What is the Data Saying?
The DNP must have a basic understanding of statistical measurements and how they apply within the parameters of data management and analytics. In this assignment, you will demonstrate understanding of basic statistical tests and how to perform the appropriate test for the project using SPSS or other statistical programs.
Use the following information to ensure successful completion of the assignment:
- Refer to “Setting Up My SPSS,” “SPSS Database,” and “Comparison Table of the Variable’s Level of Measurement,” located in the DNP 830 folder of the DC Network Practice Immersion workspace.
- Doctoral learners are required to use APA style for their writing assignments. The APA Style Guide is located in the Student Success Center.
- This assignment uses a rubric. Review the rubric prior to beginning the assignment to become familiar with the expectations for successful completion.
- You are required to submit this assignment to LopesWrite. Refer to the LopesWrite Technical Support articlesfor assistance.
- Set up your IBM SPSS account and run several statistical outputs based on the “SPSS Database” Use “Setting Up My SPSS” to set up your SPSS program on your computer or device. You may also use programs such as Laerd Statistics or Intellectus, if you subscribe to them.
- The patient outcome or dependent variables and the level of measurement must be displayed in a comparison table which you will provide as an Appendix to the paper. Refer to the “Comparison Table of the Variable’s Level of Measurement.”
- Submit a 1,000-1,250 word data analysis paper outlining the procedures used to analyze the parametric and non-parametric variables in the mock data, the statistics reported, and a conclusion of the results.
What is the Data Saying
Quantitative data results are the core framework of every research study. Consequently, the interpretation of this data is critical in fully understanding the results and their implications to the research question. Deductions related to the true meaning of data are unachievable by simply observing the raw data. Thus, statistical analysis of such raw data helps to reveal the true meaning of the data and what it is conveying. However, Liang et al. (2019) argue that checking if the data meets the parameter test’s requirements and choosing the appropriate experimental design is paramount to prevent inaccurate inferences or deductions from the collected data. By taking individual pieces of information, statistics perform analysis based on the desired outcomes, often displayed in graphic, chart, or tabular form. This report aims at reviewing various statistical tests used in analyzing both parametric and non-parametric variables of a provided mock data.
Parametric and Non-parametric Tests
Statistical testing is typically classified into either parametric or non-parametric tests. Hopkins et al. (2018) claim that, for the parametric tests, the fundamental assumption is that the sample population is normally distributed with similar parameters such as standard deviations and means in the general population. Liang et al. (2019) highlight the standard assumptions for the parametric tests as homogeneity, normality, and independence of the data variance. Parametric tests include one-way Anova, two-sample t-test, paired t-test, and one-sample t-test (Liang et al., 2019). In contrast, non-parametric tests do not rely on assumptions regarding the population distribution, parameters, or the shape of the general population. Non-parametric tests include Kruskal-Wallis, Mann-Whitney, Signed-ranked, one-sample Wilcoxon, and one-sample Sign.
Parametric tests are deemed unsuitable when the sample population deviates from the normally distributed data’s assumption, resulting in Type II error and inaccurate deductions. The researcher will be unable to detect the variances in the population although present (Hopkins et al., 2018). The same case applies to non-parametric tests. Thus, on the one hand, parametric tests are used when seeking power, the spread size for each group varies, or the sample size is sufficient (Liang et al., 2019. On the other hand, non-parametric tests are used on data with non-removable outliers, raked data, ordinal data, non-normal looking data, small sample sizes, and data better represented using median (Hopkins et al., 2018). Incorrect use of either test will result in a lack of power. Therefore, prior data assessment and identification of the data’s type, shape, and characteristics is crucial to ensure test validity and results obtained.
Paired Sample T-test
A paired sample t-test compares the two means from the same sample with two dependent groups to establish a significant difference as a parametric statistical test. According to Skaik (2015), paired sample t-tests are generally used in research studies with the same subject on different sides’ samples, experiment/control research designs, or post-test/pre-test research designs. The test is crucial in comparing the variations between a matched pair, two measurements, two conditions, or two points in time (Kim, 2015). As such, paired sample t-test cannot be used on a ranked outcome, non-normally distributed data, more than two units, or unpaired data.
A paired sample t-test was conducted on 30 patients to determine if an intervention differed in the mean weights (See Appendix A). Baseline weight is compared to intervention weight to determine if a significant difference exists for all involved in the study. This test will examine a change in weight for both the intervention and nonintervention groups. The mean baseline weight is calculated at 217.5 pounds with a standard deviation of 53.40, and the intervention weight is 178.3 pounds with a standard deviation of 44.88. The results include t=7.188 with df=29, t(df)=2.05 with a 95% confidence interval and p = 0.000. A p <.005 indicates a statistically significant difference.