COMPLEX PROBABILITY SAMPLING
Sampling: is meant to select a part of elements in population, and conclusions obtain may be about the entire population.
Probability Sampling: probability sampling is based on concept of random selection, a controlled procedure that assures that each population element is given a known non zero chance of selection.
Complex Probability Sampling: an efficient sample in statistical sense is one that provides precision while using a smaller sample size. A sample that is economically efficient and should provide desired precision and accuracy at a lower cost, that is to choose such a design that is cheap in data collecting (less travel and interviewer time). Complex probability sampling considers following four approaches; systematic sampling, stratified sampling, cluster sampling and double sampling.
A better and versatile approach in this; every kth element in the population is sampled; starting with a random start ranging from 1 to k. this approach is flexible & simple. With using this approach, there is no need to assemble cards in a file, just to choose the total cards and sampling ratio, and random start; then begin drawing sample by using every kth card. Systematic sampling is statistically more efficient than simple random sample, when dealing with similar population elements grouped together. One difficulty with systematic sampling is periodicity; that may yield bias results. Another concern is the monotone population. That may yield to similar results.
Populations are sub divided into sub populations or sub groups (starta). The approach by which this decision is made (that is: what elements to include from subgroups) is named as stratified sampling. The process follows the following steps. After dividing population into subgroups, a simple random sample can be taken out from each subgroup and there after results can be combined according to population. There are three main reasons to choose stratified sampling approach .
- This approach increase statistical efficiency
- This approach provides adequate analysis of sub populations.
- This approach enable us to use different research method and procedure for each and every subgroup /starta.
This approach is statistically more efficient as compared to simple random sampling, this approach is also helpful when research aim is to study the characteristics of a particular sub division of population. This approach is also helpful when different methods of data collection are applied in different parts of population. While using this approach, usually a primary variable is chosen, for which we believe will correlate other variables too.
This approach is defined as the division of population into groups of elements with some groups randomly selected for study. Two reasons behind using the cluster sampling are:
- This approach is more economic efficient than simple random sampling.
- This approach can be applied even in unavailability of a practical sampling frame.
This approach is statistically less efficient the reason is that clusters are usually homogeneous, but economic efficiency overcomes this less statistical efficiency.
This approach is also named as “sequential Sampling” or “multiphase sampling” and is define as “to collect some info by using sample and then to use this information as a base to choose a sub sample for in-depth study” this approach is called as double sampling. This is mostly use with combination to stratified and cluster sampling approaches.
Double sampling process includes collecting data from a sample using a previously defined technique; based on the information found, selecting a sub sample for further study.