Introduction
Sherri (2012) points out that the T-test is a statistical test where the statistic that is being tested follows a t distribution when the null hypothesis is supported. The test was first used way back in 1908 by a chemist who was working in a brewery in Ireland. This paper aims at assessing how the t-test was applied in the study entitled Characteristics of Nurse Practitioners Interested in Participating in a Practice- Based Research Network (PBRN) by Sharon M. Meyer and James J. Werner. The findings of the study were published in the 22nd volume of Journal of American Academy of Nurse Practitioners.
The above research article incorporates non- parametric tests (or what Sherri, 2012, refers to as t-tests). In this paper, the author will provide a brief summary and report of the statistics, with the aim of finding out whether the assumptions of the test were met. The author will also analyze whether the data used was appropriate for the test. In the paper, the research will describe the statistics they intend to analyze in the article, provide a brief description of the study by the two authors, describe how the statistics were used in the study as well as explain how this usage was appropriate or inappropriate for the particular study. The author will also analyze how the two authors met (or failed to meet) the assumptions made and address the appropriateness of the level of measurement. This is in addition to a discussion of the how the data was displayed and an examination of the appropriateness (or lack of it thereof) of the data displays or presentation.
Study Description
Practice- based research networks (herein referred to as PBRNs) are ambulatory practices which are mainly devoted to providing primary care to patients. PBRNs also provide information on how community- based programs are being implemented. This is especially important when it comes to monitoring and evaluation of the programs. In this regard, they have become essential laboratories that make it possible for researchers to conduct primary care studies. However, there is lack of enough information on the number of nurse practitioners (herein referred to as NPs) who are interested in participating in PBRN. The study in focus outlined the characteristics of those NPs who expressed interest in participating in PBRN in the northeast (NE) Ohio region and how they compared to the characteristics of state and national NPs. The researchers contacted 1016 NPs in NE Ohio through the internet and mail. They collected information on the characteristics of their practice and demographics.
The study was to find out whether there was a significant difference between the characteristics of the NE Ohio NPs and state and national NPs. It was also meant to determine whether the characteristics of NE Ohio NPs were significantly different from those of the control group depending on whether or not the informants would participate in future PBRN studies. The study concluded that the characteristics of the NPs in NE Ohio were comparable to those of their counterparts in Ohio and the US in general. There was no significant statistical difference between the NPs who practiced in ambulatory settings and had interests in participating in PBRN and those who had no interest (Meyer & Werner, 2010).
How t-Test was used in the Study
In this study, the t-test was used to compare the proportion data and means of the NE Ohio NPs with those of the NPs from Ohio and the entire country. The t-test- in combination with the Fisher’s test and Chi-square tests- were also used to compare the practice and demographic characteristics of the NPs in NE Ohio who were involved in ambulatory practice and were willing to participate in future PBRN studies with those from the same region who were not interested. The number of hypotheses was quite large and so the researchers chose to include Benjamini- Hochberg correction (Meyer & Werner, 2010).
The use of t-test in this particular study was very appropriate. This is because it enabled the researchers to compare the means of the three samples in focus. By using the t-test, they found that there was no significant statistical difference between the means of the samples. Keller (2012) points out that the t-test is appropriate when the variances of the samples in question are similar. This is because the test assumes that the variances are similar and the populations have a normal distribution.
However, even when this assumption is violated (like in the case of the study in focus), the t-test is still robust enough and provides useful information even when the variances are not equal. The t-test is also very appropriate because of the ease with which researchers were able to make calculations and analyze data. Many statistical software packages such as SPSS, MatLab, and MiniTab have very easy-to-follow instructions for conducting t-test calculations. The t-test was also very appropriate because the sample was large enough.
Level of Measurement
The researchers used questionnaires to gather information on the practices and demographics of NPs. These measures included gender, ethnicity, race, date of birth, membership to professional organizations, RN education, and the number of years they have been practicing as NPs. There were more questions for those NPs who practiced in ambulatory care. These included queries on practice setting, practice location, practice ownership, and hospital privileges. The last question probed whether the informant would prefer participating in PBRN studies in the future or not. This information was then compared with that gathered from NPs in the state and around the nation.
The ratio level of measurement was used in this study. This level of measurement was most appropriate because the study included the comparison of ratios, in this case means (Rubin, 2010). By using ratios, the researchers were able to compare the differences between the characteristics of the NPs in NE Ohio and the Ohio NPs in general. They were also able to compare characteristics of the NE Ohio NPs and those of nationwide NPs.
Display or Presentation of Data
The findings of the study were ‘displayed’ in three tables. The first table included a comparison between the means of the practice and demographic characteristics of NE Ohio NPs and Ohio NPs and NE Ohio NPs and national NPs. Additionally, the table contained p values from the t-test to indicate whether there was any statistical significance in the comparisons. The second and third tables contained information on the nurses who were involved in ambulatory care and were either interested or not interested in contributing to further PBRN research. The measures that were highlighted in the first table were the same as those that were presented in Table 2 and Table 3. The difference between the two tables was that Table 2 presented the means and corresponding p value while Table 3 presented the percentages and corresponding p values (Meyer & Werner, 2010).
The three tables presented the right information in a simple way but the researchers would have done better to include bar graphs for comparison of trends between the three sets of data. Lines and bars make it easier for a person to pick out the implications of the data at a glance. Additionally, the tables were long because of the large number of variables that were being investigated in the study. As a result, the reader has to scroll through the document and may easily lose track of the information presented at the top of the table. Nonetheless, the tables presented all the information that was gathered from the study in a way that could easily be understood and as such, they were appropriate. The tables were also appropriate because the variables included subsections such as Caucasian, African American, Asian and Other within sections. Such information can only be best presented and displayed using a table.
Conclusion
In conclusion, it is important to note that the researchers in the study were investigating two phenomena using the same sets of data. The t-test was very appropriate in this case because it generated comparative information that could be interpreted easily. The nature of the data and variables that were being investigated called for a ratio level of measurement that complements the t-test in comparing means. The research was above average because the statistic, level of measurement, and displays used were all appropriate.
References
Keller, G. (2012). Statistics for management and economics. New York: CENGAGE Learning.
Meyer, S. M., & Werner, J. J. (2010). Characteristics of nurse practitioners interested in participating in a practice-based research network. Journal of American Academy of Nurse Practitioners, 22(1), 156-161.
Rubin, A. (2010). Statistics for evidence-based practice and evaluation. Belmont, Calif.: Brooks/Cole.
Sherri, L. J. (2012). Research methods and statistics: A critical thinking approach. Belmont, CA: Wadsworth Cengage Learning.