In probability sampling each element has a probability of being including in sampling. It means that in probability sampling, the selection sample from the universe has an equal chance to all the members. There are four types of Probability Sampling.
- Simple random sampling
- Stratified random sampling
- Systematic sampling
- Cluster or multi-stage sampling
Simple Random Sampling
A simple random sampling is one in which each unit in the universe has an equal chance for selection. This type is based on chance sampling. The methods used in this type are: Lottery System, Table or random numbers, Arrangement of the number in the same order, Computer. These are the equipment used for selection of simple random sampling. If these are not available, we use the method of a dice put in to the container. When a dice is put only once and drawn. This method is called sampling with displacement. If the dice is putted again and again into the container and then draped out is called method of replacement.
- It is free from errors.
- It is not based on prejudice & bias.
- It is more representative to the universe.
- It is more scientific.
- Random selection is difficult.
- This is not used in heterogeneous universe.
- It does not work well in widely dispersed universe.
- There is lack of knowledge about universe.
Stratified Random Sampling
When the universe is highly heterogeneous, a stratified sampling may be used. In this type, the universe is divided into sub-universes or sub-population on the basis of some characteristics like age, sex, education, religion etc. After division of universe different strata is made and then select sample from each stratum or sub-universe. Stratification is the division of heterogeneous universe into small homogeneous social packets.
- Homogeneity out of heterogeneity is achieved.
- It is more representative.
- It provides grounds for future research.
- Statistical accuracy is achieved.
- It is more time consuming.
- It is more cost.
- It leads to errors because of unequal size of strata.
- It needs more technical skills.
Systematic sampling is very easy in the universe where it is in systematic order like Hayatabad where street and houses are numbered and in a specific order. It is difficult in villages where no order exists in houses. In systematic sampling a sample may be chosen at regular interval of time. It is also called sampling by regular or fixed interval. For example, we select a sample of 100 students out of 1200 in a school. The sample size is 100, the systematic fraction would be, 100 (size of sample)/1200 (universe) =1/12th.
The sample interval is 12. So, every 12th student would be included to our sample. The first one would be selected randomly by picking a piece of paper. If it was 6th them others will be taking intervals like, 6,18, 30 etc.
- It saves, time, cost and energy.
- It is representative to the population.
- It is easily drawn and worked out.
- It proves efficient from simple random sample.
- Finding, arranging and selecting of items is difficult.
- More heterogeneous will lead to errors.
- Higher ratio of items will lead errors.
Cluster or Multi-Stage Sampling
It is the process in which we select groups and clusters instead of individuals. In such cases we do not select units individually like a student but we select a group of units like a school. These grouping of units are called clusters. Clusters are often geographical units like (villages, district, wards) etc. Clusters mayor may not be equal in size. Suppose we select a sample of 320 dwellings out of 12000 in a city. The universe is divided into various wards. Suppose 375 such wards have been identified and the average number of dwellings in each ward is 32. (12000/375) now we select 10 wards out of 375. The number of dwellings would be 320 which is our required sample. Cluster sampling is also known as area selection sampling because it is based on a specific area. Multi-Stage sampling is like that, in some eases it is different from cluster sampling in multi-stage sampling we can pass through various stages by random method.
- It is easier in case of large sample.
- It is economical and less time taken.
- It combines units into groups & clusters.
- It is less representative.
- Its findings cannot apply for another area.
- It may contains basic elements