Objective Methodology

Study objective

The present study is intended to carry out a representative sample survey of households in the State of Telangana to obtain primary data on socioeconomic and health conditions of Muslims, Hindus and other minority religious communities in order to understand the achievement/progress of Muslims in various development indicators as compared with Hindus and other religious groups.

 

Study Methodology

The determination of the sample size for the present study is governed by several considerations, including the religious groups (Hindus, Muslims and other religious minorities), for which the estimates are required, the desired level of precision of the estimates, degree of representativeness of the sample and the availability of resources and time. A representative sample is one in which the characteristics of the sample are approximately the same as those of the target population. A good representative sample can be obtained by adjusting the sample size and using an appropriate method of selecting the sample units. Although sample size and representativeness are two related concepts, but they are in a way different issues. The size of the sample does not reflect its ability to accurately represent a target population. Large unrepresentative samples are as bad as small unrepresentative samples.

The size of the sample is determined primarily by two things: the availability of resources, which sets the upper limit of the sample size and the requirements of the proposed plan of analysis, which sets the lower limit. The decision to fix the optimum sample size cannot be made satisfactorily, because often we do not possess enough information to be sure that our estimated sample size is the best one. Broadly speaking, the sample size must be large enough (1) to undertake reliable and a meaningful analysis of cross-tabulations, (2) to provide for desired levels of accuracy in estimates of proportions/percentages, and (3) to test for the significance of differences between proportions.

 

1 Sample Size Determination

Keeping the above points in mind, we have to fix the sample size so that it represents the diverse characteristics of Muslim population in Telangana state and also the size of the sample is adequate to undertake a meaningful bi-variate analysis. According to 2011 census, 85.1% of the total population of Telangana state were Hindus, 12.7% were Muslims and the rest (2.2%) were distributed among a variety of minority communities such as Christians, Sikhs, Buddhists and Jains. The main focus of the present study is to examine religious differentials in socio-economic and educational indicators. Further, the religious differentials need to be examined separately for rural and urban areas. In the State of Telangana, Muslims are mainly concentrated in urban areas (75%), while Hindus are predominantly living in rural areas (70%). Keeping in mind the diversity in the settlement pattern of these two religious communities, we need to select sufficiently a large sample of Muslims spread over in all the 10 districts of the state.

 

A sample of 4000 Muslim households in the state of Telangana would be an adequate size for the present study. As the objective of the study is to examine religious differentials in several socio-economic and environmental variables, it is necessary to sample 3000 Hindu households to represent the population of Hindus in the state. Since the other minority communities constitute a meagre 2.2% population, it is proposed to cover 1000 HHs from other minority religious groups. Therefore, the sample size proposed for the study is as follows:

Religious Group Sample size(No. of HHs)
Muslims 4000
Hindus 3000
Other religious groups 1000
                       Total 8000

The sample of households is to be allocated to rural and urban areas based on proportion of rural and urban population according to religion in 2011. The distribution of sample households in each district will be based on proportion of rural urban population.

 

2 Sampling in Rural areas

A multi-stage stratified sampling design is proposed to be adopted for the selection of households in the sample units. In multi-stage sampling, the sample is selected in multiple stages.

Stage 1: All mandals in the district according to 2011 census frame will be arranged and 15% of the total mandals in each district will be selected by PPS sampling method. There are 464 mandals in the state and 15% of the total (70 mandals) will be selected.

Stage 2: In each of the selected mandal, villages are arranged according to population size and select a fixed number of villages randomly after stratifying the villages using certain criteria. Stratification is needed to ensure the inclusion of villages with diverse socioeconomic characteristics in the sample. All the villages will be stratified by a number of variables such as village size, percentage of population belonging to scheduled castes or scheduled tribes and female literacy. The number of villages to be selected from each stratum is proportional to the stratum size. In each stratum, a sample of villages will be drawn randomly.

Stage 3: The third stage involves the selection of households. If we want to use simple random sampling method, a list of households in the selected villages would need to be obtained. A rapid house-listing exercise will be undertaken considering the time and resources required. Alternatively, we may consider following the WHO recommended 30 by 7 cluster methodology for the selection of households.

 

3 Sampling in Urban areas

In urban areas, the sampling frame for the first stage of selection is the list of all wards in the state. The Primary Census Abstract (PCA) provides details of all the wards in each city or town in urban area. The list will be arranged by geographical regions and then by female literacy. A sample of urban wards from the list of wards will be selected by PPS sampling. From the selected ward, a sample of census enumeration blocks (CEBs) attached to the ward will be selected randomly, and then in the third stage households will be selected using the method adopted in rural areas.

 

4 Sampling in Hyderabad City

Hyderabad city is divided into five zones and 150 wards, ranging from 17 wards in the East zone to 50 wards in the Central zone as shown below:

Zone No. of wards
East zone 17
South zone 43
Central zone 50
West zone 14
North zone 26
Total 150

Stage 1: All wards with population in each zone according to 2011 census frame will be arranged and four wards in each zone will be selected by PPS sampling method. This gives us 20 wards representing all the five zones of Hyderabad city.

Stage 2: Each ward is divided into several urban blocks.  From each selected ward, 3 blocks will be selected again by PPS method. This gives us 60 blocks from 20 wards.

Stage 3: The third stage involves the selection of households. If we want to use simple random sampling method, a list of households in the selected villages would need to be obtained. A rapid house-listing exercise will be undertaken considering the time and resources required. Alternatively, we may consider following the WHO recommended 30 by 7 cluster methodology for the selection of households.