Analyzing Parametric Statistics
As a practice scholar, you are searching for evidence to translate into practice. In your review of evidence, you locate a quasi-experimental research study as possible evidence to support a practice change. You notice that the study aims to make a prediction that relates to correlation between study variables. The study sample size is large and normally distributed. Reflect upon this scenario to address the following.
- In your appraisal of the evidence, you note that an independent variable is not present and that a Spearman’s ranked correlation is used to analyze data. Is this the correct level of correlational analysis? Explain your rationale.
- Are association and correlational analysis equivalent in determining relationships between variables?
- Do these findings impact your decision about whether to use this evidence to inform practice change? Why or why not?
Differentiate selected statistical methods for the purpose of translation science
Evaluate selected statistical methods for the purposes of critical appraisal of evidence
- Synthesize literature relevant to practice problems.
Reflection on Learning
Reflect upon your Week 5 learning journey in NR714 and consider the following in one or two paragraphs.
- Provide one specific example of how you achieved the weekly objectives.
- In what ways did course information influence your understanding of parametric statistics?
- What goals will you set in accordance with what you have learned this week?
Hello there! Welcome to Week 5! This week on your journey, we’ll explore how translating best available evidence to a practice problem requires knowledge specific to the critical appraisal of research studies. Of course, critical appraisal involves data analysis, so this week we’ll spend more time learning and applying statistical tests, specifically, parametric statistical tests. Parametric statistics can be used to examine relationships, make predictions, and examine causality. While you’ll likely use technology to calculate statistics, you do need to know which test aligns with the data and is best suited to answer the practice problem. How does the practice scholar determine the appropriate statistical method? Let’s find out!
Outcomes, Objectives, and Concepts
| Weekly Outcomes | Weekly Objectives | Main Topics and Concepts |
| Differentiate selected statistical methods for the purposes translation science (PO 3, 5, 9).Evaluate selected statistical methods for the purposes of critical appraisal of evidence (PO 3, 5, 9).Synthesize literature relevant to practice problems. (PO 3, 5, 9) | Select the appropriate statistical test to answer the research question.Compare and contrast the use of parametric and nonparametric statistical tests.Critique the statistical tests used to answer the research question.Appraise the quality of data generated from selected statistical tests to guide practice.Select appropriate literature to answer research question. | Identify Parametric Statistical Procedures Used in a Published Research StudyReview Statistical Procedures for Appropriateness in Light of the Hypotheses, Research Question, and for the Level of Measurement of the VariablesComprehend the Discussion of Data Analysis Results in the StudyDetermine Whether the Researcher Interpretation of the Results Is AppropriateEvaluate the Clinical Significance of the Findings |
Foundations for Learning
Start your learning this week by reviewing the following.
Invalid Conclusions From Studies Example: https://www.khanacademy.org/math/ap-statistics/gathering-data-ap/modal/v/invalid-conclusions-studies-example (Links to an external site.)
All Khan Academy content is available for free at www.khanacademy.org (Links to an external site.)
Student Learning Activities
| Learning Activities | This week, you will complete the following. PrepareAssigned ReadingsExplore Interactive LessonTranslate to PracticeQuizDiscussion Question ReflectReflection on Learning |
| Additional Resources | Statistics How To: https://www.statisticshowto.datasciencecentral.com/parametric-statistics/ (Links to an external site.) Statistical Power: https://courses.lumenlearning.com/boundless-statistics/chapter/statistical-power/ (Links to an external site.) The t-Test: https://courses.lumenlearning.com/boundless-statistics/chapter/the-t-test/ (Links to an external site.) Statistics Terminology Resource Document: Statistics Terms (Links to an external site.) Key Points Questions (Links to an external site.) |
Learning Success Strategies
- Review key terms in additional resources above to ensure you understand the definitions and concepts related to the review and critical appraisal of statistical analysis in a published research study.
- As you review weekly content, consider how research study data management and statistical analysis directly influence the conclusions made by researchers.
- Be ready to share your thoughts through the interactive discussion. Review the discussion guidelines and rubric to optimize your performance.
- You have access to a variety of resources to support your success. Click Resources on the home page to access program and project resources.
- Your course faculty is here to support your learning journey. Reach out for guidance with study strategies, time management, and course-related questions.
Polit, D. F., & Beck, C. T. (Eds.). (2017). Nursing research: Generating and assessing evidence for nursing practice (10th ed.). Wolters Kluwer.
- Chapter 16: Descriptive Statistics
- Chapter 17: Inferential Statistics
- Chapter 18: Multivariate Statistics
- Chapter 19: Processes of Quantitative Data Analysis
- Chapter 20: Clinical Significance and Interpretation of Quantitative Result
Informing Practice Change Through Parametric Statistics
Using Statistics to Examine Relationships
Researchers use correlational analyses to identify relationships between or among variables. The purpose of this analysis may be to describe relationships between variables, clarify the relationships among theoretical concepts, or assist in identifying possible causal relationships, which can then be tested by causal analysis. Data measured at the interval level from a single population provide the best information on the nature of the relationship. Keep in mind that there are analysis procedures available for most levels of measurement. A large sample size with a wide range of data is desirable.
Two statistics are commonly used to examine relationships: Pearson Product-Moment Correlation and Factor Analysis. Pearson Product-Moment Correlation is a parametric test used to determine relationships among variables. By far the most common correlation is the Pearson correlation that measures the degree and direction of a linear relationship between two variables. The Pearson correlation is identified by the letter r. When you encounter correlations, there are four considerations you should bear in mind.
Previous
4. When judging ‘how good’ a relationship is, it is tempting to focus on the numerical value of the correlation. For example, a correlation of +0.5 is halfway between 0 and 1.00 and, therefore, appears to represent a moderate degree of relationship. However, a correlation should not be interpreted as a proportion. Although a correlation of 1.0 does mean there is a 100% perfectly predictable relationship between X and Y, a correlation of 0.5 does not mean you can make predictions with 50% accuracy. To describe how accurately one variable predicts the other, you must square the correlation. Thus, a correlate of r = 0.5 provides 25% accuracy.
1. Correlation simply describes a relationship between two variables. It does not explain why the two variables are related. Specifically, a correlation should not and cannot be interpreted as proof of a cause-and-effect relationship between two variables.
2. The value of a correlation can be affected greatly by the range of scores represented in the data.
3. One or two extreme data points, often called outriders, can have a dramatic effect on the value of a correlation.
4. When judging ‘how good’ a relationship is, it is tempting to focus on the numerical value of the correlation. For example, a correlation of +0.5 is halfway between 0 and 1.00 and, therefore, appears to represent a moderate degree of relationship. However, a correlation should not be interpreted as a proportion. Although a correlation of 1.0 does mean there is a 100% perfectly predictable relationship between X and Y, a correlation of 0.5 does not mean you can make predictions with 50% accuracy. To describe how accurately one variable predicts the other, you must square the correlation. Thus, a correlate of r = 0.5 provides 25% accuracy.
1. Correlation simply describes a relationship between two variables. It does not explain why the two variables are related. Specifically, a correlation should not and cannot be interpreted as proof of a cause-and-effect relationship between two variables.
Next
- 1
- 2
- 3
- 4
Researchers will report their correlation results directly into the text of the article or in a table format.
Factor Analysis
A factor analysis examines interrelationships among large numbers of variables and disentangles those relationships to identify clusters of variables that are most closely linked. Closely related variables are grouped together into a factor. Several factors may be identified within a data set. The researcher interprets the statistical findings in an attempt to explain why the analysis grouped the variables in a certain way.
Factor analysis aids in the identification of theoretical constructs. It is also used to confirm the accuracy of a theoretically developed construct. For example, a theorist may state that the concept (or construct) of ‘hope’ consists of the following elements: anticipation of the future, belief that things will work out for the best, and optimism. A factor analysis can be conducted on study findings to determine whether subject responses clustered into these three groupings.
A factor analysis is frequently used by researchers in the process of developing measurement instruments, particularly those related to psychological variables such as attitudes, beliefs, values, and opinions.
Using Statistics to Predict
A foundational intention of research is to make predictions. It is no surprise that researchers look for opportunities to employ a statistical analysis that lends itself to making a prediction. Let’s look at a few of these commonly used prediction-focused statistical tests.
View the following activity to investigate the use of statistics to predict.
tatistics are used to predict value in research using regression analysis. Regression analysis is used to predict the value of one variable when the value of one or more variables is known. The variable to be predicted in a regression analysis is referred to as the dependent variable. The dependent variable is usually measured at the interval level.
The researcher’s goal when selecting this analysis is to explain as much of the variance in the dependent variable as possible. In the regression analysis, variables used to predict values of the dependent variable are referred to as independent variables. If there’s more than one independent variable, the analysis is referred to as multiple regression.
In regression analysis, the symbol for the dependent variable is y and the symbol for the independent variable is x. Scatter plots and a bivariate correlation matrix often are developed before regression analysis is performed to examine the relationships that exist in the variables. The purpose of regression analysis is to develop a line that will best reflect the values on the scatter plot.
Many types of regression analysis have been developed to analyze various types of data. One type, logistic regression, was developed to predict values of a dependent variable measures at the ordinal level. Logistic regression is being used with increasing frequency by nurse researchers. The outcome of a regression analysis is the regression coefficient r.
When r is squared, it indicates the amount of variance in the data that is explained by the equation. A small sample size decreases the possibility of obtaining statistical significance.
ustifying Statistical Suitability and Interpreting Statistical Outcomes
Judging Statistical Suitability
As a practice scholar, you will read diverse research studies with varied research designs, sampling strategies, and statistical analyses. Related to your role in determining statistical suitability for each research study you read, review, and critically appraise, multiple factors are involved in determining the suitability of a statistical procedure for a particular study. These include the research study’s purpose; hypotheses, questions, or objectives; design; and level of measurement. Determining the suitability of various statistical procedures for a particular research study is not straightforward. Regrettably, there is not usually one right statistical procedure for a particular study.
Evaluating statistical procedures requires that you make a number of judgments about the nature of the data and what the researcher wanted to know. You need to determine whether the data for analysis were treated as nominal, ordinal, or interval; how many groups were in the study; and whether the groups were dependent or independent.
In independent groups, the selection of one subject is totally unrelated to the selection of other subjects. For example, if subjects are randomly assigned to treatment and control groups, the groups are independent. In dependent groups, subjects or observations selected for data collection are related in some way to the selection of other subjects or observations. For example, if subjects serve as their own control by using the pretest as a control, the observations are dependent. Also, if matched pairs of subjects are used for control and treatment groups, the observations are dependent. For example, in a study of twins, one twin may be placed in the control group and the other in the treatment group. Because they are twins, they are matched on several variables.
One approach to judging the appropriateness of an analysis technique for a critique is to use a flowchart, which directs you by gradually narrowing the number of appropriate statistical procedures as you make judgments about the nature of the study and the data. There are numerous flowcharts or algorithms available.
Click image to enlarge
Click to enlarge image
An algorithm is an essential guide as you critique a researcher’s data management and analysis.
Three assumptions underpin decisions researchers make as they select statistical tests. An algorithm takes these 3 assumptions into consideration. As you analyze quantitative research study data, the algorithm poses questions that allow you to determine if the researcher selected an appropriate parametric or non-parametric statistical test.
Let’s take a look at a few questions.
What type of data did the research team collect?
- Discrete or categorical data = Chi square tests for one and two samples are appropriate. Discrete or categorical data do not lend themselves to additional statistical testing beyond the Chi square test.
- If the researcher collected continuous data, closely read their research question.
What type of question is posed by the researchers?
- Does the research study question state an intention to predict relationships or predict differences?
- Notice the intention of the stated research question, results in a series of questions that determine if a nonparametric statistical test is appropriate as compared to its parametric counterpart.
Each question posed in an algorithm keeps your critique of data management and statistical analysis clearly focused on both the type of data collected and whether the data collected by the researchers was appropriate to answer the research question posed in the study. It is a misstep for a researcher to use a parametric statistical test when the data collected and research question do not support crossing over into parametric statistical analysis.
As you critique each quantitative research study, an algorithm guides your review and appraisal. Integrate the findings from your appraisal of data management and statistical analysis into your overall critique of the quantitative research study.
This flowchart identifies factors related to the appropriateness of a statistical procedure: the research question, level of measurement, study design, and type of sample.
Interpreting Statistical Outcomes
To be useful, the evidence from data analysis must be carefully examined, organized, and given meaning. Evaluating the entire research process, organizing the meaning of the study results, and forecasting the usefulness of the findings, all of which are involved in interpretation, require high-level intellectual processes. In this segment of the conduct of a research study, the researcher translates the results of analysis into findings and then interprets them by attaching meaning to the findings.
The researcher enters a process of interpretation that includes examining evidence, forming conclusions, considering implications, exploring the significance of the findings, generalizing the findings, and suggesting further studies. This information is usually included in the final section of published studies, which often is entitled “Discussion.”
Types of Results
Interpretation of results from quasi-experimental and experimental studies traditionally is based on decision theory, leaving the researcher with five possible results:
- significant results that agree with those predicated by the researcher
- nonsignificant results
- significant results that are opposite those predicted by the researcher
- mixed results, and
- unexpected results.
Findings
Results in a research study are translated and interpreted, then they become findings, which, are a consequence of evaluating evidence. Although much of the process of developing findings from results occurs in the mind of the researcher, evidence of the author’s thought processes can be found in published research reports.
Conclusions
Conclusions are a synthesis of the findings. In forming conclusions, the researcher uses logical reasoning, creates a meaningful whole from pieces of information obtained through data analysis and findings from previous studies, remains receptive to subtle clues in the data, and considers alternative explanations of the data. One of the risks in developing conclusions is going beyond the data, or forming conclusions that are not warranted by the data. This occurs more frequently in published studies than one would like to believe.
Considering Implications
Implications are the meanings of conclusions from scientific research for the body of nursing knowledge, theory, and practice. Implications are based on but are more specific than conclusions, and they provide specific suggestions for translating the findings to practice.
Exploring the Significance of the Findings
The significance of a research study is associated with its importance to the nursing body of knowledge. Significance is not a dichotomous characteristic because studies contribute in varying degrees to the body of knowledge. Significance may be associated with the amount of variance explained, the degree of control in the study design to eliminate unexplained variance, or the ability to detect statistically significant differences. To the extent possible at the time the study is reported, the researcher is expected to clarify the significance.
A few studies, referred to as landmark studies become important reference points in the discipline of nursing. The true importance of a research study may not become apparent for years after publication. Certain characteristics are associated with the significance of studies. Significant studies make an important difference in people. A very significant study has implications for one or more disciplines in addition to nursing. The study is accepted by others in the discipline and frequently is referenced in the literature. Over a period of time, the significance of research study is measured by the number of other studies it generates.
Generalizing the Findings
Generalization extends the implications of the study findings from the sample studied to a larger population. For example, if the study was conducted on diabetic patients, it may be possible to translate the findings to persons with other practice problems or to healthy people.
References
References
Dang, D., & Dearholt, S. (2018). Johns Hopkins nursing evidence-based practice model and guidelines (3rd ed.). Sigma Theta Tau International.
Polit, D. F., & Beck, C. T. (Eds.). (2017). Nursing research: Generating and assessing evidence for nursing practice (10th ed.). Wolters Kluwer.
White, K. M., Dudley-Brown, S., & Terhaar, M. F. (2016). Translation of evidence into nursing and health care (2nd ed.). Springer Publishing Company.
