Differences, Examples, and Investigator’s Concerns
Inferential statistic employs parametric (PTs) and non-parametric (NPTs) tests of statistical significance. PTs require either interval or ratio data (at least one of the variables has to belong to one of these groups), normal distribution, and the independence of the compared groups along with random assignment or selection (Rubin & Bellamy, 2012, p. 157). According to Rubin and Bellamy (2012), it is possible to employ PTs if some of the conditions are not met, but it is not advisable since the results may become unreliable. Similarly, Qualls, Pallin, and Schuur (2010), provide evidence to the fact that inappropriate use of PTs results in a higher probability of error. As a result, in the case when the conditions are not met, NPTs are used.
In other words, PTs are the tests that need certain assumptions about the population’s parameters to be met, and NPTs “do not make such assumptions” (Anastas, 2012, p. 481). PTs are concerned with a mean (the “measure of central tendency showing the middle of a distribution by summing up all scores and taking the average of them”), which is a suitable measure of central tendency in the case of the normal distribution (Anastas, 2012, p. 560). If the distribution is skewed, the mean is less likely to be suited for measuring it; here, in NPT, median (“mid-point of distribution”) or mode (the value that is encountered most often) should be used (Anastas, 2012, p. 560).
In general, NPTs are more convenient and can be applied to a wider range of circumstances while PTs ensure a lower risk of error. Still, many NPTs are considered equivalent to their respective PTs (Anastas, 2012). Examples of PTs include the t-test that is the most popular of the kind: it is a classic PT as it should be used to “compare the mean differences between two groups on an interval- or ratio-level dependent variable” (Rubin & Bellamy, 2012, p. 160). The NPT counterparts of t-test include, for example, the sign test that is used with ordinal dependent variables (Rubin & Bellamy, 2012, p. 172). The investigator needs to take into account the specifics of the data that they have and the data that the test is fit for and use appropriate variants that are quite numerous nowadays (Rubin & Bellamy, 2012).
Cocreation of Possibilities: Data Analysis Plan Discussion
Currently, the following plan can be proposed for the study of the Cocreation of Possibilities intervention created by Bergdahl, Benzein, Ternestedt, Elmberger, and Andershed (2013). The two tools suggested for the study will provide ordinal data that will reflect the changes in the patients’ quality of life and relationships with the nurse practitioners (NPs). The said data will then be descriptively analyzed for every patient (to define if and how their state changes), NP (to determine how their effectiveness changes), and the two groups (to describe how the performance of the group changes). In the process, the ordinal data will be turned into ratio data, and the sets of the teams will be compared with the help of a test. A PT might be used in case the rest of the conditions are present (in particular, normal distribution), but it is not very likely. If PTs are not an option, an NPT is typically used (Rubin & Bellamy, 2012, p. 157). For the same reason, it is more likely that mode will be able to become the reporting measure of central tendency, but the final data might affect this choice. From the current research, it appears that the Mann–Whitney U test will be suitable to test both null and research hypotheses (to test if the experimental group will exhibit better performance with respect to particular variables). This test is an NPT that is suitable for skew distribution, any type of data, and the comparison of two groups, which makes it fit for the current research as well (Rubin & Bellamy, 2012, p. 172).
References
Anastas, J. (2012). The research design for social work and human services. New York: Columbia University Press.
Bergdahl, E., Benzein, E., Ternestedt, B., Elmberger, E., & Andershed, B. (2013). Co-creating possibilities for patients in palliative care to reach vital goals: A multiple case study of home-care nursing encounters. Nursing Inquiry, 20(4), 341-351. Web.
Qualls, M., Pallin, D. J., & Schuur, J. D. (2010). Parametric versus nonparametric statistical tests: the length of stay example. Academic Emergency Medicine: Official Journal of the Society for Academic Emergency Medicine, 17(10), 1113-1121. Web.
Rubin, A., & Bellamy, J. (2012). Practitioner’s guide to using research for evidence-based practice. Hoboken, N.J.: Wiley.