Process of Selecting Participants in the Research - Essay Example

SELECTING PARTICIPANTS IN THE RESEARCH Sampling: this is the step of determining who will be in the research as participants –not everyone can be included a sample needs to be representative of the population to be studied as much as possible KEY CONCEPTS IN SAMPLING Sample: The sample is representative of the population. A representative sample has the same distribution of the characteristics as the population from which it is selected Results from a representative sample can be generalized to a wider population. Generalizability is important in needs assessments and so a positivist approach is more often employed as it emphasizes generalizability. Information rich the key concept in interpretive research means that the sample has cases from which you can learn about issues central to the research question Element: refers to the item under study in the population and sample in generalist social work may be the client systems e.g. individuals, families, organizations or communities. Elements depend on the unit of analysis Sampling frame: This is the list of all the elements in the population from which the sample is selected. TYPES OF SAMPLING METHODS There are 2 types : 1. Probability 2. Non-probability Probability sampling: The sample has each element with a known chance of being selected for the sample. It increases the representativeness of the sample and should be preferred in the positivist approach Non- Probability Sampling Aka purposive sampling the research picks the sample according to the nature of the research problem and phenomenon being studied more suited to interpretive studies generalizability is less important it has less representativeness. PROBABILITY SAMPLING: There are four types of probability sampling 1. simple random sampling 2. systematic random sampling 3. stratified random sampling 4. cluster sampling Simple random sampling -The simplest method – the population is a whole unit and each element has an equal probability of being selected because the sampling is random Sampling method Generalizability Simple random Can generalize; minimal limitations Systematic random Can generalize; limitations minimal-note how the elements are listed in the sampling frame Stratified random Can generalize; limitations minimal- make sure the strata involved are reflected in the analysis of the data Cluster Can generalize, but some limitations possible- note the characteristics of the elements because there is a possibility of sampling error with this type of probability sampling Systematic random sampling: using a list of the elements every nth element is chosen the nth element is determined y the size of the sampling frame ( i.e the population) and the size of the sample needed e.g. if there are 200 elements in the sample frame and a sample of 100 is needed then every second elements is chosen Stratified random sampling a modification of the two above the population is divided into strata ( layers) samples are then taken from each layer can help ensure that certain proportion of the elements is represented than in the two examples above After dividing the population into strata e.g. Native American, Hispanic and others for an after school program for 7th graders, simple for systematic random sampling is carried out from each stratum/ layer in proportion to the stratum’s representation in the population Disadvantage: 2 preconditions must be fulfilled: 1. ensure that membership in the group whose representation you are concerned about has an impact on the phenomenon you are studying 2. You have to know the proportion of this group relative to the rest of the group. Cluster Sampling a large unit with the elements of interest is random sampled , then sampling from this unit the elements to go into the final sample Useful when it is different to a sampling frame e.g. after school program- if you have to get the lists from each school it will be long instead of from the school district office. simple, systematic or stratified sampling can be used in this first sampling Problem: when few units are sampled there is more chance for sample error. NON PROBABILITY OR PURPOSIVE SAMPLING allows the researcher to select elements that are information rich and so is the best for interpretive studies e.g. after school program can decide to find out the problems and experience of families who need after school care rather than find out the proportion of families who are in need of after-school services there are 8 commonly used types of purposive sampling 1. typical cases 2. criterion 3. focus groups 4. key informants 5. community forum 6. quota 7. snowball availability Typical cases most commonly used cases are chosen from literature review, other research or consultation with relevant groups Criterion all cases that meet some criterion are picked e.g. all eligible to receive services from the agency Focus groups picked purposely from people who have an interest or expertise in the topic good to ensure participation by the community Key informant those picked are people in the community who are experts in the field of interest Community forum a series of public meetings are held, the issues are discussed ad opinions sought from the public, member cans be purposely picked Quota sampling a certain number / amount of element with specific characteristics are purposely included in the sample. same as stratified random sampling but no randomness in selection you have to know the number of the element to be considered Problems: you have to ensure that the variable is important to the study in question Snowball sampling some members of the population are picked and then used to contact others useful for less accessible populations e.g. homeless Availability sampling aka convenience sampling it involves including available /convenient elements in the sample its results cannot be generalized to the population in the study because the elements are picked for their convenience not the richness in information e.g. interviewing people at shopping mall to get a sample of the community useful for practice evaluation Purposive or Non probability Sampling Methods Method Characteristics Typical cases Those with typical characteristics Criterion Participants selected according to some eligibility criteria Focus groups Those with an interest in the research topic Key informants Those with expertise on the research topic Community forum Open to a community; some can be purposely invited Quota Certain proportions of participants from different groups selected according to specific characteristics Snowball some participants identified; these are participants then identify others with certain characteristics Availability Those selected because they are available STUDYING COMPLETE POPULATIONS It is sometimes necessary to study the whole population than a sample as in the case of a program evaluation al the elements of the population should be included Examples include clients served in the last 6 months clients served since the start of the program all children with don’s syndrome in a medium –sized city SAMPLE SIZE -deciding on sample size is not a problem that usually confronts generalist social workers because their sample is fixed i.e. own practice the size depends on its homogeneity/ the similarity among different elements, the closer it is the smaller the sample can be - the size depends also on the research approach positivist: when using probability samples need larger samples Interpretist: in these studies, the sample can be small because what matters is the information richness of the sample. any larger sample than needed leads to redundancy -sampling error which is the extent to which the values of the sample differ from those of the population. -The margin of error means the precision needed for the research e.g. a 5% margin of error means that the variation in either direction could be as much as 5% .If 55 % clients are satisfied it could be as much as 50- 60 % in reality the larger the sample the less the margin of error and the sampling error The extent of sampling error should be reported in large scale surveys. also you can say that the results have a 95 %confidence level i.e. how often the same results can be obtained if research is repeated Table: size of sample required at 5% confidence level population size sample size 50 44 75 63 100 80 150 108 200 132 250 152 300 169 400 169 500 217 750 254 1,000 278 2,000 322 4,000 351 5,000 357 10,000 370 15,000 375 20,000 377 25,000 378 50,000 381 100,000 384 1,000,000 384 Source: Educational and Psychological Measurement, R.V Krejcie and D.W. Morgan, pp. 607-610.1970 sage publication. THE AGENCY AND SAMPLING Generalist social workers sometimes have to use sampling methods that are not text book examples e.g.: a. limited probability sampling b. combined sampling methods Limited Probability sampling A limited probability sample from a neighborhood or agency can be used to compare its characteristics with those of a sample from a larger population. This is to allow for a generalization over a bigger area community or county. Without a large sample it is not possible to do so. This is because it is not possible to do i.e. agencies lack time and money Problems: - Often the researchers assume that they know what the important characteristics are when comparing a smaller sample with a larger one Combined Sampling: - Whenever it is not practical to use a planned method another sampling method may be used sometimes a different method for different groups. Smith and Kronauge (1990), studied decision making in abortion choice .with a random sample of women who received at one agency in st. Louis, Missouri they compared two characteristics of this sample –age and status- with a national sample. The similarity allowed them to tentatively generalize the findings from the sample to the wider population (147). Kruk (1994) while studying disengaged noncustodial fathers in United states , Canada, and Great Britain managed to get court records in Canada for a probability sample but not so in Great Britain ,so he use availability sampling methods : he placed ads in the newspapers for referrals from legal and social work practitioners (148). ETHICAL ISSUES IN SAMPLING 1. responsible reporting 2. informed consent Reporting the sampling method -the researcher should report the research methods as accurately as possible -it is unethical to claim that the results of a no probability sample are generalizable Sometimes authors report random methods of sampling which is actually availability or other form of non probability sampling. Watch the sampling method and generalization limitations when reading reports. Informed consent no coercion should be used inform the participants of the purpose and goals of the study Dilemmas: 1. Disclosing information about the study can affect its results e.g. participants can change their behavior if they know they are being observed 2. when subjects know their participation is voluntary they can choose no to Solutions: 1. Creates an egalitarian relationship between researcher and subject. The subject participates actively in the research and the need for consent becomes irrelevant. Methods like focus groups, key informants and community forum. This approach is a feature of participatory research. 2. stress the integration between research and practice when explaining to subjects .show that they will benefit too Advantages of egalitarian relationship: 1. increased responsiveness 2. validity may be enhanced 3. important for social work where relationship with the client is emphasized HUMAN DIVERSITY ISSUES IN SAMPLING 1. Generalizations made from studies of homogenous populations should not be done e.g. a) white middle-class samples have been studied and findings generalized and used as the norm to evaluate minorities b) Sometimes when programs are effective with one minority group then it is assumed to be so with other minority groups which may not be the case 2. Researchers should be conscious of other diversity issues other than race and ethnic, like ageism and homophobia which are still common even among researchers. The exclusion does not have to be conscious Read More