In this type of sampling, the researcher makes a choice of his subjects depending of how easily he can access them. It is usually easy to perform since it does not require the researcher to use complex mechanisms of sampling. He/she needs to choose from the subjects who present themselves where the researcher is stationed. For example in this research aimed at measuring attitudes about tuition reimbursement, amongst all employees at a national cable company, convenience sampling may involve issuing questionnaires at the entrance, whereby employees will be accessed with ease. On the other hand, the employees can be asked to volunteer for interviews. The employees who will volunteer are the only ones who will be interviewed.
VeritableEssays is the most reliable custom writing service on the internet today
The results of this kind of sampling may not be representative since when the questionnaires are issued at the entrance; it will be assumed that all the employees will pick one. Some may be field workers and therefore may not have a chance of being sampled. On the other hand, asking employees to volunteer for a survey regarding reimbursement may cause over representation of the volunteers who posses tough opinions since they are most likely to volunteer. Some may not volunteer since they lack the courage. In general, the results of such a survey are likely to be biased and unrepresentative of the overall population.
This is sampling in which each individual has a chance of being chosen as a subject. There are several kinds of random sampling that can be applied. These include; “simple random, stratified random, systematic, and multistage sampling” (Salant and Dillman 1994). A researcher who uses random sampling is likely to acquire unbiased results due to the fact that the researcher has no influence in the choice of samples. The samples become true representatives of the overall population in the sense that any of the employees can be interviewed. This helps in avoiding the problem of under-coverage bias. The results are also likely to be precise especially due to the high rate of response that is characteristic of random sampling.
Even though the results are largely unbiased, there is always a possibility of non-response bias. This situation in the research may be caused by some of the employees who may not adequately respond to questionnaires due to lack of sufficient information concerning the subject. The sampling error depends on the sample size in random sampling and therefore the researcher may increase the accuracy of the results through using a large sample size.
Stratified Random Sampling
This is sampling whereby the population is divided in to strata and then a simple random sample selected from each stratum. These strata consist of groups of people with a common characteristic. For example in this case, the employees can be divided in to groups depending on their salary brackets, or the number of years they have worked for the company. The results obtained from such kind of sampling are a good representation of the total population as well as the critical smaller groupings present in the population. In the survey, the researcher will be able to represent the minority groups amongst the employees.
The results are likely to be relatively precise than other kinds of random sampling. Another important factor in regard to the results obtained is the fact that only the relevant subgroups will be in focus. The irrelevant ones may increase non-response bias, a problem that this sampling helps to avoid. The researcher can also use varied sampling techniques, thereby improving the accuracy of his/her inference. The statistical power of testing variations between the strata is more impartial since the samples taken are equal regardless of the size of the strata. Stratified random sampling will generally have more statistical precision than simple random sampling. The results are therefore likely to be unbiased.
Salant, P. and Dillman D. A. (1994). How to conduct your own survey. John Wiley & Sons, Inc.