The population is divided into several groups known as strata. If a given population has distinct characters, it is represented as a heterogeneous population. In such instances adequate care shall be taken to ensure proper representation to all the distinct characters of the universe. In order to ensure proportionate representation of all the characters, the technique of simple random sampling is not considered suitable.
Therefore, an improved device is adopted in the name of stratified random sampling. In applying this technique, the population is divided into several groups called strata. Each group of the population is called a stratum. A group comprises units with a particular homogeneous character. All the units of the universe possess one chief homogeneous character but can be divided as separate parts, each of which again possesses one uniform character. For instance, a sample of 100 students has to be selected out of 1,000 students in a college. From among 1,000 students, which constitute the universe, say we find 600 boys and 400 girls. Broadly, all are students and possess the homogeneous character. But this universe can be divided into two groups, viz; boys and girls, i.e., strata. If proportionate representation has to be given to each group,, the sample shall include 60 boys and 40 girls (on 10% basis). By making use of the proportionate stratified random sampling technique, the sample observations can be drawn on an equal proportion basis from each stratum with the help of simple random sampling technique. This proportionate stratified sampling procedure is expected to yield good results specially when there is no great difference in dispersion from one stratum to the other. However, in case of larger variations among different strata, this is certainty not the most efficient procedure. In such cases it is necessary to assign greater representation to a stratum with a larger dispersion and a smaller representation to one with small variation. Such sample selection is expected to lead to maximum efficiency.
In disproportionate stratified sampling, an equal number of cases will be taken from each stratum without any consideration for its percentage share in the universe. In the sense, smaller groups are giving relatively greater weight when compared to the larger groups in the universe.
In either case, it is necessary that stratification shall be made with utmost care so that observations in each group comprise homogeneous elements. The application of simple random technique for selection of observations in each stratum will ensure equal chance to all the observations. The investigator is concerned more about adequate representation to all the characters of the universe and it is immaterial whichever unit is selected at random from each homogeneous group.
Merits:
Therefore, an improved device is adopted in the name of stratified random sampling. In applying this technique, the population is divided into several groups called strata. Each group of the population is called a stratum. A group comprises units with a particular homogeneous character. All the units of the universe possess one chief homogeneous character but can be divided as separate parts, each of which again possesses one uniform character. For instance, a sample of 100 students has to be selected out of 1,000 students in a college. From among 1,000 students, which constitute the universe, say we find 600 boys and 400 girls. Broadly, all are students and possess the homogeneous character. But this universe can be divided into two groups, viz; boys and girls, i.e., strata. If proportionate representation has to be given to each group,, the sample shall include 60 boys and 40 girls (on 10% basis). By making use of the proportionate stratified random sampling technique, the sample observations can be drawn on an equal proportion basis from each stratum with the help of simple random sampling technique. This proportionate stratified sampling procedure is expected to yield good results specially when there is no great difference in dispersion from one stratum to the other. However, in case of larger variations among different strata, this is certainty not the most efficient procedure. In such cases it is necessary to assign greater representation to a stratum with a larger dispersion and a smaller representation to one with small variation. Such sample selection is expected to lead to maximum efficiency.
In disproportionate stratified sampling, an equal number of cases will be taken from each stratum without any consideration for its percentage share in the universe. In the sense, smaller groups are giving relatively greater weight when compared to the larger groups in the universe.
In either case, it is necessary that stratification shall be made with utmost care so that observations in each group comprise homogeneous elements. The application of simple random technique for selection of observations in each stratum will ensure equal chance to all the observations. The investigator is concerned more about adequate representation to all the characters of the universe and it is immaterial whichever unit is selected at random from each homogeneous group.
Merits:
1. Administrative Convenience:
As compared with simple random sample, the stratified samples would be more concentrated geographically. The time and money involved in collecting the data and interviewing the individuals may be considerably reduced. The supervision of the fieldwork could be allotted with greater ease and convenience.
2. Representative:
Stratified sampling ensures a desired representation of various strata in the population. It overrules the possibility of any essential group of the population being completely excluded in the sample. Stratified sampling thus provides a more representative cross-section of the population and is frequently regarded as the most efficient system of sampling.
3. Greater Accuracy:
Stratified sampling provides estimates with increased precision. Moreover, stratified sampling enables us to obtain results of the known precision for each of the stratum.
4. Sometimes the sampling problems may differ markedly in different part of the population, e.g., a population under study consisting of
(i) literates and illiterates or
(ii) people living in institutions, hostels, hospitals, etc. and those living in ordinary homes.
In such cases, we can deal with the problem through stratified sampling by regarding the different parts of the population as strata and tackle the problems of the survey within each stratum independently.
5. Stratification is of great advantage when the distribution of the universe is skewed.
Demerits
As compared with simple random sample, the stratified samples would be more concentrated geographically. The time and money involved in collecting the data and interviewing the individuals may be considerably reduced. The supervision of the fieldwork could be allotted with greater ease and convenience.
2. Representative:
Stratified sampling ensures a desired representation of various strata in the population. It overrules the possibility of any essential group of the population being completely excluded in the sample. Stratified sampling thus provides a more representative cross-section of the population and is frequently regarded as the most efficient system of sampling.
3. Greater Accuracy:
Stratified sampling provides estimates with increased precision. Moreover, stratified sampling enables us to obtain results of the known precision for each of the stratum.
4. Sometimes the sampling problems may differ markedly in different part of the population, e.g., a population under study consisting of
(i) literates and illiterates or
(ii) people living in institutions, hostels, hospitals, etc. and those living in ordinary homes.
In such cases, we can deal with the problem through stratified sampling by regarding the different parts of the population as strata and tackle the problems of the survey within each stratum independently.
5. Stratification is of great advantage when the distribution of the universe is skewed.
Demerits
- It is a very difficult task to divide the populations into homogeneous strata. This may require considerable time, money and statistical expertise.
- If different strata of population overlap such a sampling would not be representative.
- The supplementary information to set up strata is not available sometimes.
No comments:
Post a Comment