Research Methods: The Systematic Sample, The Stratified Random Sample

The Systematic Sample

Journalpapers.org When a population can be registered accurately or is limited, some sort of systematic selection provides what is closer to a random sample. The systematic sample consists of the selection of each list member. For example, if a sample of 200 is to be selected from a 200,000 phonebook list, the first name is selected by selecting the name randomly from the randomly selected page. Then each thousandth name next to the sample is selected

200 names are complete. If the last page is reached before the desired number is selected, the number will continue from the first page of the directory. A systematic sample of car owners can also be selected from a register or file from the state registry or from a sample of eighth-graders from the school attendance list.

The Stratified Random Sample

Sometimes it is advisable to divide the population into smaller homogeneous groups for a more accurate representation. This method generates stratified random samples. For example, in a study of the income of wage workers in a city, the correct sample is approximately equal to the relative sum of each socioeconomic level throughout the city.

If the proportion in society is 15% skilled workers, 10% managers, 20% skilled workers, and 55% unskilled workers, the sample should contain roughly the same proportions that are considered representative. Random selection should be made in each subgroup.

For a sample of 100, the researcher will randomly select 15 professionals from a subpopulation of all professionals in the community, 10 managers from that sub-population, and so on. This process provides the researcher with a more representative sample than a selected sample of the entire community, which may be overly weighted by the preferences of unskilled workers.

 The Area or Cluster Sample

Area or cluster samples are a variation of the simple random sample that is particularly suitable if the population of interest is not limited if there is no list of population members, or if the geographic distribution of individuals is widespread.

For example, if the researcher wanted to interview students in 100 high school dormitories in a large school district, they could first randomly select 10 schools from all secondary schools in the district. Then they can choose 100 randomly from a list of rooms in 10 schools.

 Nonprobability Samples

Unverified samples should be selected so that they do not match the sample. This sampling technique includes easy, voluntary, targeted sampling, and snowball sampling. Some uncertainty sampling techniques can produce samples that do not accurately reflect the characteristics of the population of interest. Such samples can yield unsubstantiated summaries and should not be used when random sampling is possible.

The appropriate sample consists of those available for the survey. Educational researchers often use appropriate samples (eg available classes) because of administrative limitations in random selection and assignment of people to experimental and control groups.

The state of the groups can be equated by statistical means such as analysis of covariance (see Chapter 11). In some types of descriptive studies, the use of an appropriate sample may limit generalizability to that population. For example, if a psychology professor uses Introductory Psychology as a class subject, the professor can safely generalize only to other groups of similar psychology students.

Sample Size

There is usually a trade-off between the desirability of a large sample and eligibility of a small sample. The ideal sample is large enough to provide a fair representation of the population the researcher intends to summarize and small enough to be economically selected - in terms of the availability and cost of quantities in terms of time and money.

There is no fixed number or percentage of subjects to determine the appropriate sample size. This may depend on the type of population of interest or on the data to be collected and analyzed. A national public opinion poll randomly selected a sample of about 1,500 subjects to reflect the opinions of a population of over 150 million voting-age US citizens with an error rate of 2-3%.

 More important than size is the careful selection of the sample. The ideal method is a random selection, which allows the law of chance or probability to determine which members of the population to select. Using a random sample, regardless of whether it is a large or small sample, can assess sampling error by giving researchers an idea of ​​the confidence they can have in their results.

In summary, here are some practical observations about sample size:

1. The larger the sample, the lower the sample error rate and the more likely the sample is representative of the population.
2. Study studies should usually contain a larger sample than is needed in experimental studies because the study results come from those in the volunteer sense.
3. If the sample needs to be divided into smaller groups for comparison, the researcher must first select a sample large enough so that the subgroup has a size appropriate for its intended use.
4. For questionnaire surveys, since response rates can be as high as 20-30%, a large initial sample should be selected to receive the email so that the final number of questionnaires returned is large enough to allow the researcher to make minor sampling errors.
5. Subject availability and cost factors are reasonable considerations in determining an appropriate sample size.

Share this post

Berlangganan update artikel terbaru via email: