1. Non Probability Sampling
The non probability sampling may or may not necessarily represent the whole population as in this type of sampling the sample selection is not done by the random means; unlike the probability sampling. Owing to the fact that random selection procedure is tyrannous and time consuming as it needs the scattered samples from the whole population, the non probability sampling is done purposively, according to the judgment of the researching team that which population or group can fulfill their most of the research demands. The generalization phenomenon cannot be preferably employs on that as it may not necessarily represent the whole population as the selection of sample was biased.
2. Suitability of Non-probability Sampling
Non probability sampling is not applicable every where rather there are certain limitations and criteria specified for the usage of this sampling considering the requirements of the research being carried on.
1.The preferred use of sampling would be in the case when there is need of demonstrating the specific characteristics in certain groups or populations.
2.The studies, in which the parameters like exploratory research, qualitative or the pilot study is to be incorporated, are facilitated by the use of non probability sampling.
3.When the random selection is not possible as in the case of unlimited population then the selection is made according to the principles of non probability sampling to cover almost all the population in to the research.
4.The researches where generalization should be made should avoid the incorporation of this type of sampling as the findings of the non random selection could hardly infer the generalized results applicable to entire population.
5.The main advantage connected with this type of sampling is that all the techniques are less time consuming, less budget requiring and less tiring.
6.The initial work out plan for the probability sampling can be prepared through the non probability sampling as the less time and resources are required.
3. Types of Probability Sampling
Main categories of probability sampling are accidental (haphazard or convenience) sampling and purposive sampling and the other sampling division includes include Snowball Sampling, Quota Sampling and self selection sampling. Purposive sampling however has a wide number of applications in research as the problem plan is being followed up in this type of sampling.
3.1. Haphazard, Accidental or convenience Sampling
In the convenience sampling as indicated by the name, the sampling is done according the convenience of the researcher. If we want to access any population we may stand at any point of our choice and chose out the subjects for our sample. If the rules are followed, the convenience sampling is most suitable for execution and in contrast to probability sampling its cost can easily be managed at comparatively low level. However the main disadvantage associated with this type of sampling is that the samples are not selected at random and the biasness is very much intervened in it. The easy access and recruitment of the subjects is its main advantage. For instance if the research club for the environment is to be formulated, it would be easy to collect the volunteer subjects by standing at the department entrance and asking them whether they would like to be a part of the research or not.
3.2. Purposive sampling or Judgment Sampling
The purposive sampling or the judgment sampling is more related to the useful purpose or the aim in mind and the subjects are chosen accordingly. The main belief in this type of sampling is that all the individuals would not be safe for sampling, few individuals can stay fit in the criteria of research, aiming the subjects to be purposive. The purposive sampling is done in the limited population where the subjects are chosen according to their research expertise.
3.3. Snowball Sampling
The snowball sampling is carried on according to the reference of the first selected person to the next person known to him and then he refers to the other known and so on making the research possible. The snowball sampling effectively proceeds when the sampling size is preferably small as snowball sampling could not effectively incorporate the large population size. In snowball sampling the population representation could not be attained as the subjects identification is done in chain and most probably the identified people are from the same group having more or less similar attributes. The people that are represented in snowball sampling are hard to achieve otherwise as they are involved in socially restricted characteristics, taboos or hidden realities like that of HIV patients, drug addicts, prostitutes etc.
3.4. Quota Sampling
The subjects are given proportionate chances of representation in the quota sample providing equity in different sectors on the basis of quota. The quota is usually applied in the divisions of gender, religion, socioeconomic status, education, race, and age. The researchers usually take the quota system into integration when they need specific involvement of different parameters like gender while studying the college students, the classes when researching about their behaviors and so on.
3.5. Self Selection Sampling
The samples taken from the institutions or organizations are in the form of units or case. The sampling that involves organizations in the study seeks the involvement of case or the unit for their research is made in accordance with self selection sampling.
The methodologies involved in the self selection are of wide range, of which the preferable method is online questionnaire in which the subjects of specific division of certain entrepreneurs are invited to take part according to their own will as they are not forced to become units for research. [linkunit]Another method for selection of the case or unit may be the advertisement for the experimentations of effects of certain parameters; for instance the drug trials and the associated effects of it on the subjects. The main interest of the self selection sampling is that the research is conducted on those volunteers that are not selected by the personal contact or motivation rather because of the volunteer interest of the person to become a part of that research. The main advantage of the this type of sampling is that it requires less or no time and energy in locating the volunteers; however the disadvantage is that the finding could not be applicable to the whole population, could be self biased or the inferences may be based on the exaggeration of certain factors as posed by the cases or the units.
•Deville Jean-Claude (1991), "A Theory of Quota Surveys", Survey Methodology, Vol. 17, pp 163-181.
• Deming WE (1960), "Sample Design in Business Research", John Wiley and Sons, New York.
• Neyman J, "On the two different aspects of the representative method: the method of stratified sampling and the method of purposive selection", Journal of the Royal Statistical Society, Vol. 97, pp558-606.
• http://www.experiment-resources.com/non-probability-sampling.html. Retrieved on May 15, 2011.