Online Learning

# Descriptive statistics

Differences in the mean pretest and posttest scores were computed to find out the extent in the change of confidence level brought about by the CRRP course. A higher mean difference value would indicate a higher degree of change in confidence level brought about by the CRRP course. Range and standard deviation measured the variability of the computed values in the study (Agresti amp. Finlay, 2009). A nurse leader may use descriptive statistics in cases when the average result is helpful in determining a course of action. In such cases, descriptive statistics are persuasive enough because it is able to give an overall picture of the data set in discussion. However, descriptive statistics, as the name implies simply provides a description of the data set and does not allow the nurse leader, to make inferences regarding the data (Malone, 2001). Based on my personal experience, we use descriptive statistics (particularly mean values) to find the prevalent cases in the nursing unit. Our department also routinely conducts a nurses’ evaluation assessment and our mean performance scores are usually given to us. Usefulness of Confidence Intervals in Determining Clinical Significance Confidence intervals indicate how variable the study data are, that is, the average distance of the data set values from the mean (Lee amp. Zelen, 2000). It should be noted that the true condition of a given population would be almost impossible to determine. Thus, researchers rely on the condition of a sample to provide a picture of the population. Confidence intervals aid researchers, analysts and practitioners in making decisions with regards to the clinical relevance of the data at hand. For example, if a study indicates a confidence interval of 95%, then the reader is able to determine that the values or the assessment given in the study is true for the population 95% of the time. The shorter the confidence interval, the more accurate is the assessment (Maki, 2006). For example, suppose a trial was conducted on the effectiveness of a weight loss pill against a placebo. Results of the study indicate that at a 95% confidence level, the weight loss was given to be 9 lbs. This means that the weight loss range would be between 4 to 14 lbs. Another interpretation of this information would be that it is highly likely for the pill to reduce one’s weight by at least 4 lbs, but highly unlikely for it to reduce one’s weight by more than 12 lbs. In this case, although the 9 lb weight loss arrived at was essentially just an estimate, the confidence interval that was set for the trial was able to quantify the uncertainty that was associated with that estimate (Malone, 2001). Clinical Significance vs. Statistical Significance Statistical significance measures the likelihood that the differences in the results of a particular test is due to the intervention applied on the treatment group and not simply due to chance (Malone, 2001). The most common measures of statistical significance, or hypothesis testing, are confidence intervals and p-values. On the other hand, clinical significance measures the magnitude of the differences created by the intervention on the daily lives of the participants (Agresti amp. Finlay, 2009). One controversy surrounding the issue between clinical and statistical significance is that statistical significance does not provide a clear picture of how large is the

Descriptive statistics