There are different types of research, and the choice of proper tools and tests can be listed among the key components of success. Taking into account that the reliability of study results strictly depends on methodological decisions, significance tests can never be selected without preliminary research. In order to choose the most appropriate tests, any researcher is required to take the key properties of study samples into consideration.
All types of tests serve specific purposes, and prior to test selection, it is extremely important to generalize on the specific characteristics of samples. As it is stated by Cooper and Schindler (2014), the first issues to be considered are the number of data sets that need to be analyzed (one, two, or more) and the type of data (numerical or categorical). Nowadays, numerous tools such as between-groups t-tests allow working with a few data samples.
Apart from that, the great role of possible links between individual cases in one data set or between separate samples is to be recognized. Decisions related to significance testing also need to be made with reference to the properties of data, data collection methods, and measurement scales used.
Thus, the combination of the type of scale, the degree to which samples are independent, and the number of samples is the basic information needed for decision-making. For beginners in academic research, it can be quite difficult to generalize on the properties of different significance tests and keep this information in mind, and this is why special guides and tools remain extremely popular.
Among them are decision-making trees used by students and specialized algorithms. To use some tools, it can be necessary to include other details such as the exact or approximate sample sizes, similarities or differences between samples, or data transformations.
