Sampling methods

Résumé du document

As time and resources available are limited, we cannot interview all members of a population. Therefore, researchers use sampling to obtain the information they need. They select a limited number of people (a sample) to represent the characteristics of a whole population. There are different methods to select a sample. We can divide them into two major groups: the probability sampling and the non-probability sampling. A good sample survey can give important and interesting results. However, a bad sample survey (wrong people interviewed) can have disastrous results on a company, for example. When you use a probability sampling method, you must find some process or procedure that assure that the different units in your population have equal probabilities of being chosen in the sample for the survey being conducted. It is scientific, operationally convenient and simple in theory. You give a number to every member of the population. Then you use a table of random numbers, a computer random number generator or a mechanical technique (to close your eyes and pull out the numbers which refer to people) to select your sample. Finally, you pull out a number and you interview the person who has this number.

Extraits

[...] It means that you have to divide the population into groups. However, the persons of these groups have a common characteristic like the age, the race, the socio-professional group ( Then you use the technique of the simple random sampling: You give a number to every member of your strata. Finally, you close your eyes and pull out the numbers that refer to people, and you interview them. This method is quite interesting because you are able to represent key subgroups of the population, especially small minority groups. [...]

[...] Errors from chance will cancel each other out in the long run, those from bias will not. Standard error It is the standard deviation of the values of a given function of the data (parameter), over all possible samples of the same size. Sources Internet http://trochim.human.cornell.edu/kb/sampling.htm http://www.ubmail.ubalt.edu/~harsham/stat-data/opre330Surveys.htm http://www.isuma.net/v02n03/mendelsohn/mendelsohn_e.shtml http://www.cems.uwe.ac.uk/~pwhite/SURVEY1/node30.html http://www.cems.uwe.ac.uk/~pwhite/SURVEY1/node31.html http://www.education.tas.gov.au/internalaudit/manual/10Techniques.htm http://www.jesus.ox.ac.uk/~barron/eap/data%20collection.pdf http://www.cems.uwe.ac.uk/~pwhite/SURVEY2/node10.html Books Marketing of Geoff Lancaster & Paul Reynolds Elements of marketing of A.R. Morden Marketing of Steven J. Skinner Marketing: a case study approach of David Stokes Marketing research of George Breen & A.B. [...]

[...] Quota Sampling This method implies personal judgement on the researcher's part. Quota sampling has some similarity to stratified sampling however the selection of respondents within strata is non-random, but is left to the interviewer. This non-random element is its greatest weakness and quota versus probability has been a matter of controversy for many years. Statisticians criticise it for its theoretical weakness. Market and opinion researchers defend it for its cheapness and convenience. In quota, sampling each interviewer is given an assignment of interviews based upon quotas that are representative of the whole population. [...]

[...] In addition, you have a quick reply and you do not need any sampling frame. Nevertheless, you have no idea how representative the information collected about the sample is to the population as a whole. In addition, there are many biases, because the persons interviewed, are not representative of the population. Moreover, it is difficult to check your answers. Therefore, the quality and the accuracy of your survey are limited. However, this method is often the only feasible one, particularly for students or others with restricted time and resources, and can legitimately be used provided its limitations are clearly understood and stated. [...]

[...] - Your frame sampling: your frame is not accurate. A respondent A respondent is a person who is interviewed in an opinion poll. Researchers are taking a keen interest in his opinion and in his behaviour. A sample It is a limited number of people who represent the characteristics of a whole population. A random sampling You select your sample by chance. You use techniques in which each member of the population as the same probability to be interviewed. For example, you pull out a number in a box and you question the person who has this number. [...]