Socio-Economic Status (SES) Origins: The case of Tanzania Secondary School Students G.A. Malekela-LeclUrer, lJ.tafiti Vol viii No. I, 1986, Journal Department of Educatiun of the Faculty of Arts and Social Sciences University oj Dar es Salaam University of Dar es Salaa~ Introduction The aim of this study is to investigate whether or not children from differ- ent SES origins have similar chances of being selected for secondary eductttion in public schools both at O.-lev:l, where selection is based on regional merit, and at A-level where selectlon IS governed by merit at the national levels. It is hypothesised that, although the majority of students would come from relatively humble backgrounds, children in this category would in fact be un- der represented. Findings from elsewhere, in the West in Africa, as well as in Tanzania, would lead to this expectation. Findings from samples of 1,186 Form IV students (14070 of the popuia- tion), representing the a-level and 1,727 Form V students (70% ofthe popula- tion), representing the A-level, who were attending the two levels in 1982 reveal that at both 0- and A-levels, selectivity indices increase with increasing parental education and occupational status. The more schooling the parents have, the more likely it is that their children wi!l be selected. Similarly with occupations, students whose parents are employed III the modern sector have a greater chance of being selected than' those whose parents are still in the traditional sector . This shows that there is inequality in access to secondary educiltion in Tanza- nia as children from different SES do not have equal chances of being selected. Though the majority of the students at both levels come trom relatively hum- ble backgrounds,. they are actually underrepresented, whereas children from higher SES are overwhelmingly represented .. The data in this study is based on a survey of 1,186 Form IV students 14% of the population), and 1,727 Form V students (70%) who were attending pl,lb- lie schools in the 1982 school year. Before this data, collected between July and October 1982, a pilot study was carried out in May' 1982 to find if there were ambiguities in any of the questions, to clarify ideas relevant to the reasearch problem, and to determine the best procedures to be adopted when carrying out the actual study.This was consistent with suggestions made by notablescho- lars, such as Gay (1981), Leedy (1980), and Bradburn (1982), that it is always useful to pre-test questionnaires whenever a descriptive survey method like this one is used. The Form V sample was adequately representative. It was drawn from 25 of the 31 A-level schools covering all regions except Dodoma, Mara, Shinyan- ga and Singida, where there are such schools. All Form V students in the 25 schools present on the day the questionnaire was administered filled out the questionaire. It was stratified sexually and by specialization (Science and Arts). Thesamplewas picked specifically to measure the extent to which regional quo- tas at Form I selection four years earlier had been maintained when the region- al quota system was abandoned and when selection for Form V was made. There were however, some minor problems with the Form IV sample. Although the sample was stratified by sex and residence (boarding/clay), hoard- ers were slightly underrepresented by about 110,10, while day scholars were over- 67 repre.sented, again oy 11f1Jo. Girls in day'schools were oversampled. These undesirable proportions occured for reasons beyond the aut,hor's control. On going to some schools in order to have students fill out the questionnaire, it was discovered that all schools in the region were holding their Form IV mock examinations, or that teachers were too busy to release their students for an hour to complete the questionnaire as the teachers were trying to comp~te th.eir syllabi in time for the Form IV national exams. The main interest in having this sample was to measure general background characteristics so that they could be compared with those of the Form V students. Both boarding and day schools were identified because it is known that boarding schools cater mainly for stu- dents who completd their primary education in rural areas, while the few ur- ban day schools serve only the urban dwellers. We assumed the rural and urban populations to be different in many aspects. Data for the Form IV sample was collected from 31 public secondary schools scattered all over the country, but students in private schools were deliberately left out as they do not go to thIJse schools through ther'igorous quota selection. With the Form IV data it is pos- sible to c6mpare social selectivity of the two levels of secondary education wi- thin the publ!ic system where the governmerit has instituted regional equity at Form I but ~as no social equity criteria when selection is made. Why SOfia! seletivity? Scholars are concerned with whether meritocratic selectign is filir or biased. They are also concerned with whether, by equalizing educational ~pp<>rtunity (Le. access), children from different social origins will have equal rFsuIts(Le. achievement) when they have similar initial inputs (Le. ability <;lnd~otitation). Findings from both the Western capitalist and Eastern socialist c0t.lntr~s have revealed a strong association between social origins, SCholastic achievement, and/or status attainment (Bourdieu & Passeron, 1977; CoIlins, 19711;Dobson & Swafford, 1980; Entwistle, 1978; Epstein, 1972; Fisz- man, 1982; Kerckhoff, 1972: Kozakiewicz. 1980; Sewell & Shah, 1967). A higl).er percentage. of children of high SES finish secondary and college .education having higher intelligence than children of lower SES. Higher SES parents can ensure tfiat their offspring get superior education by providing the material con- ditions and an environment conducive to the development of the skills, values, attitudes, and motivations that are associated with achievement in life. The issue at stake is fo finJ the distribution of income and services among social groups. Since the taxpayers 'money is used in one way or another in main- taining educational systems, there is concern over distributive justice among the different categories of taxpayers. What is the situation in developing countries where western educational systems are recent imports? To date, there is a continuing debate among scho- lars as to . whether or not children from different social. origins have equal chances of being selected. Heyneman (1979), for example, concludes that children from poor and humble backgrounds in Uganoa have equal chances of being selecte~I to secondary school due to the fact that the elites have not yet developed a distinct culture and that children from all backgrounds are equal- ly motivated to take the primary seven examinations. Other scholars, such as Cooksey ';.\1981) and Niles (1981) have disputed such findings ~s misleading as in their :itudies it was evident that children from higher SES had ;nore chances of being selected that those from lower ones. Weis (1979) found a similar situa- tion in Ghana where children from higher SES seemed to have cemented their advantage over others from lower SES when she compared her data to. that of Foster (1965). 68 Tanzania is aspiring to build an egalitarian SOClety.Consequently, regional and district quotas have been instituted for equity consideration when selection for Form I is made. Thus far, there has been no effort to extencl the same SES. I There are complaints by some Tanzanians that most children from high SES families and their relatives gain selection to Form I (Malekela 1977' Mbilinyi 1976). In spite of this, merit in the final exam is not suppos;d to be the only criterion to be considered when selection is made. Continuous assessment character, and nation-building variables are to be considered as well. To what <;:xtenthas the inclusion of other variables besides academic excellence contribut- ed towards gr&ter equality, if any? Although merit has been found to be cIass- biased in developed countries, will this be true for Tanzania which has a very tiny elite and whose growth into a privileged group has been hijacked by the Arusha Declaration since 1967? Before we go into the proper socio-economic status characteristics of the students, it is a good idea to have a general background to Tanzanian society. With such knowledge at hand comparisons can thus be made more meaningful. Tanzania, like any other society, is heterogeneous. It has diverse popula- tions with different cultures and subcultures. Economically, over 85% of the population consists of peasants who depend on subsistence agriculture. Indus- tries are very few and are concentrated in the larger towns such as Dar es Salaam, Mwanza, Tanga, Morogoro, Arusha and Moshi. There are very few people in the modern sector of the economy. By 1978 there were only 511,310 people in wage and salary employment, which is about 60/0of the total labour force (Unit- ed Republic of Tanzania, 1982). Most of the people in wage and salaried occu- pations are in the services sector .. Concerning the provision of education, the majority of the adult popula- . tion was iliterate at independence in 1961. Six years later in 1967 when a census was taken, it was found that 69% of the Tanzanian population was illiterate. ButJiteracy levels were much higher in urban areas, where.61% of the adult population was literate, as compared to 29% in rural areas (Dey & Mogil, 1973). The male-female disparity in literacy was considered too: The level of liter- acy was much higher for males than females: in 1967 it was 45OJofor males and only 19% for females. The disparity between the two sexes is a historical phenomenon, althou!;h it tends to be exacerbated in developing countries where women are still more oppressed than in developed countries. Following President Nyerere's declaration in his new year's address of 1969 that 1970 was to be the 'Adult Education Year' in Tanzania (Nyerere, 1973) and subsequent efforts in the 1970's to give the general masses the rubrics of reading, writing and arithmetic, the efforts paid some dividends. By 1980 it coulc be reported that 64.556,10percent of the adult population could read and write (Wizara ya Elimu ya Taifa, 1981), an increase of 33.55% over 1967. In her 1982/83 budget estimates, the then Minister of National Education reported that 79% of the adult population was literate (Siwale, 1982). One looks at this achievement with some scepticism given the fact that there are not enough follow- up materials such as books, newspapers, and magazines to reinforce the ac- quired skills. One is afraid that as many as half these adults might relapse into illiteracy a year or so after being certfied that they have acquired the basic skills of reading, writing, and numeracy. In 1980, as was the case in the 1967 census, the "ratios of theiIliterates and literates varied widely across the country. For example, while Dar ~ Salaam (un- ban centre and capital) was leading in 1967 with 6OOJoof its adult population 69 literate, Shinyanga region w..s bringing up the rear with only 16% and males were ahead af females in all ~8 pravinces. Variatian between regians was obvi- aus in 1980. Dar es Salaam was again leading with 92.63070 of its population literate, while Lindi regian was at the bottam with anly 44.25'10 ..Unfartunate- Iy, the 1980 report does nat break dawn the fIgures far males and females, while the report af 1982 just gives tbe natianal percentage af literate people withaut shawing the regianal variatian and the sex ratias. By 1978, 87.83% af the ecanomically active'populatian was in agricultUre, mast af whom were invalved in subsistence farming. Only 1.45070o.f the'agricul- turallabour force was emplayed in plantations such as the sisal and tea planta- tions where machinery is used. Mast af the peasants in Tanzania depend an tbe hand-hoe. The distribution af other occupatians and the breakdawn of their labour farce is shown in Table 1. Table 1. ECONOMICALLY ACl1VE POPULATION BY OCCUPATION IN MAINLAND TANZANIA.' 1971, PERCENT . MajOr Occupation group Male Female Total 1. Agriculture 79.81 95.34 87.83 2. Craftlrnachine'~tor 5.94 0.18 2.97 3. Sales/clerical 3.54 0.92 2.19 4 .. Professional/1'echnicians/teachers 3.49 1.11 2.26 5. Managers/administrators 0.45 0.02 0.23 6. Non-agricu1turallabourers 2.74 0.49 1..58 7. Other workers 4.02 1.93 . 2.94 Total 99.99 99.99 100.00 3715173 3971580 7686753 Source: The figures have been computed by the author from 1978 Population Census Volume VII, p. 239 for male, p.260 for female, and p. 218 for the total. Note: Those categorized as students, other unoccupied, and not elsewhere stated have been exclud- ed in the. computation. Incames amang the salaried vary in the ratio. 9.1 (Nyerere, 1977). In rural areas, incames also. vary depending an the nature af the sails, climate, whether ar nat ane en~a~es in cash-crap production, the amaunt of land one has and so forth. Looking into income differences, urban salaried warkers had incames four times more than a peasant (Whitaker, 1978). This does nat differ much from the ILO (1978) study which showed tbat 40.60'10 af farmers enjoy a high stan- dard of living, compared to 30 x 40% af the wage eaners. When ane reads about the exodus of )'Oung people from the rural to the urban areas in develaping coun- tries, the glaring gaps in income life style and availability af amenities; most af which are concentrated in urban areas, are a driving force. It was hypothesized that although the majarity af the studt:l1Lswould came from relatively humble backgraunds, children in this categary would be under:" represented. Findings from elsewhere, in the West and hi"Africa, and those af Sa:naff (1979) and UNESCO (1981) studies an Thnzania wauld lead one to. this expectatian. In aur questionnaires for both samples, we included questians to measure students' parents' educatianallevel, occupations; and other variables such as sectars af emplayment, items found in the parents' hames; students' brathers, sisterS and clase relatives having secondary ar university educatian, wage-emplayment, and reasanably good income, and the languages a student speaks with his parents an the ane hand and with brothers and sisters. an the other. 70 ABto the educational level of parents of students, a question was asked of students about their tather's and mother's levels of schooling. We wanted to measure the impact of the parent's education on the selection of the Offspring to continue with schooling. Using educational data for the 1978 census we have tried to estimate the age-cohort .of the parents of the students in our samples. Given the fact that the mean ages for both Form IV and V students in our sample lie between 19 and 21, their parents on average should be between 40-49 years. Marriage aae for men is 23 years while for women it is much lower as v.umen do not pay dowry. By using the age-cohort of their parents, we can estimate the representation of students in each of the schooling levels of their parents. If we use the national average, the results will be deflated; hence the cohort. Thble 2 looks at the level of schooling of the fathers of Form IV students. TABLE 1 SOCIAL SELECTIVITY OF FORM IV STUDENlS BY FATHERS' FDUCATION 1978 PopulatioB Students by Repraeatatioa Ages 40 - 49 Father'. EdUcadOB IadIceI N 'I. N 'I. / / None 335,717 54.14 285 24.48 0.45 Primary 1-4 180,464 29.10 331 28.44 0.98 Primary 5-8 88,756 14.31 346 29.72 2.08 Form 1--4 I 12,147 1.96 111 9.54 4.87 Above Form 4 2,981 0.48 91 7.82 16.29 All males age 40--49 620,065 99.99 1,164 100.00 Other & N.E.S. 3,219 22 Note: Data for the 1978 population hAve been computed by the author from 1978 Population Cen- sus Vol. VII, P. U4. ' Of the fathers who had no formal schooling in 'Dlble 2, at least 209 (73 Of.) had adult education, which means they are perhaps able to read and scribble their names. However, girls had more fathers who had had schoo1.ing. For ex- ample, while 64 (13%) ofthe fathers of girls had no formal education, 221 (33"-') .of the fathers of boys had none. On the other hand, while only 28 (4'0) of the fathers of boys had above Form IV education, 63 (13'1.)" of the fathers of girls had had that level of schooling. It would seem that girls from humble families are much less likely than their brothers to be sent to school, while the gap is not so great in higher SES families. From Thble 2 it is quite clear that the higher the level of schooling the father has, the higher the selectivity. Those with fathers who have .... education beyond Form IV are 36 times as likely to be selected to ~on~aTy school as are those whos~ fathers have had no formal eduCSltion. 71 Table 3 Social Selectivity of Form IV Students by Mothers' Education 1978 Popu- Students by lation Mother's Representa- Ages 40—49 Education tion Indices N % N None 564,521 87.84 563 48.33- 0.55 Primary 1—4 61,694 9.60 367 30.64 3.19 Primary 5—8 15,147 2.36 174 14.94 6.33 Form 1—4 984 0.15 42 3.60 24.00 Above Form 4 328 0.05 29 2.49 49.80 All Females age 40—49 642,674 100.00 1,165 100.00 — Others & N.E.S. 1,117 21 Note: Data for the 1978 population computed by the authorfrom 1978 Population Census Volume VII, p. 156. Of the 563 mothers of Form IV students grouped under *non' in Table (3), 446 {79%) have had adult literacy classes. Looking at the mother's level of edu- cation we find similar trends, with girls having more educated parents than boys. Only 165 (34%) of the girls' mothers had no formal schooling whereas 398 (59%) of the boys' mothers had none. Seventeen (14%) of the girls' mothers had more than Form IV education, and only 12 (2%) of the boys' mothers had this level of schooling. Comparing selectivity for the same schooling level, mothers have greater impact on selectivity than fathers. Whereas those with mothers with above Form IV education have 90 more chances of being selected than those with uneducat- ed mothers, the comparable ration for fathers is only two-fifths as large. A plausible explanation for this is that educated women of the parent's generation certainly come from families more privileged than the rest of the population. Hence, mother's education is a strong indication of the long standing prestige of the family and probably also of the level of family resources. It is more likely that such mothers socialize their offspring to value achievement in school to perpetuate the presitige of the family. Furthermore, since mothers tend to be at home more often than fathers, they can probably monitor their children's schoolwork and homework and give directions and assistance, par- ticularly if they are educated; hence the difference. However, in our two sam- ples, nearly all mothers with schooling were married to men of similar or higher educational attainment. All the highly educated mothers were married to high- ly educated men and therefore these students also came from elite paternal back- grounds. If we compare our findings of the Form IV students* selectivity by paren- tal education with those of UNESCO (1981), the message we get is similar; the higher the parental level of schooling, th ehigher the selectivity. However, UN- ESCO's findings show a slightly higher selectivity by increasing level of school- ing than ours. The differences could be attributable to sampling problems or differences in handling non-response answers of either the parent cohort or the sample students. 72 TABLE 4 SOCIAL SELECTIVITY OF FORM V STUDENTS BY FATHERS' EDUCATION 1978 Population Students by Father's Representation Ages 40 - 49 Education Indices N 070 N 070 None 335,717 54.14 427 25.27 Primary 1-4 180,464 0.47 29.10 471 Primary 5--8 88,756 27.87 0.96 14.31 513 Form 1-4 12,147 30.35 2.12 1.96 135 7.99 Above Form 4 2,981 4.08 lJ.48 144 8.52 All males age 40-49 620,065 17.75 99.99 1,690 100.00 Other & N.E.S. 3,219 37 TABLE 5 SOCIAL SELECTIVITY OF FORM V STUDENTS BY MOTHERS' EDUCATION 1978 Population Students by Representation Ages 40-49 Mothers' Indices Education N 070 N 070 None 564,521 87.84 822 48.41 0.55 Primary 1-4 61,694 9.60 565 33.27 3.47 l'rilmif, 5 --8 15,147 236 252 14.84 5.29 Form; 984 0.15 31 1.83 12.20 Above Form 4 328 0.05 28 1.65 33:00 All female, a1'-' I'" 49 642.674 100.00 1,698 100.00 Other & N.F.S. 1,1/7 29 As is the case with fathers of Form IY.'lUdents, some fathers of the Form V students in Table 4 have received an adulJ lJ!cracy education. Of those grouped under the 'none' level, 309 (72 percent) have had such education, The sex patterns are similar to those of the rorm IV sample. Whereas 51 (12 percent) and 37(, (30 perccnt) of the girls' and boys' fathers respectively have had no for- mal schooling. 72 (17 percent) and 72 (6 percent) of (he girls and boys' fathers respectiwly had more than !'orm IV edll;,~i()l1. From Table 4 we can see that selectivity increases as Ihe fat he! " level at education gets higher. The biggest gap, as in the 101m IV sample, is between Furm 1--4 and above Form 4. Those "tldents \\"ho comc from h,lIllcS \\here falll\.."' ban' above Form IV educa- !Jon have 4 m.'r\.: ,ll;l!,C,'S 01 bcing ,elected thcllI the adjacent group. Giv\.: ':1. nature of lanLanian o;ocicty, Llthers with su~h an educational level, this b...::i~ at the ,'" of independence, fended to assume lOp ICldership positions :n the government and civil service. During the colonial period, education for tile few who managed to gL't it was more a vehicle of change from traditional society to the modern sector than for social reproduction. Though it is still much the same today, there are signs of education beginning to play a reproductive role, especially among the highly educated. Some mothers in the 'none' group in Table 5 have had adult literacy train- ing; 640 (78 percent) of the mothers have attended such classes. Parents of the girls in the sample again seem to be n^ore educated than those of the sampled boys. Whereas 113 (27 percent) and^09 (55 percent) of the girls' and boys' mothers respectively had no schooling, 16 (4 percent) and 12 (1 percent) of the girls' and boys' mothers respectively had above Form IV education. If we compare these results with those of the mothers of the Form IV stu- dents (Table 3), we find minor differences in selectivity. The disparities for the Form IV students tend to be larger. This difference is probably due to biases in the sample, with an overrepresentation of the Form IV day scholars, particu- larly the girls' sample. Mothers of the day scholars are in urban areas, and women like men in urban settings, tend to be more educated than their counterparts in rural areas. , . From Tables 2, 3, 4, and 5 it is evident that the lower the parents educa- tional level, the lower are the chances of being selected either into O —level o r » A » _ level secondary levels. The more schooling the parents have, the-more, chances their children have of being selected. Here one discerns SES bias despite regional quota in favour of those with more schooling at the outset of secon- dary school and persists with only minor changes at the transition from Form IV to Form V. T «KI« For ease of comparison between the two sampled forms and sexes, labie 6 summarizes what we have been discussing concerning parents educational level and selectivity. TABLE 6 Students Selectivity by Parents' Education, Percent Students Selectivity by Fathers' Education None Form IV Form V Form IV Form V Primary 1—4 32.64 29.54 12.14 12.23 Primary 5—8 32.35 29.85 23.00 21.82 Form 1 — 4 24.52 28.04 36.96 37.41 Above Form 4 6.35 6.91 13.96 11.27 Total 4.14 5.66 12.94 17.27 N 100.00 100.00 100.00 100.00 677 12.73 487 417 74 Students' Selectivity By Mothers' Education Boys Girls Form IV FormV Form IV Form V Non 58.53 55.43 34.02 26.97 Primary 1—4 27.21 29.87 35.46 43.67 Primary 5—8 10.74 12.12 20.82 23.15 Form 1—4 1.76 1.64 6.19 2.39 Above Form 4 1.76 0.94 3.51 3.82 Total 100.00 100.00 100.00 100.00 N 680 1279 485 419 The data in Table 6 shows that 10.5% and 12.6% of the Form IV and Form V boys came from elite paternal backgrounds (at least having secondary educa- tion), while the respective percentages for girls were 26.9 and 28.5. Looking at maternal schooling level, 3.5% and 2.6% Form IV and Form V boys had elite mothers while for girls the respective percentages were 9.7 and 6.2. The differ- ences in selectivity between children of the elite and non-elite fathers and mothers at the Bonn IV and Form V levels based on the two sexes were statistically insig- nificant. If we combine the two sexes, although 17.4% and 16.5% of the Form IV and Form V students respectively came from elite paternal backgrounds, the difference is insignificant at the .001 level. Whereas 6.1% of the Form IV stu- dents came from an elite maternal background, only 3.5% of the Form V stu- dents came from such backgrounds. This means that differences in maternal level of schooling have a greater impact in selectivity .at Form I and decline as one moves up the schooling ladder. 75 If we again look at tne parental educational level as a whole, the differ- ences are significant at .05 level. The representation of students from elite back- grounds falls from 11.7^t in Form IV to 10.0% in Form V. If there are no sampling biases, the data suggest that there might be some favouritism when selection for Form I is made, or since most of the students pass through board- ing schools while in secondary school where all students are exposed to similar learning environments, unlike in primary schools where students live with their, parents, differences in home environment that are conspicuous in primary schools tend to be wiped out. Since these results include students from both private and public schools., we shall here consider those who completed form IV in public schools. Neverthele_ss, even after this consideration, the percen- tage still falls. Children from elite back grounds whose fathers are better edu- cated are 202, accounting for (17.4 percent) of the form IV sample, while then are only 198 (15.6 percent) of the form V students who completed form IV in public schools. Although the differences are statistically insignificant, the data suggests there might be favouritism in selection for Form I as people have often com- plained (Malekela, 1977; Mbilinyi, 1976; Uhuru, May 11, 1981). Besides parents' educational level, we also asked students about their par- ents' occupations. In a society where occupations in the modern sector are tagged to credentials, one's level of schooling canlenlalively suggest the type of possi- ble occupations. There are only seven occupational classification! in the 1978 census. This narrow classification makes it difficult to properly categorize some of the oc- cupations. One is not sure whether business should be classified under "sales/clerical" or "other workers." Due to this problem miscategorization Of some occupation may have arisen. This should not, however, blur the find- ings as we are more concerned with the major occupations available to the majority of Tanzanians. TABLE 7 SOCIAL SELECTIVITY OF FORM IV STUDENTS BY FATHER'S OCCUPATION 1978 Popu- Students by Represe- lation Father's ntation Ages 40-49 Occupation Indices N N % Farmers 509,016 82.41 664 57.00 0.69 Artisans 35,517 5.75 72 6.18 1.07 Businessmen/Clerical 18,262 2.96 159 13.65 4.61 Profession 17,872 2.89 164 14.08 4.87 Manag/Admin. 3,815 0.62 90 7.72 12.45 Other workers 33,180 5.37 16 1.37 0.26 All males age 40 49 617,662 100.00 1,165 100.00 Students, unoccupied, & N.E.S. 7.296 21 A Includes all other categories not in the researcher's agenda. The general picture is that the children of farmers are under represented and that the children of businessmen, professionals, and those in decision-making positions are over represented. The children of artisans are more or less fairly represented. 76 Mbilinyi (1976) reports that in her 1974 study of Form 1II students in pub- lic schools, 42.60/0, 45.8%, 9.5%, and 1.6% of the students' fathers in the sample were peasants, wage-earners, traders, and self-employed respectively. Since her sample was largely drawn from students in urban day schools, the higher per- centage of fathers in wage employment than in our case is to be expecteJ. The general trend, howevf;f, remains similar: children whose fathers are employed in the modern sector are over represented. / With occupations, as with education, girls' parents tend to be more privileged than those ofthe sampled boys. Whereas 200 (41%) of the girls' fathers were peasants, only 464 (68*) of the boys fathers were. At the other extreme, 157 (32%) of the girls had fathers who were in prof<:ssional and managerial or administrative occupations, while only 97 (14%) of the boys' fathers were in such occupations. The findings confirm that girls at the secondary level are more privileged than boys in terms of their social backgrounds. From Table 8 it is clear that students whose mothers are farmers are under- represented, whereas all other groUDS are overpresented. Only 341 (710/0) of tile mothers of the sampled. gIrlS were peasants. For boys, 604 (89%) of their mothers were in this category. In the top occupations were 75 (16%) and 31 (5%) of the girls' and boys' mothers respectively. TABLE 8 SOCIAL SELECTIVITY OF FORM IV STUDENTS BY MOTHERS' OCCUPATION 1978 Population Students by Representa- Ages 40-49 Mother's tion Indices Occupation N 070 N 070 Farmers 603,177 97.96 947 81.43 0.83 Artisans 444 0.07 19 1.63 23.29 Businesswomen/ Clerical 2,469 0.41 64 5.50 13.41 Professional 2,245 0.36 93 8.00 22.22 Manag/ Admin. 135 0,02 13 1.12 56.00 Other workersa 7,291} 1.18 27 2.32 1.97 All females age 40-49 615,769 100.00 1,163 100.00 Students, unoccupied & N.E.S. 28,509 23 a Includes all other categories not in the researcher's agenda. If we compare the results ill Table 9 with those of the Form IV fathers' occupations (see Table 7), the major patterns ramain more or less the same: children of farmers are under represented. The boys' fathers tend to be of low- er status than those of the girls. 840 (660/0) of the boys' fathers and only 174 (42070) of the giris' fathers were peasants. At the end of the scale, 226 (18%) of the boys' fathers and 144 (34%) of the girls' fathers were employed in profes- sional and executive positions. 77 TABLE 9 SOCIAL SELECTIVITY OF FORM V STUDENTS BY FATHERS' OCCUPATION 1978 Population Students by Representa- Ages 40-49 Father's Occupation tion Indices N % N 0/0 Farmers 509.016' 82.41 1,014 60.14 0.73 Artisans 35,517 5.75 88 5.22 0.91 Business/ clerical 18,262 2.96 180 10.68 3.61 Professional 17,872 2.89 270 16.01 5.54 Manag/ Admin. 3,815 0.62 100 5.93 9.56 a Other worker:; 33,180 5.37 34 2.02 0.38 All males age 40-49 617,662 100.00 1,686 100.00 Students, unoccupied & N.E.S. 7,296 41 a See note to Table 6 ~h __ TABLE 10 SOCIAL SELECTIVITY OF FORM V STUDENTS BY MOTHERS' OCCUPATION 1978 Population Students by Representa- Ages 40-49 Mother's Occupation tion Indices N % N % Farmers 603,177 97.96 1,450 85.50 0.87 Artisans 444 0.D7 12 0.71 10.14 Businesswomen I clerical 2,469 0.41 62 3.65 8.90 Professional 2,245 0.36 130 7.67 21.31 Manag/ Admin. 135 0.02 7 0.41 20.50 a Other workers 7,299 LIS 35 2.06 1.75 All females age 40-49 615,769 100.00 1,696 100.00 Students, unoccupied & N.E.S. "?8.509 31 a See note to Table 8. It would seem that agriculture IS the major occupation for the wumen. Although in the sample the percentages tend to be slightly lower than in the age cohort, its strong position is nevertheless remarkable. Those who are in non- agricultural occupations are so few that when the indices are calculated they tend to rise more sharply than when dealing with the fathers of the students. In litble 10, all others groups are over represented except the children whose mothers are farmers. 78 Girls' preeminence in privilege is reflected here too. Only 318(77 percent) of the mothers of girls were peasants, while 1,132(88 percent) of the mothers of boys were peasants. With the prestigious occupations, 58 (14 percent) and 79 ~ percent) of the girls' and boys' mothers respectively were professionals and' executives. Thus girls tend to come from more privile~ed backgrounds than boys at secondary levelsin Thnzania, as elsewhere(Bowman & Anderson, 1980). The provision of equality of access has some way to go, especially looking at ,the education of girls from an average family, who have very limited opportuni- ,ties to participate. Some research done in the past, such as Foster's (1965), Clignet's and Foster's (1966), lumped peasants or farmers into one group without making an effort to differentiate them. Recognizing the fact that farmers are not a , homogeneous group, in this study we tried to go deeper into the peasant ques- tion so that we could' say something more about those students whose parents are farmers. Thble II tries to clarify the issue. From Thble ll, it seems that only 61 percent and 67 percent of the Forms IV and V parents are real peasants as they depend on family labour only for their subsistence. Those who depend on family labour and casual labourers ,are relativelybetter-off farmers and comprise 29 percent and 28 percent of the Forms IV and V students' parents respectively.These can afford to hire casual labour- ers, probably their poor neighbours to work on farms which need extra hands. This could be done during harvesting seasons. TABLE 11 The Status of The Parents of Form IV and Form V Students Categorized as Farmers Form IV Form V N OJ. N % I. Dependent on family labour only 497 61 822 67 2. Dependent on family labour and casual labourers 233 29 342 28 3. Dependent on family labour and permanent labourers 61 8 SS 4 4. Dependent on labourers only 14 2 13 I 5. Have a plough 151 19 298 24 6. Own or hire a tractor 95 12 213 17 7. Grow cash crops 749 93 I,OSO 85 NOTE: Percentages can add up to 100 only from the first to the fourth row. The answers for the rest of the rows do not necessarily depend on what one answered in the previous questions. There were 805 and 1,232 students in the Ftrms IV and V samples respectively who said that their parents were farmers. The few farmers who depend on family labour and permanent labourers and those depending on labourers only are probably very rich ones. It seems they own considerable amount of land with perennial crops which require the presence of labourers throughout the year. Those few parents owning ploughs and owning or hiring tractors are the _riche~far~ers. Loo~ing at Thble II; pro~ably the majority, if not all of them 79 are drawn from rows two through four. In a country where the majority of the peasants are still dependent on the hand-hoe and family labour (Koley, 1973), the use of a plough or tractor signifies a higher status within the peasant com- munity. That 93 percent and 85 perce~lt respectively of the parents of the Forms IV and V students who are farmers grow cash crops is not surprising. Peasants were urged and are still compelled by the colonial and post-colonial governments to engi:1~ein cash-c~(Jp production (Coulson, 1982). It is the peasants' surplus productIOn that mamtams states and the unproductive bureaucrats in agricul- tur~lly based economies. With the little they get paid, the peasants have to pay their taxes to the government and the school costs to meet the education of their children. The few who do not grow cash crops may be getting some assistance through other means to educate their children. Of those claiming to be peasants, then, we can safely conclude that only 610le and 67% of form IV and Form V students' parents were either small hold- ing farmers or poor. The rest were relatively richer. Conclusion Judging from the student representations in the samples by parents' educa- tion and occupation, we may conclude that equality in access to secondary edu- cation has some way to go in Tanzania. Students from lowly educated parents and those engaged in occupations whose incomes are low are under-represented in the study. The higher the parents' educational level the higher is the represen- tation. Selection at Forms I and V tends to favour those who come from higher SES. Furthermore, is an imbalance between the two sexes. Girls form 32% of the students in public secondary schools (Siwale, 1982). But of those who get this opportunity of education, a good number of them come from relatively better-off backgrounds. Girls from poorer hackgrounds are hit hardest in this inequality of access to secondary education. Although some research in developing countries (see Heyneman, 1979; Saha, 1983) suggests that SES is not very important in scholastic achievement because there is less differentiation by home and language, parental expectations, and classes, the evidence from Tanzania indicate~ the contrary. Children from more educated homes have more chances of being selected at the two levels of secon- dary education than those from families with no more than a few years of school- ing. Selection to the two levels'ls governed largely by examination scores. lnvery selective systems like Tanzania's, the more educated parents know what is at stake and take the necessary measures, such as sending their children to 'cram' class- es after school hours, or making their children repeat (even if illegally) so that the chance of their children being selected is improved. As the privileged groups grow in number, it seems they are going to take up some of the places now taken by the offspring of the less privileged groups. (Studies by Foster 1965 and Weis (1979) attest to this trend in Ghana). While the professional and semi-professional groups' children constituted 23.5% of fifth form students in 1961 in Foster's study, the same group had 35,9% in 1974 in Weis' study. (The i