Nationwide Student Sampling

Continuing with Basic Concepts of Sample Design for Educational Survey Research by the International Institute for Educational Planning, we get to one of the toughest aspects of survey research – finding an accurate sample from clusters spread out over a wide area. Focusing on the educational system, an ideal survey would have a very large sample that draws from all students in every school within the country. Yet this is implausible, as the nationwide effort alone necessary to handle that style of research would be substantial. So the author suggests the following method of developing a sample from such a large, diverse, and spread out population.

Define the Basic Sample Design Your first step is to develop the basic characteristics of the sample, including the strata included, the number of schools, the target population, the number of strata, the selection equation, and pseudo schools. Describe the Population Explain the population you are looking to study, exclusion criteria, and what the desired result of the sample would describe. Prepare the List of Schools Based on the target population you are hoping to address, prepare a list of all of the schools that are included within that criteria, along with relevant details.

Turn the Information into Data The next step suggested by the author is to take all of the definitions and explanations listed above and turn it into relevant data that can be used as reference. Select the Variables Each strata is going to have its own characteristics that may be broken up into strata. Use the data to group them into variables that are going to be included in your data.

Apply Variables to the Spreadsheet Once again, the next step is to turn the information you have found into data that will be included in the spreadsheet by each of their strata. Allocate the Samples The strata you’ve included will need to be parceled out for the target population, in order to account for any weighing that needs to be placed on the samples. Consider Alternate Data The author describes the optional step of breaking the samples out further for analyses.

Breaking out the samples should only be used for research purposes, but should not alter the way you have broken out the sample. Identify the Pseudo schools Pseudo schools are schools that need to become a part of a larger sample in order to collect relevant data, often because the size of the school is too small. Distribute Random Sample Using what the author refers to as “lottery tickets” the researcher is then expected to distribute these tickets to the relevant schools in order to prepare selection of the school aspect of the sample.

Select the Sample Use the lottery tickets to select the schools that will be asked to provide students for the sample. Select the Students Finally, select the students that are going to participate in the research project, based on the schools that turned up within the lottery tickets.

The specific details of the design can be found in the article. However, the above 12 steps are the suggested methods that the author uses to describe how best to take a sample from schools across the country.