics although gender differ-ences were also associated with school levels and subjects (Farkaset al., 1990; Kutnick, 2000; Woessmann, 2003; Hungi, 2008; Hungiand Thuku, 2010). Regular student attendance is crucial forlearning as it affects academic success (Gray and Partington, 2003;Gray and Beresford, 2008). In Vietnam, Hungi (2008) found thatstudents who were never (or were rarely) absent exhibited higher

levels of mathematics and reading achievement compared tostudents who were frequently absent from school.

Grade repetition is another characteristic that negatively affectsstudent self-esteem and academic motivation and is associatedwith long-term problems such as poor academic performance,behavioural problems and dropping out of school (Hungi andThuku, 2010). In Nepal, research performed by Save the Children(2003) found that students with preschool experience had a 26percent higher rate of promotion from grade 1 to grade 2 thanchildren who had not attended preschool (UNESCO, 2003, p. 182).Student study behaviours – in particular, time spent on homework– is associated with greater academic success (Keith et al., 1986).Students with access to textbooks and reference materials alsoperform better than students with access to fewer materials(Veenstra and Kuyper, 2004). Akabayashi and Psacharopoulos(1999) in Tanzania, Heady (2003) in Ghana and Ray and Lancaster(2003) in India found that children’s involvement in economicactivities or household chores was also significantly associatedwith reading and mathematics achievement.

Parents are the most important individuals in children’s lives,they influence children’s development by providing opportunitiesfor them to explore, imagine and express themselves (CERSOD,2005). As their understanding expands, children observe thebehaviours of parents, siblings and other adults and relate it totheir learning. At this stage of development, parents provide vitalsupport for children’s learning while continuing to serve as rolemodels at home. Parents with lower SES are less involved in theirchildren’s education than the parents with higher SES (Sui-Chu andWillms, 1996); however, parental verbal stimulation and support-ive behaviour can mitigate the effects of SES (Macbeth, 1994).Children are more likely to attend school regularly when theirparents encourage and monitor school attendance and create ahome environment conducive to higher academic performance(Teachman, 1987). Parents’ involvement in their children’seducation fosters positive attitude towards school, increasedhomework completion, reduced absenteeism, lower levels ofdropping out, and enhanced academic achievement (Astone andMcLahanan, 1991; Epstein, 1987; Fehrmann et al., 1987; Keithet al., 1986; Lareau, 1987; Stevenson and Baker, 1987).

The preeminent role of the school system as the institution thatformally serves to provide instruction to children began to bedebated after Coleman et al. (1966) found that family character-istics contributed more to academic success than school char-acteristics in the United States. However, Heyneman and Loxley(1983) stressed the importance of the school, particularly indeveloping countries. Hanushek (1986) argued that the traditionalapproach of simply providing resources is frequently insufficientand that differences in educational quality between schools are notdue to differences in expenditures, class size, and other commonlymeasured attributes of teachers. Although the quality of resourcesis important, improving academic outcomes requires effectiveeducational processes (Chapman and Adams, 2002) and educa-tional quality is not merely due to interactions between subjectsand objects but influences the interactions between subjects andobjects (Scheerens, 2004).

Among the various school characteristics, student teacher ratiohas been a widely examined variable; however findings regardingthe effects of student teacher ratio on school achievement areinconsistent. Hanushek (1997) claimed that class size was notimportant, but Woessmann (2003) found that smaller class sizewas significantly related to inferior student performance in mathand science on the TIMSS in 39 countries in 1994/1995, whereasLindhal (2005) found that minority and economically disadvan-taged children benefited from smaller class sizes in Sweden. Thesefindings were consistent with the results of Angrist and Lavy(1999) in Israel and Krurger (2003) in the United States. However,

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–41 33

Hoxby (2003) found that class size did not significantly influencestudent achievement in the United States. Kingdon (1999) did notfind a significant relationship but found that class size wasnegatively associated with numeracy skills and positively relatedto reading skills in India.

In addition to the student teacher ratio, instructional materials,teacher training were found important by Willms and Somers(2001) in 13 Latin American countries. Other major factors relatedto educational quality found by Fuller (1987) were instructionaltime spent on subjects, homework assignments, feedback andactively involving students in the classroom. Similarly, in-serviceteacher training, adopting curriculum and teaching materials tolocal practices, community support and supervision, sufficientresources, and motivated school leadership were required forschool success (Dalin, 1994). Verspoor (1989) argued that teachers’knowledge of the subject matter and materials, teachers’expectations regarding student performance, pedagogical process-es, teaching practices, classroom organisation and students’ timeon task in the classroom were associated with educational quality.Similarly, other studies have shown management types of school(Psacharopoulos, 1987; Govinda and Varghese, 1993) and locationof school (Woessmann, 2003) was associated with educationalquality.

4. Methods

4.1. Instruments

A mathematics test, student survey, and questionnaires forhead teachers and mathematics teachers provided the data for thisstudy. The mathematics test was designed to assess the level ofacademic content knowledge that was consistent with the nationalcurriculum goals and was based on discussions with mathematicsspecialists, the mathematics head of the curriculum developmentcentre (CDC) and eighth grade mathematics teachers. The 20-itemmathematics assessment had a maximum possible score of 50, andthe weight accorded to each item reflected the national curriculumgoals. The test consisted of five arithmetic questions, eight algebraquestions, two statistics questions and five geometry questions,and questions assessed the cognitive domains of knowledge (fourquestions), understanding (five questions), skills (six questions)and problem solving (five questions). However, because thepurpose of the present study was to identify the factors influencingmathematics achievement, content and cognitive domains ofmathematics achievement were not analysed. Marking schemewas developed to maintain consistency for correcting answersheets of students. The mathematics assessment was performed atthe end of academic term to ensure that each school had theopportunity to complete the instruction of the mathematicscurriculum.

The student survey consisted of three sections. The first sectionfocused on family characteristics such as the number of familymembers, father’s and mother’s education, travel time to school,the number of books in the home and the presence of certain items(e.g. computers) in the home. The second section focused onpersonal characteristics such as gender, the number of days absentfrom school, grade repetition, homework completion, the numberof hours spent studying, the number of hours spent on householdchores, preschool experience and academic motivation. The finalsection focused on students’ perception of their parents’ involve-ment in their education.

Questionnaires were also administered to head teachers andmathematics teachers to obtain information about the school andtheir personal and professional characteristics. To ensure thereliability of the measures, pilot testing was performed at a publicsecondary school in the Bardiya district in the mid western region

of Nepal, which was not included in the study sample. Cronbach’salpha was used to assess the reliability and internal consistency ofthe instruments. For family characteristics, the alpha coefficientwas 73.4, for student characteristics, alpha was 76.4, for parentalsupport, alpha was 86.1, and for school characteristics, alpha was70.3; the overall reliability of the scale was 72.3, which isconsidered acceptable for research purposes (Muijs, 2004, p. 72).

4.2. Sample

A multi-stage random stratified sampling method was used.Initially, two stages – ecological zones and districts – wereidentified. Based on the proportion of students from threeecological zones – mountain, hill and lowland (tarai) – ninedistricts were selected: one mountain district, four hill districtsand four lowland (tarai) districts. Altogether, the sample consistedof 21 secondary schools with 762 eighth grade students (400 girls).Schools within the districts were selected purposefully takingaccount of their characteristics such as public and private, ruraland urban and so on. To ensure that schools were representative ofall five regions of the country, three schools were selected from thefar western region, seven schools from mid western region, twoschools from the western region, seven schools from the centralregion and two schools from the eastern region. The type ofschool (public or private) and school location (rural or urban)was also included because a current educational issue in Nepalconcerns the extent to which differences in academic achieve-ment are related to school characteristics. Of the 21 schoolsin the sample, four were private schools and 10 were locatedin urban areas (i.e. district headquarters). One head teacher andone mathematics teacher from each school participated in thestudy.

4.3. Dependent and independent variables

The outcome variable in this study was students’ mathematicsscore. Based on theoretical and empirical considerations, four setsof variables – family characteristics, student characteristics,parental involvement and school characteristics – were identifiedas independent variables for this research. Family characteristicsconsisted of the number of family members, father’s level ofeducation, mother’s level of education, travel time to school, thenumber of books in the home and the presence of certain items inthe home. Travel time to school was also included because in Nepal,students in different ecological zones have different commutingtimes to school that might influence academic outcomes.

The second independent variable was student characteristics:gender, ethnic group, absenteeism, grade repetition, homeworkcompletion, time spent on studying, time spent on householdchores, preschool experience, perception of the mathematicsinstructor and academic motivation. The third independentvariable was children’s perception of parental involvement intheir schooling, which primarily consisted of parental support forhomework, students’ perception of parental support and parentalmonitoring. The final independent variable was school character-istics, which consisted of school type, school location, teachertraining, number of school days in the academic year, student–teacher ratios, and school physical facilities. Definitions anddescriptive statistics for the study variables are presented inTable 1.

5. Data analysis

Data analysis is presented in three sub-sections. The firstsection describes the distribution of scores in mathematics; secondsection explains the differences in mathematics achievement

Table 1Variable definitions and descriptive statistics (N = 762).

Variables Variable definitions Percent Mean SD

LEARACHIV Children’s mathematics test score 11.55 9.28

NUMFAM Number of family members in the household 6.23 2.05

EDUFATH Father’s education, 1 = secondary and above, 0 = otherwise 43.4

EDUMOTH Mother’s education, 1 = secondary and above, 0 = otherwise 24.8

TRVLTIMESCH Travel time to school 19.34 11.14

ADDBOOKHOM Number of books at home, 1 = 1–5 books, 2 = 6–10 books, 3 = more than 10 books, 4 = no books 2.74 1.17

MATERLHOMa Presence of certain items in the home, available = 1, not available = 0 5.05 2.46

GENDER Child sex, 1 = male, 0 = female 47.5

ETHNSTU Child ethnicity, 1 = high caste, 2 = middle caste and 3 = low caste

GREREP Number of grade repetitions since school admission 1.39 .62

ABSNT Number of days absent during the past month 2.83 4.66

HOMWORK Complete homework, 1 = never do, 2 = sometimes do, 3 = everyday 2.71 .56

STUTIMHOM Number of hours spent studying 2.68 .85

STUPERTEACH Student perception of teacher, 1 = difficult, 2 = easy, 3 = very easy 1.97 .71

HOUSWORK Number of hours spent on daily household chores 2.34 .75

EXPECD Preschool experience, 1 = yes, 0 = otherwise 43

STUMOTIVA Student motivation to study, 1 = strongly disagree to 5 = strongly agree 3.88 1.12

PARSUPHOM Parental support of homework, 1 = yes, 0 = otherwise 41

PARENCOURG Parental encouragement for studying, 1 = strongly disagree–5 = strongly agree 4.08 1.08

PAREMONIT Parental monitoring of studying, 1 = strongly disagree–5 = strongly agree 3.59 1.27

TYPSCH Type of school, 1 = public, 0 = Private 84

LOCSCH School location, rural = 1, 0 = urban 46.5

GEOLOCSCH Geographical location of school, lowland = 1, hill = 2 and mountain = 3

TEACTRIN Teacher training, 1 = untrained, 2 = partially trained, 3 = trained 2.42 .77

INSERVTRIN In-service teacher training, yes = 1, otherwise = 0 32.4

TEACHDAYS Number of school days in the academic year 181.7 4.3

STR Student–teacher ratio in a particular grade 60.5 10.9

PHYFACILb School physical facilities, available = 1, not available = 0 7.80 2.34

a Bicycle, radio, television, CD player, water tap, gas stove, bio gas, motorbike, computer.b Library, separate room for head teacher and teacher, playground, drinking water, toilets, electricity, telephone.

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–4134

across gender, ethnicity of student, types of school (public andprivate) and locations. In the final section, a multiple linearregression analysis was performed that tested four models ofrelationship of different sets of variables with students’ mathe-matics outcomes.

5.1. Distribution of scores in mathematics

Fig. 1 illustrates the distribution of scores of mathematicsachievement of eighth grade students in Nepal. The average scorewas 11.55 out of possible 50 points with the skewness of .96,indicating that majority of students positively skewed in the leftlow end.

It explains that majority of students (72.3 percent) scoredbetween 11 and 15 marks which is considered unqualified (orfailed) for the purpose of promotion into the next grade. Only 9.4percent of students scored more than half of the total possible 50marks.

31.1

55.9

72.3

84.3 90.6

94.

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

<5 6- 10. 11 -15. 16 -20. 21 -25 26 -30

Score in mathematics (ou t of 50)

Fig. 1. Distribution of scores in mathe

5.2. Differences in mathematics achievement

Table 2 presents the means and standard deviations of eighthgrade students’ mathematics scores. It indicates that overallmathematics achievement was low however, there were differ-ences in performance related to student gender, ethnic group,school type (public or private), and school location.

5.3. Factors related to students’ mathematics performance

A multiple regression analysis was used to predict relationshipbetween independent variables (family characteristics, studentcharacteristics, parental involvement and school characteristics)and the dependent variable of mathematics performance. Becausethe dependent variable (mathematics score) was continuous, alinear functional relationship with the explanatory variables couldbe assumed (Muiijs, 2004). The hypothesised relationship ispresented in the following equation:

8 97.0 99.1 99.8 100.0

31-35 36 -40 41 -45 >45

Mean 11.55

Skewness 0.96

matics (cumulative percentage).

Table 2Mean and standard deviation of mathematics achievement across gender, ethnicity,

types of school and location.

Categories Mean Standard deviation

Gender

Girls 10.08 8.53

Boys 13.13 9.74

Ethnic group

High caste 15.54 10.27

Middle caste 9.60 8.15

Low caste 8.28 5.70

Type of school

Private 22.31 10.29

Public 9.47 7.44

School location

Urban 13.27 9.88

Rural 9.52 8.01

Ecological zones

Mountain 5.32 5.13

Hill 11.42 7.54

Lowland (Tarai) 12.15 10.50

Mean 11.55 9.28

Boys (mean score = 13.13) outperformed girls (mean score = 10.08). High caste

children (mean score = 15.54) outperformed middle and low caste children (mean

scores = 9.60 and 8.28, respectively). Private school students (mean score = 22.31)

outperformed public school students (mean score = 9.47). Similarly, there were

significant differences in mathematics scores between students of urban and rural

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–41 35

LEARACHIVi j ¼ b0 þ ðb1 NUMFAMi j þ b1 EDUFATHi j

þ b1 EDUMOTHi j þ b1 TRVLTIMESCHi j

þ b1 ADDBOOKHOMi j þ b1 MATERLHOMEi jÞ

þ ðb2 GENDERi j þ b2 ETHNSTUi j þ b2 GREREPi j

þ b2 ABSNTi j þ b2 HOMEWORKi j

þ b2 STUTIMHOMi j þ b2 STUPERTEACHi j

þ b2 HOUSWORKi j þ b2 EXPECDi j

þ b2 STUMOTIVAi jÞ þ ðb3 PARSUPHOMi j

þ b3 PARENCOURGi j þ b3 PAREMONIi j

þ ðb4 TYPSCHi j þ b4 LOCSCHi j

þ b4 GEOLOCSCHi j þ b4 TEACTRINi j

þ b4 INSERVTRINi j þ b4 TEACHDAYSi j þ b4 STRi j

þ b4 PHYFACILi jÞ � � � þ . . .

where LEARACHIV is the outcome variable, learning achievementof children j of school i, b0 is the intercept of regression model andb1, b2, b3 and b4 are the slope of the set of explanatory variables,family characteristics, student characteristics, parental involve-ment and school characteristics of child j of school i, respectively.Variables were entered into models simultaneously because thegoal of the study was predictive rather than explanatory (Cohenand Cohen, 1983). The causal priority of independent variables wasspecified in advance because it is assumed that the set ofindependent variables entered at a given step does not affectthe set of independent variables entered previously (Ma, 1997).The causal priority of the independent variables was based uponthe theoretical considerations discussed in the literature review aswell as the Nepalese context. The causal priorities of the four sets ofvariables were as follows:

(a) Family characteristics influence student characteristics, such astime spent on household chores, time spent studying,homework completion, preschool experience, and academicmotivation.

schools and between different ecological zones.

(b) Family characteristics and student characteristics determineparental involvement in children’s schooling.

(c) Family characteristics, student characteristics and parentalsupport influence school characteristics.

Prior to the multiple regression analysis, correlations betweenstudy variables were examined to assess the extent of multi-collinearity (Muijs, 2004); zero-order correlation coefficientsamong the variables were found to be less than .50.

Similarly, values for variance inflation factors (VIF) rangedbetween 1.15 and 1.62, and all tolerance statistics were greaterthan .3, which indicates that the final model did not exhibitmulticollinearity (Field, 2009). The Durbin–Watson statistic of 1.27was also between one and three, implying that errors in regressionwere independent. Standardised residuals were examined todetect the presence of outliers. In five cases, standardised residualswere greater than .3 but Cook’s distance estimates indicated thatthese cases did not exert undue influence on the regression model(Field, 2009). Furthermore, to test the assumptions of normality,linearity and homoscedasticity, standardised scatterplots of theresiduals were examined to determine whether the residuals werenormally distributed. The residuals were found to exhibit a linearrelationship with the outcome variable, and the variance ofresiduals on the predicted score was the same for all predictedscores.

In the first model (model I), family characteristics weresignificantly related to children’s mathematics performance(R2 = .30.5, adjusted R2 = .299; F(6756) = 1.36; p < .001). Themultiple regression analysis indicated that family characteristicsaccounted for 29.9 percent of the variance in children’s academicoutcomes. The correlation coefficient for the set of familycharacteristics was 55.2. Standardised regression coefficientsfor model I indicated that the number of family members, father’sand mother’s education, the number of books in the home andpresence of certain household items – but not travel time toschool – were significantly related to children’s mathematicsperformance in Nepal.

In the second model (model II), the set of student characteristicswas added to family characteristics, and the regression analysisfound a significant linear relationship with mathematics achieve-ment (R2 = .490, adjusted R2 = .479; F(10,752) = 2.38; p < .001).Student characteristics accounted for 18 percent of the variance inmathematics achievement, with a correlation coefficient of .70.Standardised regression coefficients for model II indicated thatabsenteeism, homework completion, students’ perception of theirmathematics teacher, time spent on household chores, preschoolexperience, academic motivation and gender – but not time spentstudying at home – were significantly related to mathematicsperformance.

In the third model (model III), the set of parental involvementmeasures was added to the model that included family and studentcharacteristics. The linear combination of family characteristics,student characteristics and parental involvement was significantlyrelated to mathematics performance (R2 = .513, adjusted R2 = .50;F(3759) = 2.39; p < .001). The regression correlation coefficient forthe model III was .71.6 and accounted for 2.1 percent of thevariance in children’s mathematics performance. Parental home-work involvement, support and monitoring were significantlyrelated to students’ academic outcomes.

The comprehensive model (model IV) that combined schoolcharacteristics with family characteristics, student characteristicsand parental involvement was independently and linearly relatedto students’ mathematics performance (R2 = .601, adjustedR2 = .586; F(8754) = 5.93; p < .001). The multiple regressioncorrelation coefficient for model IV was .775, which accountedfor 8.6 percent of the variance in students’ academic outcomes. The

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–4136

type of school (public or private), school location (rural or urban),geographical location (mountain, hilll, lowland/(tarai), teachertraining, in-service teacher training, number of school days,student–teacher ratios and school physical facilities were signifi-cantly related to students’ mathematics performance.

6. Discussion

This section discusses the findings presented in Table 3. First,the relationship of family characteristics to students’ academicachievement is discussed, followed by discussion of the relation-ship of student outcomes to student characteristics, parentalsupport and school characteristics.

6.1. Family characteristics

Family characteristics accounted for 29.9 percent of thevariance in students’ mathematics performance, and childrenfrom homes with fewer family members, father’s and mother’seducation at the secondary level or above, more books and morepossessions at home were likely to outperform their counterparts.Having a father or mother with a secondary or higher level ofeducation was positively associated with mathematics perfor-mance. The analysis indicated that an increase of one standarddeviation in mother’s education increased the child’s academicachievement by .220 (p < .001); similarly, an increase of onestandard deviation in the father’s level of education increased thechild’s academic achievement by .170 (p < .001), other things areheld constant. The effect of mother’s education was greater thanthe effect of father’s education, although fewer mothers (24

Table 3Standardised regression coefficients for family characteristics, student characteristics,

Variables Model I Model II

b t b t

NUMFAM �.118*** �3.750 �.059** �EDUFATH .170*** 4.647 .067**

EDUMOTH .220*** 5.916 .163***

TRVLTIMESCH �.048 �1.527 �.061** �ADDBOOKHOM .240*** 7.224 .133***

MATERLHOM .090** 2.516 .040

GENDER .083**

ETHNSTU �.096*** �GREREP �.050* �ABSNT �.185*** �HOMWORK .126***

STUTIMHOM .036

STUPERTEACH .153***

HOUSWORK �.061** �EXPECD .086**

STUMOTIVA .135***

PARSUPHOM

PARENCOURG

PAREMONIT

TYPSCH

LOCSCH

GEOLOCSCH

TEACTRIN

INSERVTRIN

TEACHDAYS

STR

PHYFACIL

Multiple R 55.2 7R2 30.5 4Adjusted R2 29.9 4Variance 29.9 1

N = 762.* p < .010.** p < .050.*** p < .001.

percent) had completed secondary education compared to fathers(44 percent). In Nepal, mothers are less likely to be employed thanfathers, regardless of their level of education, which might providemothers with more opportunities to spend time on their children’sschooling; mothers who are more educated are also likely to havehigher SES, which would benefit their children. This result isconsistent with the findings of other studies of the effects of familySES on children’s educational outcomes (Woessmann, 2003; Fuchsand Woessmann, 2004; Guncer and Kose, 1993; McEwan andMarshall, 2004; Schiller et al., 2002; Willms and Somers, 2001).

There was a negative relationship between family size andstudents’ academic achievement. Children in smaller families (3–5family members) outperformed children in larger families (�.118,<000). The effect of family size was diminished for private schoolstudents compared to public school students and greater forstudents in rural schools compared to urban schools. The resourcedilution theory (Heer, 1985), which explains the effect of familystructure, claims that children benefit less from available resourcesin larger families. Children in smaller families have greater accessto resources and intellectual contexts that provide academicmotivation (Blake, 1989), such as economic resources (Olneck andBills, 1979), parental aspirations and parental support (Behrmanand Taubman, 1886). However, Buchmann and Hannum (2001)found that the number of siblings did not affect parents’ focus onchildren’s schooling in developing countries.

Because Nepal has a diverse geography, students living in themountain and hill areas must spend more time travelling to schoolthan students in lowland areas (tarai). There was a negativerelationship between travel time to school and children’s academicachievement, however the relationship was not statistically

parental involvement and school characteristics.

Model III Model IV

b t b t

2.126 �.056** �2.027 �.032 �1.24

2.043 .034 1.032 .001 .032

4.975 .139*** 4.289 .069** 2.17

2.191 �.058** �2.116 �.067* �2.63

4.492 .122*** 4.180 .106*** 3.97

1.271 .031 .989 .018 .59

3.044 .072** 2.666 .063* 2.47

3.377 �.095*** �3.418 �.076** �2.94

1.755 �.042** �1.497 �.047* �1.81

6.514 �.169*** �6.037 �.153*** �5.97

4.492 .099*** 3.539 .067* 2.56

1.300 .023 .840 .006 .243

5.252 .126*** 4.327 .090*** 3.31

2.106 �.056** �1.973 �.022 �.82

3.098 .052* 1.843 .037 1.39

4.645 .102*** 3.268 .068* 2.38

.131*** 3.991 .145*** 4.80

.072** 2.040 .058* 1.79

.083** 3.092 .049** 1.98

�.120** �2.77

.169*** 5.08

�.152*** �4.93

.130*** 3.65

.077** 2.80

.119*** 3.72

�.102** �2.45

.089** 2.33

0.0 71.6 77.59.0 51.3 60.17.9 50.0 58.68.0 2.1 8.6

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–41 37

significant. When this measure was combined with other familycharacteristic measures in the model IV, the variable travel time toschool was significant and consistent with the findings of Hungi(2008) in Vietnam. The finding that the number of books at homewas significantly related to mathematics performance confirmedthe findings reported in the sociological literature, whichdocumented that the number of books in the home is a proxyfor family educational SES. In the analysis, increasing theavailability of books by 1 standard deviation increased students’academic achievement by .240 standard deviations. Regardless ofthe type of school, students in homes with more than 10 booksoutperformed students in homes with fewer than 10 books orstudents in homes with no books. Families with parents withhigher levels of education also tended to own more books.

The presence of certain items in the home – e.g. a radio, bicycle,water tap, cassette player, television, telephone, gas stove,computer and motorbike – reflected family SES and wassignificantly related to students’ academic achievement. Morethan 60 percent of the students lived in homes with a radio, bicycle,water tap, television and telephone, but only 25 percent of thestudents lived in homes with computers and motorbikes. Studentsin urban schools and private schools were more likely to have acomputer in the home than students in rural schools or publicschools. For the various possessions identified in this study, familyownership of a radio, bicycle and computer was significantlyrelated to students’ mathematics performance in Nepal. Thepresence of a computer in the home was found to be significantlyrelated (.214, <.001) to students’ mathematics performancecompared to ownership of other items. This finding is consistentwith the finding that the facilities available in the home werepositively related to student achievement in Jordan (Al-Nhar,1999).

6.2. Student characteristics

Students’ knowledge, interests and capabilities are related toeducational quality. The second model, which incorporatedstudent characteristics, accounted for 18 percent of the variancein students’ mathematics performance. The analysis found thatgender was significantly related to mathematics achievement (.83,p < .050). Overall, boys outperformed girls, with a mean differenceof �3.05 (see Table 2). However, for students attending privateschools, girls (mean score = 24.37) outperformed boys (meanscore = 21.27). Similarly, middle and low caste girls outperformedboys. In Vietnam, Hungi (2008) also found that boys outperformedgirls in mathematics (�3.91, .59), while girls outperformed boys inreading (8.51, .63). Fuchs and Woessmann (2004) obtained similarresults using PISA data. However, Farkas et al. (1990) found thatgirls outperformed boys in 12 subjects with a median difference of�4.5 points in the United States. Sammons’ (1995) longitudinalanalysis of student achievement over the course of 9 years foundthat girls outperformed boys in reading and mathematics in theUnited Kingdom. However, Veenstra and Kuyper (2004) found thatstudent characteristics accounted for more variance than familycharacteristics in the Netherlands and that boys outperformed girlsin mathematics but not in language skills.

Reducing inequalities in school attendance and academicachievement in students from different ethnic groups is a criticaleducational issue in Nepal. The present study found thatdifferences in outcomes were related to children’s ethnicity. Thehigh caste group outperformed the middle and low caste groups.However, the effect of ethnicity differed in public and privateschools. Ethnicity was related to mathematics performance inpublic schools but not in private schools. Parents of children whoattend private schools might have higher SES, which could mitigatethe effects of ethnicity. This finding is consistent with Jamison and

Lockheed’s (1987) finding that land ownership, literacy level, castemembership and district of residence were strongly related tochildren’s school attendance in Nepal.

Although school attendance at all levels has improved, graderepetition and school absenteeism are still a critical problem inNepal. Despite targeted interventions to minimise grade repetitionand school absenteeism, 9 percent of school children repeat eighthgrade (DOE, 2010). However, grade repetition depends on the typeof school and its location. Grade repetition was negatively relatedto mathematics performance (see Table 3, model II). In theanalysis, children who repeated a grade one or more timesdecreased scores by �.050 standard deviations compared tochildren who never repeated a grade. In the study, 31 percent ofstudents reported that they had repeated a grade at least once; 37percent of public school students had repeated a grade comparedto 6 percent of private school students. Similarly, there was lessgrade repetition in the high caste group (20 percent) compared tothe middle caste (30 percent) and low caste (35 percent) groups.Finally, students in urban areas were less likely to have repeated agrade than students in rural areas. This finding is consistent withthe findings of Hungi (2008) in Vietnam and Marshall (2003) inHonduras, who found that repeating a grade more than once wasassociated with poorer academic outcomes.

Because students who regularly attend school receive morehours of instruction than students who are absent, absenteeism isnegatively related to learning outcomes. Analysis showed thatdecreasing school absenteeism by one standard deviation in-creased students’ academic achievement by �.185 standarddeviations (�6.819, p � .001). The study found that 32 percentof the private school students reported absenteeism compared to53 percent of public school students. In the mountain district, 74percent of the students reported absenteeism. One reason that wasoften mentioned by school head teachers and teachers for schoolabsenteeism, particularly in Nepal, is children’s involvement inhousehold chores.

Homework, which is an important point of contact betweenparents, children and schools, was significantly related to students’mathematics performance. In the analysis, increasing student timeon homework by 1 standard deviation increased academicachievement of children by .126 standard deviations (see modelII in Table 3). Ninety percent of private school children reporteddoing homework every day compared to 70 percent of publicschool children; none of the private school children reported thatthey never did homework. However, this study finding contrastswith the findings of Engin-Demir (2009) in Turkey and Keith et al.(1986) in the United States, although it supports Cooper (1989),who found that students who did homework scored two thirds of astandard deviation higher than those who did not.

Expanding quality early childhood education and development(ECED) programme for all children is a policy priority that isdesigned to enhance children’s overall development and establisha foundation for school success in Nepal. More than 50 percent ofthe students in the study reported preschool experience throughthe ECED programme; 70 percent of private school students hadpreschool experience compared to 40 percent of public schoolstudents. There was significant, positive relationship betweenpreschool experience and academic achievement (.086, p < .001),which was consistent with the Save the Children (2003) findingthat students with preschool experience were 26 percent morelikely to be promoted from grade 1 to grade 2 than childrenwithout preschool experience in Nepal (UNESCO, 2003, p. 182).

Other student characteristics that were examined werestudents’ perception of their mathematics teachers, academicmotivation and involvement in household chores. In the analysis,students who perceived that it was easy to understand theirteacher’s instruction had mathematics scores that were .153

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–4138

standard deviations higher (p < .001). Student motivation alsoplays an important role in learning (Schunk, 1990) because itencourages students to engage in learning activities. The analysisfound a significant relationship between student motivation andoutcomes (.135, p < .001). However, the number of hours spentstudying was not significantly related to students’ academicachievement, which contrasts with the findings of Engin-Demir(2009) in Turkey.

Children often carry out household chores in Nepal, andincreased involvement in household chores was associated withfewer opportunities to study. In the analysis, increasing children’sinvolvement in household chores by one standard deviationdecreased academic achievement by �.6 standard deviations(p < .050), which is consistent to Akabayashi and Psacharopoulos’(1999) findings in Tanzania. Children who spent only one hour ondaily household chores scored 3.53 points higher on average thanchildren who spent two or more hours per day. Private schoolstudents reported less time spent on household chores(mean = 1.77 hours per day) compared to public school students(mean = 2.45 hours per day). Involvement in household choressignificantly affected academic outcomes for public schoolstudents but not private school students.

6.3. Parental support

Parents are considered to be children’s first teachers, andchildren’s academic performance is better when parents areinvolved in children’s schooling. However, measures of parentalinvolvement in children’s education vary across studies, anddifferences are associated with SES and ethnic background(Boethel, 2003). The third model, which incorporated parentalinvolvement as well as family and student characteristics,accounted for 50 percent of the variance in students’ mathematicsachievement. However, parental support only accounted for 2.1percent of the variance in academic achievement (see Table 3). Thestudy found a significant relationship between parental involve-ment in homework and children’s school performance (1.31,p < .001). Many studies that have investigated the effects ofparental involvement in children’s homework have found con-flicting results; although some studies have found an association(Cooper, 1989; Fehrmann et al., 1987), others have not (Epstein,1987; Engin-Demir, 2009). Hoover-Dempsey et al. (1995) sug-gested that parental involvement entails a multi-dimensional setof tasks and that parents often felt ill-prepared due to limitedknowledge and demands on their time and energy. The finding ofthis study indicated that parental support significantly related tochildren’s mathematics performance (.72, p < .50). Children whostrongly agreed that their parents encouraged them to studyobtained scores that were 6.37 points higher on average thanchildren who strongly disagreed. In addition to encouragingchildren to study, parents might also monitor their children’sefforts by asking questions and observing their activities. Theanalysis found that parental monitoring was positively related tochildren’s academic outcomes (.83, p < .50).

6.4. School characteristics

In addition to the effects of family characteristics, studentcharacteristics and parental support, children’s learning outcomesare also associated with school characteristics because instruc-tional activities occur in school. The fourth regression model,which combined these four independent variables, explained 58.6percent of the variance in students’ mathematics achievement.However, school characteristics accounted for only 8.6 percent ofthe variance in mathematics performance (see Table 3). It isimportant to note that some of the individual variables included

within the set of student characteristics such as absenteeism, graderepetition and homework completion are closely related to schoolcharacteristics because they reflect the quality of the school thatthe student attends which might be a possible reason for lowvariance of school characteristics.

There were substantial differences in academic performanceassociated with school characteristics in Nepal. Traditionally,governments have preferred to retain control over the educationalprocess to maintain the quality of public education. However,budgetary shortfalls and increasing student enrolment have ledmany governments to re-examine the role of the private sector ineducation (Kingdon, 1996). After democracy was established in1990, the Nepalese government introduced liberal policies thatemphasized the role of the private sector, and currently 11 percentof students are enrolled in the more than five thousand privateschools that have mainly been established in the urban area.However, this has also increased public dissatisfaction due toissues of cost and access to quality education. The results of theanalysis presented in Table 2 revealed that private school students’scores (mean = 22.31 points) were higher than public schoolstudents’ scores (mean = 9.47 points), and this difference wassignificant (�.120, p < .001). Because public and private schools inNepal differ in terms of the quality of education they provide,policy makers in Nepal should adopt measures that reduce thelearning gap between the two types of schools. Study findingsindicated that the higher performance of private school studentsmight be due to parents’ higher SES, as well as private schools’lower student absenteeism, lower student–teacher ratios, locationof private schools in the urban setting and greater number ofschool days. Another possible reason might be competitive natureof schools as they face pressure to maximize students’ perfor-mance while using resources efficiently (Woessmann, 2003). Thisfinding is consistent with Psacharopoulos (1987) and Govinda andVarghese (1993), who found that children in private schools incountries as diverse as the UK, Tanzania, Australia, India, andColombia performed better than their public school counterparts,even after controlling for student and family characteristics.

Scarce resources in developing countries often lead to pooreracademic performance due to dilapidated schools, insufficientnumbers of teachers, and crowded classes. However, this situationcan be mitigated by effective instructional processes andimprovement in equipment and instructional materials ratherthan smaller class size (Fuchs and Woessmann, 2004). This studyfound a negative relationship between student–teacher ratio andstudents’ academic achievement (�.102, <.05), indicating thatchildren in smaller classes outperformed children in larger classes.In Nepal, schools lack instructional materials, and larger class sizesmight affect student learning due to less interaction betweenstudents and teachers, a noisier and more disruptive classenvironment, and teachers’ inability to carry out activities orfocus on students’ individual needs. This result contradictsHanushek’s (1997) findings but is consistent with Woessmann’s(2003) finding that smaller class sizes were significantly related tothe poor performance of students in math and science on the TIMSSin 39 participating countries in 1994/1995. Kingdon (1999) foundthat smaller class size was negatively associated with theacquisition of numeracy skills but positively associated withreading skills in India. Lindhal (2005) found that larger class sizewas detrimental for minority and economically disadvantagedchildren in Sweden.

Although research findings have not been consistent, teachertraining is a widely used intervention to improve instructionalpractices. The analysis found a significant, positive relationshipbetween teacher training and students’ academic achievement(.77, p < .05). Children taught by trained teachers were likely toscore 10 points higher than children taught by partially trained or

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–41 39

untrained teachers. This result is consistent with Biniakunu’s(1982) findings in the Democratic Republic of Congo. However,Craig et al. (1998) found that teacher training did not necessarilyimprove educational quality. An academic year that includes agreater number of school days provides more opportunities forlearning to occur in the classroom. In Nepal, the Education Act andgovernment regulation requires a minimum of 220 school days inan academic year. However, the study found that the number ofschool days varied from 175 to 193 days and that the number ofschool days during the academic year was significantly related tostudents’ academic achievement (.119, p < .001).

7. Conclusions and implications

The purpose of this study was to identify factors related toeducational quality by determining the effects of family char-acteristics, students’ characteristics, parental involvement andschool characteristics on students’ mathematics performance.Family characteristics were responsible for a large proportion ofthe variance in students’ academic achievement. This finding hasprofound implications for the education of the ethnically andgeographically diverse population in Nepal where more than 40percent of parents are still illiterate. The study found that childrenwith educated parents and in homes with greater materialresources exhibited better academic performance, which suggeststhat educational policies should focus on families as well asschools. In short-term, few strategies can only be adopted tominimize the effects of family characteristics but in long-termmuch can be done. To provide more opportunities for studentsfrom disadvantaged communities, educational programmes simi-lar to those implemented in the United States in the 1960s and theconditional cash transfer programmes that were successfullyimplemented in Mexico and Brazil could be introduced. Studyfindings revealed a significant relationship between parentalinvolvement and school outcomes, which indicates that educa-tional policies should promote the partnership of schools,municipalities, non-governmental organisations and other socialagencies to increase parents’ awareness of their role in fosteringchildren’s academic achievement, particularly in disadvantagedcommunities.

The study finding that gender and ethnic background weredifferentially associated with mathematics performance in Nepalraises another issue. In the current political climate, the issue ofequal opportunity in education must be addressed to maintainsocial equity and harmony. The Ministry of Education shouldinvestigate this issue thoroughly and determine the extent towhich teacher training promotes gender sensitivity and schoolsprovide inclusive and gender-neutral environment and gender-balanced curriculum and textbooks. Similarly, the study foundsignificant differences in academic achievement between publicand private schools, which is a serious concern for policy makers,teachers and parents. It means than Nepalese government shouldadopt policies to manage both types of schools (private and public)regarding the quality of education they are offering.

Results of this study indicate strong negative relationship ofstudent absenteeism to mathematics performance. In the contextof high absenteeism in both private and public schools and itsdiverse nature, further study is needed to thoroughly understandthis phenomenon. However, districts and schools with high levelsof absenteeism should be identified, and programmes that fosterschool attendance such as day meals, scholarships and conditionalcash transfer could be introduced. Similarly, schools shouldaddress the problem by working with parents and communitymembers through the School Management Committee (SMC),Parent Teacher Association (PTA), local agencies such as the VillageDevelopment Committee (VDC) and municipal agencies. Grade

repetition, which was negatively related to academic outcomes,was related to other issues, and schools that improve educationalquality will automatically reduce grade repetition. Studentperception of teachers was significantly related to learningoutcomes, which suggests that teacher training should focus notonly on instructional knowledge and skills but also on attitudesthat maintain a positive atmosphere both inside and outside theclassroom.

Providing training to teachers is an important policy for betterschool outcome but only teacher training may not work properlybecause it is strongly related with other teacher developmentpolicies such as teacher licensing, hiring and professionaldevelopment which may make important difference in thequalifications and capacities that teachers bring to their work.The study found a positive relationship between teacher trainingand academic outcomes. However, due to students’ poor anduneven mathematics performance, further investigation is neededto determine the extent to which teachers master training contentand incorporate it into classroom practices.

Nationally, the average student–teacher ratio is 44:1 at thebasic education level. However, the average student–teacher ratiowas 60:1 in the schools participating in the study, which is muchhigher than the national average and indicates that humanresources in education are not equally distributed. Policies suchas teacher redeployment within and between districts should beintroduced to distribute existing resources more equitably. Studyfindings also revealed that schools’ academic year did not provideminimum number of school days required by educationallegislation and regulations, which is a serious concern for policymakers, educational administrators and other stakeholders. In thisregard, roles of School Supervisors (SS), Resource Persons (RP),School Management Committees (SMC) and Parent TeacherAssociations need to be re-examined, and further investigationis needed to identify effective monitoring mechanisms. Schools’physical facilities might gradually be improved by regularlysurveying schools’ physical facilities to establish that they meetminimum criteria.

Finally, it came to notice from a comprehensive regressionmodel (model IV) employed in this study that could only explainless than 60 percent of variance in academic achievement amongstudents. It means that 41 percent of the variance in students’mathematics achievement in Nepal was unexplained by theselected independent variables of this study, which indicates thatfurther theoretical and empirical research is needed to identify thecontribution of other factors related to academic outcomes, such asclassroom processes, teacher characteristics and motivation. Inaddition, more sophisticated methods of analysis should beadopted to explore the complex and reciprocal relationship amongvarious factors.

Acknowledgements

I am grateful to Professor Shoko Yamada of the Graduate Schoolof International Development, Nagoya University, Japan, forinvaluable advice and guidance that contributed to the successof this study. I acknowledge the support of District EducationOffices and the schools that provided permission for the study.Finally, I thank the head teachers, teachers and students whoparticipated in this study.

References

Akabayashi, H., Psacharopoulos, G., 1999. The trade-off between child labour andhuman capital formation: a Tanzanian case study. The Journal of DevelopmentStudies 35 (5) 121–140.

Al-Nhar, T., 1999. Determinants of grade eight achievement in Jordan: a multilevelanalysis approach. Educational Psychology 19 (1) 37–45.

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–4140

Angrist, J.D., Lavy, V., 1999. Using Maimonides rule to estimate the effect ofclass size on scholastic achievement. The Quarterly Journal of Economics114, 533–575.

Astone, N.M., McLahanan, S.S., 1991. Family structure, parental practices and highschool completion. American Sociological Review 56 (3) 309–320.

Barnard, M.W., 2004. Parent involvement in elementary school and educationalattainment. Children and Youth Services Review 26, 39–62.

Behrman, R.J., Taubman, P., 1886. Birth order, schooling, and earnings. Journal ofLabor Economics 4 (3) 121–145.

Becker, G., 1964. Human Capital. A Theoretical and Empirical Analysis, with SpecialReference to Education. NBER.

Biniakunu, D.D., 1982. In-service teacher training improves eight grades’ readingability in Zaire. Journal of Reading 25, 662–665.

Blake, J., 1989. Family Size and Achievement. University of California Press,Berkeley.

Boethel, M., 2003. Diversity: School, Family and Community Connections. AnnualSynthesis, Institute of Education Sciences, Washington, DC.

Buchmann, C., Hannum, E., 2001. Education and stratification in developingcountries: a review of theories and research. Annual Review of Sociology27, 77–103.

Buchmann, C., 2002. Measuring Family Background in International Studies ofEducation: Conceptual Issues and Methodological Challenges. InternationalComparative Studies in Education, National Academy Press, Washington DC.

Cameron, V.S., Heckman, J.J., 2001. The dynamics of educational attainment forBlack, Hispanic and White males. Journal of Political Economy 109 (3) .

Centre Bureau of Statistics, 2011. National Census. Thapathali, Kathmandu, Nepal. .Center for Educational Research and Social Development (CERSOD), 2005. Parental

support and student learning. A research report. Submitted to Department ofEducation, Kathmandu, Nepal.

Chapman, D., Adams, D., 2002. The Quality of Education: Dimensions and Strategies.The Asian Development Bank/Hong Kong. Comparative Education ResearchCentre, The University of Hong Kong.

Chapman, W.D., Weidman, J., Cohen, M., Mercer, M., 2005. The search for quality: afive study of national strategies to improve educational quality in Central Asia.International Journal of Educational Development 25, 514–530.

Cohen, J., Cohen, P., 1983. Applied Multiple Regression/Correlation Analysis inBehavioural Sciences, 2nd ed. Lawrence Erlbaum, Hillsdale, NJ.

Coleman, J.S., Campbell, E.Q., Hobson, C.J., McPartland, J., Mood, A.M., Weinfield,F.D., 1966. Equality of Educational Opportunity. National Centre for EducationalStatistics, Washington, DC.

Coleman, J.S., 1988. Social capital in the creation of human capital. American Journalof Sociology 94, 95–120.

Coleman, J.S., 1991. Parental involvement in education. In: Policy Perspective, Officeof Educational Research and Improvement. US Department of Education,Washington, DC.

Cooper, H., 1989. Synthesis of Research on Homework. Educational Leadership.Longman, New York.

Craig, H.J., Kraft, J.R.J., Plessis, J.D., 1998. Teacher Development Making an Impact(19009: USAID). World Bank.

Dalin, P., 1994. How Schools Improve? An International Report. Cassel, London andNew York.

Downey, D.B., 1995. When Bigger is not better: family Size. Parental resourcesand Children’s Educational Performance. American Sociological Review 60,746–761.

Downey, D.B., 2001. Number of siblings and intellectual development: the resourcedilution explanation. American Psychologist 56 (6) 497–504.

Department of Education, 2010. Flash Report. Sanothimi, Bhaktapur, Nepal.Department of Education, 2011. Flash Report. Sanothimi, Bhaktapur, Nepal.Educational and Developmental Service Centre, 2008. National Assessment

of Grade 8 Students. Submitted to Department of Education, Kathmandu,Nepal.

Engin-Demir, C., 2009. Factors influencing the academic achievement of the Turkishurban poor. International Journal of Educational Development 29, 17–29.

Epstein, J.L., 1987. Parent involvement: what research says to administrators.Education and Urban Sociology 19, 277–294.

Farkas, G., Sheehan, D., Grobe, R.P., 1990. Coursework mastery and school success:gender, ethnicity, and poverty groups within an urban school district. AmericanEducational research Journal 27 (4) 807–827.

Fehrmann, G.P., Keith, Z.T., Reimers, M.T., 1987. Home influence on school learning:direct and indirect effect of parental involvement on high school grades. Journalof Educational Research 806, 330–337.

Field, A., 2009. Discovering Statistics Using SPSS, third ed. SAGE Publication.Fuchs, T., Woessmann, L., 2004. What accounts for international differences for

student performance? A re-examination using PISA data, CESifo Working PaperNo. 135.

Fuller, B., 1987. What school factors raise achievement in the third world? Reviewof Educational Research 57 (3) 255–292.

Govinda, R., Varghese, V.N., 1993. Quality of Primary Schooling in India: A CaseStudy of Madhya Pradesh. International Institute for Educational Planning, Parisand National Institute of Educational Planning and Administration (NIEPA),New Delhi.

Gray, J., Partington, G., 2003. Attendance and non-attendance at school. Reform andresistance in aboriginal education 133–163. University of Western AustraliaPress.

Gray, J., Beresford, Q., 2008. Australia’s quest for equity in Indigenous education.Australian Journal of Education 52 (2) 197–223.

Gregg, P., Machin, S., 1999. The Relationship between Childhood Experiences,Subsequent Educational Attainment and Adult Labour Market Performance.Centre for Economic Performance, London School of Economics.

Guncer, B., Kose, R., 1993. Effects of family and school on Turkish students’ academicperformance. Education and Society 11 (1) 51–63.

Hanushek, E., 1986. The economics of schooling: production and efficiency in publicschools. Journal of Economic Literature 24 (3) 1141–1177.

Hanushek, E.A., 1997. Assessing the effects of school resources on studentperformance: an update. Educational Evaluation and Policy Analysis 19 (2)141–164.

Heady, C., 2003. The effect of child labour on learning achievement. World Devel-opment 31 (2) 385–398.

Heer, M.D., 1985. Effect of sibling number on child outcome. Annual Review ofSociology 11, 27–47.

Heyneman, S.P., Loxley, W.A., 1983. The effect of primary-school quality on aca-demic-achievement across 29 high-income and low-income countries. Ameri-can Journal of Sociology 88 (6) 1162–1194.

Heyneman, S.P., 2002. Defining the influence of education on social cohesion.International Journal of Educational Policy, Research and Practice 3, 73–97.

Hungi, N., 2008. Examining differences in mathematics and reading achievementamong grade 5 pupils in Vietnam. Studies in Educational Evaluation 34, 155–164.

Hungi, N., Thuku, F.W., 2010. Differences in pupil achievement in Kenya: implica-tions for policies and practice. International Journal of Educational Develop-ment 30, 33–43.

Hoxby, C.M., 2003. School choice and school competition: evidence from UnitedStates. Swedish Economic Policy Review 10, 9–65.

Hoover-Dempsey, V.K., Bassler, C.O., Burow, R., 1995. Parents’ reported involvementin students’ homework: strategies and practices. The Elementary School Journal95 (5) 435–450.

Iverson, B.K., Walberg, H.J., 1982. Home environment and learning: a quantitativesynthesis. Journal of Experimental Education 50, 144–151.

Jamison, D.T., Lockheed, M.E., 1987. Participation in schooling: determinants andlearning outcomes in Nepal. Economic Development and Cultural Change 35,2–279.

Johnston, S.J., Thompson, H.G., 2006. Variation in the relationship between non-school factors and Student Achievement on International Assessments. Insti-tute of Education Sciences, Department of Education, United States.

Keith, T.Z., Reimers, T.M., Fehrmann, P.G., Pottebaum, S.M., Aubey, L.W., 1986.Parental involvement, homework, and TV time: direct and indirect effects onhigh school achievement. Journal of Educational Psychology 5, 373–380.

Kingdon, G.G., 1996. The quality and efficiency of private and public education: acase-study of urban India. Oxford Bulletin of Economics and Statistics 58 (1)57–82.

Kingdon, G.G., 1999. How much do school matters to pupil achievement in India? .Konu, A., Rimpela, M., 2002. Well-being in schools: a conceptual model. Health

Promotion International 17 (1) 79–87.Krurger, B.A., 2003. Economic consideration and class size. The Economic Journal

113–485.Kulpoo, D., 1998. The quality of education: some policy suggestions based on a

survey of schools. SACMEQ Policy Research No. 1: Mauritius. IIEP, Paris.Kutnick, P., 2000. Girls, boys and school achievement: critical comments on who

achieves in schools and under what economic and social conditions achieve-ment takes place—a Caribbean perspective. International Journal of EducationalDevelopment 20, 65–84.

Lareau, A., 1987. Social class differences in family–school relationship: the impor-tance of cultural capital. The Sociology of Education 60 (2) 73–85.

Lindhal, M., 2005. Home versus school learning: a new approach to estimating theeffect of class size on achievement. The Scandinavian Journal of Economics 107(2) 375–394.

Macbeth, A., 1994. Involving parents. In: Pollard, A., Bourne, J. (Eds.), Teaching andLearning in the Primary School. Routledge, London and New York in Associationwith Open University, pp. 304–313.

Marshall, H.J., 2003. Grade repetition in Hunduran primary schools. InternationalJournal of Educational Development 23, 591–605.

McEwan, J.P., Marshall, J.H., 2004. Why does academic achievement vary acrosscountries?Evidence from Cuba and Mexico. Education Economics 12(3) 205–217.

McNeal, R.B., 1999. Parental involvement as social capital: differential effectivenessin science achievement, truancy and drop out. Social Forces 78 (1) 117–144.

Ma, X., 1997. A multiple regression analysis of mathematics achievement in theDominican Republic. International Journal of Educational Development 17 (3)313–321.

Muijs, D., 2004. Doing quantitative research in Education with SPSS. SAGE publica-tion, New Delhi.

National Living Standard Survey, 2011. Central Bureau of Statistics, StatisticalReport, Kathmandu, Nepal..

Olneck, M.R., Bills, D.B., 1979. Family configuration and achievement: effect of birthorder and family size in a sample of brothers. Social Psychology 42, 135–148.

Organization for Economic Co-operation and Development, 2003. Learning forTomorrow’s World, First Result from PISA..

Patrinos, H.A., Psacharopoulos, G., 1995. Educational performance and child labourin Paraguay. International Journal of Educational Development 15 (1) 47–60.

Parcel, L.T., Menaghan, G.E., 1994. Early parental work, family social capital, andearly childhood outcomes. American Journal of Sociology 99 (4) 972–1009.

Psacharopoulos, G., 1987. Public vs. private schools in developing countries: evi-dence from Colombia and Tanzania. International Journal of Educational De-velopment 7 (1) 59–67.

K.P. Pangeni / International Journal of Educational Development 34 (2014) 30–41 41

Ray, R., Lancaster, G., 2003. Does Child Labour Affect School Attendance and SchoolPerformance? Multi Country Evidence on SIMPOC Data. Unpublished Report.ILO/IPEC.

Research Center for Educational Innovation and Development (CERID), 1999.Assessment of Learning Achievement of Lower Secondary Children: Grades 6and 8. CERID in Association with MOES/Nepal.

Rothstein, R., 2000. Finance Fungibility: Investing Relative Impacts of Investmentsin Schools and Non-school Educational Institutions to Improve StudentAchievement. Centre on Educational Policy Publications, Washington, DC.

Sammons, P., 1995. Gender, ethnic and socio-economic differences in attainmentand progress: a longitudinal analysis of student achievement over 9 years.British Educational Research Journal 21, 465–485.

Save the Children, 2003. What is the difference? An ECD impact from Nepal,Kathmandu.

Schiller, K.S., Khmelkov, V.T., Wang, X.Q., 2002. Economic development and theeffects of family characteristics on mathematics achievement. Journal of Mar-riage and Family 64, 730–742.

Scheerens, J., 2004. The quality of education at beginning of the 21st century.In: Apaper commissioned for the Global Monitoring Report 2005, Paris, UNESCO. .

Schunk, D.H., 1990. Goal setting and self-efficacy during self-regulated learning.Educational Psychology 25, 71–86.

Slavin, E.R., 1997. Design competitions: a proposal for a new federal role ineducational research and development. Educational Researchers 26 (1)22–28.

Stash, S., Hannum, E., 2001. Who goes to school? Educational stratification bygender, caste, and ethnicity in Nepal. Comparative Education Review 45 (3)354–378.

Stevenson, D.L., Baker, P.D., 1987. The family–school relation and the child’s schoolperformance. Child Development 58, 1348–1357.

Sui-Chu, E.H., Willms, J.D., 1996. Effects of parental involvement on eight gradeachievement. Sociology of Education 69 (2) 126–141.

Teachman, J.D., 1987. Family Background. Educational Resources, and EducationalAttainment. American Sociological Review 52 (4) 548–557.

Teachman, J.D., Paasch, K., Carver, K., 1996. Social capital and dropping out of schoolearly. Journal of Marriage and the Family 58, 773–783.

UNESCO, 2003. Education for All, Global Monitoring Report, Why are Girls HeldBack? UNESCO, Paris.

UNESCO, 2005. Education for All, Global Monitoring Report, The Quality Imperative.UNESCO, Paris.

Veenstra, R., Kuyper, H., 2004. Effective students and families: the importance ofindividual characteristics for achievement in high school. Educational Researchand Evaluation 10 (1) 41–70.

Verspoor, M.A., 1989. Pathways to Change: Improving the quality of education indeveloping countries, World Bank Discussion Paper 53. The World Bank Pub-lications, Washington, DC.

Willms, D.J., Somers, M.A., 2001. Family, classroom, and school effects on children’seducational outcomes in Latin America. School Effectiveness and School Im-provement 12 (4) 409–445.

Woessmann, L., 2003. Schooling resources, educational institutions and studentperformance: the international evidence. Oxford Bulletin of Economics andStatistics 65 (2) 0305–9049.

Yayan, B., Berberoglu, G., 2004. A re-analysis of the TIMSS 1999 mathematicsassessment data of the Turkish students. Studies in Educational Evaluation30, 87–104.