The accuracy of a research report depends on several interrelated things, like:
- Research objective must be clearly articulated
- Research design must be correctly accomplished
- Correct sampling procedures and techniques must have been used
- Data collected correctly and analyzed correctly
Mistake at any point can negate excellent design, sampling technique used and questionnaire design. Beyond the conceptual differences, many kinds of error can help explain differences in the output of the programs that generate data on income. They are often classified into two broad types: sampling errors and non-sampling errors.
“Sampling error” occurs when a researcher gathers information from respondents who are in some ways different from the true population of interest. A sample error is a statistical error resulting from the fact that any given sample has an opportunity produce different estimate of a parameter of a problem. It is because inferences about the entire population are based on information obtained from only a sample of that population.
The coefficient of variation is a measure of the extent to which the estimate could vary, if a different sample had been used. This measure gives an indication of the confidence that can be placed in a particular estimate.
Therefore, one must consider in the broadest of terms the universe of potential respondents from which a sample should be drawn. Thinking broadly about your target population will help ensure you attain a representative sample. It may not always be easy to reach the correct market research survey sample frame but not doing so can lead to missing information and missed opportunities.
Non sampling errors are numerous in breed, they are pervasive is in nature and are highly arbitrary. Unlike sampling errors they do not decrease the sample size; in fact they tend to increase the sampling size.
Although sampling errors can be estimated their direction, magnitude and impact are generally unknown. As a result these are the most frequent errors in any research design. Their reduction is dependent on improving the research methods ad procedures dealing with the handling of data itself.