- The number of the measurement units in the target population is the first factor on which the approximation of the sample size will depend.
- The second factor on which the determination of the sample size depends is the initial level of the indicator.
- The third factor on which the approximation of the sample size depends is the magnitude of the change or comparison group differences that are expected to be reliably measured.
- The degree of confidence with which it can be expected that a significant change or a significant group difference will not have occurred by chance is the fourth factor on which the sample size depends.
- The degree of confidence for which it is expected that the significant change will be detected is the fifth factor on which the sample size depends.
For assistance in calculating your dissertation or research project sample size, click here.
The first two factors on which the sample size depends belong to the population characteristics. The last three factors on which the sample size depends are chosen by the evaluator or the survey designer.
Generally, the requirements for each indicator are considered in approximating the sample size needs for any particular survey. However, this task in relation to the sample size is tedious if the number of indicators is large.
This problem can be addressed with the help of the following two approaches:
The first approach is to approximate which of the indicators is expected to be most demanding in terms of the sample size, and to use the sample size required for that indicator. The biggest advantage to this type of approach is that it will automatically assure an adequate sample size for all the indicators to be measured.
The second approach is to identify a small number of indicators that are thought to be more important for any particular evaluation purpose and to limit the sample size computations. This approach assures an adequate sample size for the key indicators.
The drawback of this type of approach is that an adequate number of sample sizes might not be the same for other indicators that are more demanding in terms of sample size requirements.
An appropriate approximation of the sample size is crucial for economical reasons. If the investigator extracts a sample size that is smaller than the desired sample size, then the inference of the sample will not be appropriate or valid. If, on the other hand, the investigator extracts a sample size that is much larger than the desired sample size, then obtaining the inference of the sample would cost the researcher a lot and be tedious as well.
Generally, there is a budget for the study and this also affects the sample size to a great extent. Knowledge about the sample size is crucial in cases when data collection is expensive.
According to Peers (1996), sample size is referred to as one of the unified features of a study design that can influence the effect of significant differences, associations, or interactions.
