Descriptive statistics build the foundation for quantitative studies because they collect various features of gathered data. For example, average aspects of a population or a percentage of women in a sample are types of descriptive statistics (Polit & Beck, 2017). There are four ways, in which information can be collected and measured. The first one is nominal – objects or situations are assigned numbers (or letters) that have no actual value. For instance, males and females in a study can be represented by numerals 1 and 2. The next level is ordinal where the order of objects has value, but the connection between them is unclear. Here, satisfaction or agreement ranges can be used as examples. The third type is interval – both the order of and correlations between values are known, thus providing a measurable scale. Finally, the last level of measurement is a ratio, which adds an absolute zero to the calculations.
The accumulated information can be analyzed using univariate and bivariate statistical methods. The first type, univariate statistics, interprets only one variable collected for one’s research. For example, scholars can gather data about participants’ age in order to find the average age of a particular group or its youngest members. Univariate analyses include a central tendency which shows the mean, the mode, and the median. The mode determines the most common measures; the mean presents an average score from all values; the median shows data that is placed in the middle of all variables. For instance, Jennings, Clifford, Fox, O’Connell, and Gardner (2015) use the median to determine ages of practitioners in different studies. Another univariate analysis is dispersion – Beauvais, Stewart, DeNisco, and Beauvais (2014) use it to find the variance and standard deviation of their sample.
Bivariate descriptive statistics, on the other hand, examine two variables. They are usually utilized to find a connection or an association between multiple sets of data. Analyses such as correlation indexes and contingency tables are examples of bivariate statistics. Beauvais et al. (2014) use a specific correlation index, Pearson’s correlation, to show the connection between nurses’ success and other factors such as their preferred classes and ways of learning. It measures the association between two values, creating a line from one variable to another. González, Manzanares, and Peinado (2017) utilize a contingency table, a more complicated tool, to investigate a relationship between students’ satisfaction and different learning exercises by presenting the values in a matrix. The resulting table presents many sets of data arranged according to their proximity to and interconnectedness with each other.
These methods do not measure significance and require additional statistical tests for that. Their advantage lies in the use of specific data, while their drawback is the lack of connection between some variables. However, the correctness of results can be strengthened by calculating risk indexes – estimated proportions to reduce errors in interpretations of data. Absolute risk is the percentage of participants who experienced a negative outcome during the study’s experiment (Polit & Beck, 2017). For example, it is the rate of people who did not stop smoking after an intervention. It can help one see a difference between the results of focus and control groups, also called absolute risk reduction. Here, the difference between continuing smokers in two groups is measured.
Relative risk is a proportion of people with adverse outcomes that were in the intervention group (Polit & Beck, 2017). Relative risk reduction is used to show the impact of the intervention on the baseline risk – the effectiveness of a smoking cessation campaign on the group. The odd ratio shows the relation between people with and without adverse outcomes (people who continued and stopped smoking). Finally, the number needed to treat represents all people who need to be exposed to the intervention to have a positive effect on one person. In the example, three people receiving anti-smoking treatment are required to stop one person from smoking (Polit & Beck, 2017).
References
Beauvais, A. M., Stewart, J. G., DeNisco, S., & Beauvais, J. E. (2014). Factors related to academic success among nursing students: A descriptive correlational research study. Nurse Education Today, 34(6), 918-923.
González, V. S., Manzanares, M. T. L., & Peinado, J. A. A. (2017). Nursing students’ satisfaction during their first year of study in a private university as regards the integration of ICTs. Educational Excellence, 3(2), 35-74.
Jennings, N., Clifford, S., Fox, A. R., O’Connell, J., & Gardner, G. (2015). The impact of nurse practitioner services on cost, quality of care, satisfaction and waiting times in the emergency department: A systematic review. International Journal of Nursing Studies, 52(1), 421-435.
Polit, D. F., & Beck, C. T. (2017). Nursing research: Generating and assessing evidence for nursing practice (10th ed.). Philadelphia, PA: Wolters Kluwer.