Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Transcript of SAMPLING TECHNIQUES
Method: One method is to take the list or map and give each unit a number, write the numbers on individual slips of paper, put them in a bag and mix the slips up thoroughly, and then draw out the number of slips required. Alternatively a random number table can be used. If no suitable list or map exists, it may be possible to use participatory methods to solve this problem.
Systematic Random Sampling
Uses: where very large numbers are included in the target population and simple random sampling is difficult. Or where lists are already grouped into sections or classes.
Method and challenges: there are many possible systems e.g. by taking every tenth name for every fifth name.
Uses: where it is particularly important to explore the range of different potential impacts eg ensuring that the quota for women includes a selection of single women, very old women, a literate woman and so on.
Stratified Random Sampling
Uses, Methods and Potential Problems
Uses: when the target population is very large and/or geographically dispersed making simple random sampling extremely expensive and time-consuming.
Quota sampling: quotas for certain types of people or organizations are selected for interview.
a complete set of elements (persons or objects) that possess some common characteristic defined by the sampling criteria established by the researcher
The selected elements (people or objects) chosen for participation in a study; people are referred to as subjects or participants
Potential problems: may be very costly particularly where populations are geographically dispersed and/or individuals are difficult to trace because of for example marriage or migration. Even apparently complete lists may systematically exclude some relevant categories of respondent. In particular lists of registered entrepreneurs are likely to exclude women in enterprises. Conversely lists of female credit beneficiaries may not be a reliable basis for selection of credit users. Whether or not this matters will depend on the nature of the inquiry.
a group of people are selected in a systematically random manner from a complete list of a given population.
Potential problems: Similar to simple random sampling. It is also crucial that the system selected does not bias the sample. For example selecting every tenth name from a list compiled of groups of ten members where the first name in each group is that of the President.
When populations are divided into subgroups depending on particular characteristics.
Uses: when the nature of the issues to be investigated means that it is important to give respondents from particular subgroups an equal chance of representation and this would not happen through random sampling.
Method: the relevant characteristics to be used for stratification are identified on the basis of the questions to be asked e.g. membership or non-membership of an organisation, female or male members. A random list is then drawn up for each subgroup and respondents chosen randomly within each.
Potential problems: the identification of the characteristics for classification of respondents is crucial and may need to be refined during investigation.
Where clusters are randomly selected and all individuals or households in particular clusters are interviewed
Method and challenges: Clusters maybe geographical, for example villages or markets. They may also be for example microfinance groups or particular social categories within geographical locations e.g. all upper caste households.
Potential problems: It is important to ensure that important subgroups are not left out and also to consider any potential bias in analysis. For example if all the clusters thrown out by random selection are large villages, are the results likely to be different if some of the villages have been very small e.g. because of few facilities or different social structure.
Non-random sampling is a sampling technique where the samples are gathered in a process that does not give all the individuals in the population equal chances of being selected.
Uses: when the nature of the issues to be investigated means that it is important to give respondents from particular subgroups a chance of being selected which is disproportionate to their numerical strength e.g. where it is important to include a significant number of respondents from minority populations, female entrepreneurs etc.
Method and challenges: The categories for which quotas are to be used and the quotas to be allocated are determined based on the issue to be addressed. Common criteria are age, gender, occupation and whether people live in project or non-project areas. The quotas are fixed depending on the types of issues to be investigated but respondents within each quota category are selected randomly.
Potential problems: The categories on which quotas are based are crucial and may need to be refined as the investigation progresses.
similar to quota samples but where respondents within each quota are selected to represent diversity.
Method: selection of respondents is based on prior analysis and hypotheses of the different possible types of impact on different stakeholders.
Potential problems: it is important to be continually reflexive in response to information as it is obtained to ensure that diversity is properly understood and captured.
Chain sampling or snowballing: A first contact is selected and interviewed and then asked to suggest other interviewees and so on.
Uses: This method is useful for identifying minority groups or occupations within communities.
Method: it is important that all suggested interviewees are followed up in order to avoid bias. Questions may be cumulative to build up a complete picture of the particular population under study.
Potential problems: The chain may be biased because of the particular networks chosen. This can be overcome through probing investigation and/or combining with eg a random walk or selecting a number of such chains by another random method.
-Sherelyn C. Nuñez