From Indicators of Enrolment to Attendance Rate: A Critical Review of Frequently Used Indicators



National Institute of Educational Planning and Administration

17-B, Sri Aurobindo Marg, New Delhi – 110016 (INDIA)




Free and compulsory education to all children up to the age fourteen is constitutional commitment in India. At the time of adoption of the Constitution in 1950, it was decided to achieve the goal of universal enrolment within a period of ten year. Keeping in view the educational facilities available in the country at that time, the goal to achieve universal enrolment was ambitious to achieve it in a short span of ten years. Therefore the target date was revised a number of times despite significant improvement in all the spheres of elementary education. The government has recently launched a new programme Sarva Shiksha Abhiyan to achieve the goal of universal primary education by 2007 and that of universal elementary education by 2010. To review the progress made in the area of elementary education, a set of indicators concerning to different aspects, such as, universal access, enrolment, retention and quality of education are constructed over time and analysed at different levels. To assess the progress made during the decade 1990-2000, a lot of activities were initiated globally as a part of EFA: The Year 200 Assessment.  As a part of this exercise, in India a set of 18-core indicators concerning to elementary education and adult literacy were developed both at the all-India and state levels. Keeping in view the availability of data, a few of the 18 indicators were not constructed but at the same time a few additional indicators were constructed with special reference to conditions in India. Computation of out-of-school children, transition from primary to upper primary level of education and a few indicators concerning to access were form part of this additional set of indicators.


So far as the goal of universal enrolment is concerned, a variety of indicators were constructed to measure the progress made between 1990 & 2000 among which enrolment ratio is the most prominent one.  A variety of enrolment ratios are available all of which have some sort of limitations. In this note, first all such indicators are critically analysed and then the discussion is focused on Attendance Rate that is proposed a better indicator of children coverage than the traditional enrolment ratios. This is followed by a discussion on possible limitations in getting data on attendance.

Review of indicators of enrolment

The basic indicator that gives idea about coverage of child population (in a system) is the intake (entry) rate which is simply division of enrolment in Grade I to corresponding population at which a child is supposed to enter into the system (in most of the cases it is either '5' or '6').  However, while calculating the entry rate, repeaters are not considered and only fresh (new) entrants in Grade I are considered. This is because of the fact that repeaters are not the members of the present cohort but they have entered into the system some one or two years back. In case of gross enrolment (including children below & above ‘6’ in Grade I), the entry rate calculated is known as Gross Entry Rate otherwise it is known as Net Entry Rate. Entry rate is also known as Admission or Intake rate that demonstrates capacity of the system with regard to availability of schooling facilities. While calculating net entry rate, net enrolment (new entrants) in Grade 1 of age '6' is considered. A gross entry rate of 80 per cent means that about 80 per cent children (of entry age) including the overage and underage one are enrolled but a net entry rate of 80 per cent means that only 20 per cent children of entry age  ('5' or '6') are out of the system or are yet to be enrolled. Net entry rate is considered a better indicator of student coverage at the entry point (Grade I) than the gross entry rate. Unless the net entry rate is brought to hundred percent, the goal of universal enrolment cannot be achieved. Entry rate is also useful in knowing likely enrolment in subsequent grades in years that follow. However, in many systems age-grade matrix is not available and hence net entry rate cannot be calculated.

By enrolling all children of age-6 do not guarantee itself that the goal of universal enrolment will be achieved at its own, it is a necessary condition but not the sufficient condition. Children are to be retained in the system and should also acquire minimum levels of competencies. For that purpose other indicators, such as, Gross and Net enrolment ratio, dropout & retention rate, transition from primary to upper primary level and achievements levels should also be analyzed. The intake rate gives idea about coverage of child population of entry age-6 in Grade I but it fails to give any idea about children those who entered and then remained in the system in years that follow. For this purpose indicators concerning to enrolment ratio and retention need to be analyzed. A variety of ratios, such as, Overall, Gross (GER), Net (NER) and Age-specific enrolment ratios are available for this purpose. The overall enrolment ratio presents the overall view of the entire education system where as GER and NER presents information about the coverage of child population at a particular level, such as, primary and upper primary level of education. On the other hand age-specific enrolment ratio presents information about coverage of a particular age or age group. While assessing progress made between 1990 & 2000, as a part of EFA 18-core indicators, GER and NER were computed and analyzed.


            The GER is division of enrolment (total) at school level í' in year 't' by a population in that age group á' which officially correspond to that level í'. Thus for calculating GER at primary level, total enrolment in primary Grades I-V irrespective of ages is considered which is then divided by the corresponding age-specific population, 6-11 (6+ to 10+) years to obtain GER. Similarly, total enrolment in upper primary grades VI-VIII is divided by the corresponding population 11-14 years (11+ to 13+) to obtain GER at upper primary level. This means that overage and underage children are included in GER, which resulted into GER more than hundred percent in many locations. In locations with small population, a slight over reporting of enrolment may also result into GER more than hundred. The GER is therefore considered crude indicator of child coverage and may present misleading picture of the true situation.  A GER more than hundred do not imply that the goal of UPE is achieved because of the overage and underage children. Alternatively, net enrolment of a particular age group is considered in place of total enrolment. One such indicator is Net Enrolment Ratio, which is an improved version of GER.


In NER, overage and underage children are excluded from enrolment and then ratio to the respective age-specific population is obtained. For example, enrolment in Grades I-V of age 6-11 years is considered which is than divided by 6-11 years population to obtain NER at the primary level. Similarly, NER at the upper primary or the entire elementary level can also be worked out. A NER of 77 per cent at the elementary level implies that 23 per cent children of age 6-14 years are still out-of-school. Unless these children are brought under the education system, the goal of universal elementary enrolment cannot be achieved. Achieving hundred percent NER does not itself guarantee that the goal of UEE will be achieved at its own. Those who enrolled will have to retain in the system up to the end of an educational cycle. The NER and other indicators should be calculated separately for boys and girls and in rural and urban areas and also at different administrative levels, as it would help to identify areas/locations that need immediate attention.


NER is considered a better indicator of enrolment than the GER. However, the limitation of NER is that it excludes overage and underage children from enrolment though they are very much in the system. The calculation of NER requires age & grade matrix that in most of the systems is not available. Alternate to GER and NER, age-specific enrolment ratio may be considered which gives enrolment ratio for a particular age or age group.  For example, an age-specific enrolment ratio of age '7' will include total enrolment of age '7' irrespective of grades which is then divided by the single age population '7' to obtain the ratio. The limitation of this ratio is that it considers total enrolment than enrolment in a particular grade that corresponds to age '7'. The calculation of age-specific ratio requires age-grade matrix, which as mentioned above is not readily available in many locations. An age-specific enrolment (age-7) of 67 per cent implies that 67 per cent children of age-7 are enrolled but it is not known in which grade are they enrolled. Or alternatively it can be said that 33 per cent children of age-7 are yet to be enrolled (in Grade I).


As it seems from the above discussion that Net Enrolment Ratio is a better indicator of enrolment than other indicators of enrolment. It presents coverage of child population of a specific age group in relation to corresponding grades.  In other words, it gives in percentage terms how many children of a specific age group are enrolled and at the same time also presents estimates of out-of-school children at that point of time. The calculation of net enrolment ratio needs age & grade matrix, which as mentioned above is not available in most of the cases. Sporadic attempts have been made to collect information on age & grade matrix but the same is not available on regular basis both at the state as well as the country level. Information on age & grade matrix is being collected in the DPEP districts but the same cannot be used to generate state-specific estimates of overage and underage because of the limited coverage of districts in a state. Till such time, the existing estimates from the Sixth All India Educational Survey conducted in 1993-94 should be used to know the percentage of overage and underage children both at the primary and upper primary levels of education. However, the same is not readily available at the district level as the publications containing district-specific data in case of most states is either not available or they do not contain this set of data. Whatever limited data is available on age & grade matrix is not free from the errors of measurement. For instance in India, enrolment is collected from the recognised schools only where as the unrecognized private institutions which are large in number is not included in the annual collection of statistics. Data on age & grade matrix is obtained from the class registers where the date of birth of each and every child enrolled is written.  But in the process of transmitting age (in year) from the date of birth many approximations take place; hence the age & grade matrix is not free from errors (lot of confusion prevails so far as 5+ or 6+ or 6-11 or 5+ to 10+ population). Further, the date of birth it self may not be correct especially in rural areas where birth certificates are generally not available. On the discretion of the parents or even teachers the date of birth is recorded in the school registers.


Can attendance be a better indicator of enrolment?

The discussion presented above suggests that unless all children of age 6-11 years are enrolled, the goal of universal primary enrolment cannot be achieved. This is also true for other age groups, like 11-14 and 6-14 years. However, by enrolling children it self does not guarantee that the goal of universal enrolment will be achieved. It has been observed that children those who are enrolled do not attend schools regularly. For instance in India, compared to a GER of above 90 per cent at the primary level, the corresponding attendance rate is only 65 per cent. At the upper primary level also, the attendance rate is much lower than the corresponding GER and NER. Therefore indicators, such as, GER and NER cannot be considered better indicators of enrolment. Alternatively, it would be better to consider average attendance rate at different levels of education, which can be calculated either on daily, monthly, quarterly or even on annual basis. Keeping in view the availability of data, the attendance rate may either be gross or net in nature. The attendance rate is one of the important indicators of monitoring. For that purpose, it should be calculated separately for boys and girls and also at different levels. The school-specific attendance rates will help to identify schools that need immediate attention. Monthly attendance, if monitored properly will highlight possible reasons of low attendance and whether it is because of boys or girls, harvest season, festival season or because of the migratory population can also be known. All this is not possible to analyse in traditional enrolment ratios. Across countries, attendance rate is generally not available and is not part of the regular collection of statistics. In fact, it should be built-in in the information system itself.


Attendance rate can be calculated in relation to the number of school working days and children actually attending a class. For example, in a Class of 45 students in a school that functioned for 22 of 30 days in a month, attendance rate can be calculated in accordance to the actual number of days children attended schools. Some of them might have attended school for all the 22 days while others may not have. First, the maximum possible present days (attendance) is calculated by multiplying number of school days to number of students in a class. In this case it would come out (22 x 45) a total of 990 present days (care should be taken in case of schools that have tradition of marking attendance twice a day, in the first and last period. In that case both the maximum possible attendance days as well actual present days will be changed accordingly).  Now actual number of present days (number of days students actually attended a class) is counted in that month by observing the class register. Let us suppose that it come out to be 600 student present days. The average is calculated simply by dividing 600 by the maximum possible present days (990). This will give an average monthly attendance of 60.61 per cent in a class. By following the same procedure, average attendance in other classes and separately in case of boys and girls can be obtained either on daily, monthly, quarterly or annual basis. Once the average attendance is obtained in all the classes of a school, the same may be used to obtain average attendance for that school. In that case, first total student present days in a month are obtained by adding present days in different classes, which is then divided by the maximum possible present days (all classes) in that month. This can be obtained by multiplying school working days to total number of students in different classes in a school. Once the school-specific average attendance rates are calculated, it can be used to calculate the same at different levels. The above set of attendance rates are based on school registers, which should be built-in, in management information system.  Alternatively, attendance rates can also be worked out on household sample basis.  This was initiated recently in India and Gross, Net and Age-specific attendance rates were worked out. These rates are worked out in relation to total number of children in an age group attending school. If the attendance rate is calculated by considering all the children in Classes I-V, including the overage and underage children, the rate obtained is called Gross Attendance Rate. Otherwise if overage and underage children are not considered and only enrolment of a specific age group is considered in calculating rate, the rate thus obtained is termed as Net Attendance Rate. Similarly, age-specific attendance rate can also be calculated by considering a specific age children attending schools.


The GER, NER and Age-specific Enrolment Ratio can be adjusted in the light of actual average attendance. A GER of 95 per cent at primary level with 65 per cent attendance will give an adjusted-GER of 62 per cent. Similarly a GER of 59 per cent at upper primary level with 43 per cent attendance will give an adjusted-GER of 25 per cent. The adjusted-GER suggests that though 95 per cent children (including overage and underage) are enrolled in primary classes but only 62 per cent of them attend schools regularly. The corresponding figures at the upper primary level is 59 per cent against adjusted-GER of 25 per cent. But how ‘average attendance’ should be defined is an important question. Similarly who will be termed as ‘regular student’ and how migratory and nomads children will be treated are another important areas of concern.


Can reliable attendance rate be generated?

However, obtaining accurate attendance rate is a challenging task. Data users often question reliability of educational data and the official set of enrolment is found inflated. This is also reflected if the official set of data is compared with the corresponding statistics of the NCERT collected through All India Educational Surveys. A significant gap irrespective of an educational level is noticed both at the all-India and provincial levels and also in case of boys and girls. Information on attendance can be collected through teachers only, which like enrolment may not always present the real picture. Generally, three sets of enrolment are available in schools. First, the number of students whose names are written in the class register, second those who are marked present and third those who are physically present in the class on the day of visit. The third one in most of the cases is found lower than the second one and the second one lower than the first one. It may also be recalled that in developing countries, specifically in South Asia a number of incentives are being offered to children to improve both the enrolment and attendance. For instance, in India mid-day meal is one such scheme under which all primary school children are entitled to receive rice/wheat at the rate of 100 grams per day provided that they attend school for not less than 80 per cent of the total working days in a month. This has suddenly increased both enrolment as well as attendance across the country. Independent observers are of the opinion that in many cases the improvement in attendance is not genuine and like enrolment it is also inflated. The entire country is covered under the mid-day meal scheme. Schools that are covered under the scheme and have lifted the grains have at least 80 per cent attendance by default. In many locations, even it is found above 90 and even hundred percent that may be genuine or may also even be inflated. Thus obtaining attendance data from school registers through teachers may not bring forth the real picture about children attending schools. The same if collected from households may also not likely to improve the reliability of the attendance rate. However, advantage of HH survey is that children those who are enrolled in private unrecognized institutions are also covered in survey, which is not true in case of information collected from schools as a part of regular collection of statistics. The respondent in household surveys in most of the cases is the head of the household and not the members of the house. The head of the household is authorized to provide answer whether children in his/her house attending schools regularly. But how ‘regular’ is defined and interpret is an important mater. A student attending school for 50 per cent of the working days in a month will be considered regular or a student who attend schools for 75 or 80 per cent of the total working days. Can the head of the household provide this information accurately? This is doubtful especially when a large number of head of the households themselves are illiterate or literate without completing any schooling level.  The only option therefore left to collect reliable information on attendance is through visiting schools without prior notice. Naturally, this can be done on sample basis only. But who will conduct the survey is a moot question. Community, as it seems is the only option left for this purpose. What would be the frequency of such surveys and feedback mechanism are other important questions which needs to be properly addressed before such surveys are launched.