The hypothesis testing alone is not sufficient to offer a complete picture of what is being tested. Instead, confidence intervals (CI) complement the hypothesis-testing process by providing a range of values within which the true population parameter falls. The range of values provided by the confidence intervals includes the accurate value of statistical constraint within the population targeted. Researchers often use 95% CI. An investigator can set any level between 90%CL to 99% CL. The CL of 95% shows that suppose the study is undertaken 100%, the range would have an actual value of 95. Hence, CL offers more evidence concerning the precision of an estimate in comparison to the p-value (Shreffler & Huecker, 2023). These are critical facts to know when using the CL and hypothesis testing:
  1. A researcher has a bad CI if the null hypothesized value is present in the CI because it will lead to a high p-value.
  2. The null hypothesized value will be at a point of no difference or zero value if the CI is zero, predicting a chance of finding no difference.
  3. If the null hypothesized value falls within the CI, the p-value will be greater than 5%.
Suppose the null hypothesized value falls outside the presented C; the p-value will be less than 5%. Therefore, both CI and hypothesis intervals can be used to support conclusions (Shreffler & Huecker, 2023). Examples of how the hypothesis and confidence intervals are utilized together in healthcare research include: Comparative studies where health researchers are comparing two groups like treatment. The hypothesis testing assists in determining whether there are significant differences between the two groups, while CI estimates the magnitude of the difference. For example, hypothesis tests may confirm the superiority of one technique over the other in a clinical trial comparing two surgical techniques. At the same time, confidence intervals will specify the range of anticipated improvement. Another example is predictive modeling, where the hypothesis testing validates the significance of the predictor variable while CI quantifies their predictive accuracy. For example, in a model predicting heart attack, the hypothesis confirms the significance of variable factors like cholesterol levels, while CI shows the precision of the predictors (Hespanhol et al., 2019). An example in the workplace is when testing the effectiveness of a drug on a patient’s cholesterol level. The hypothesis is supported when the p-value is less than the predetermined significant level of 0.05. In this example, the confidence interval could indicate that the new drug reduces collateral levels by 5 to mm Hg. Such an interval quantifies the uncertainty associated with the presented estimate. Therefore, both the hypothesis and confidence intervals are inferential techniques. They utilize a sample to test the strength and validity of the hypothesis of the population parameter. The only difference is the point of reference, whereby the hypothesis tests focus on the null hypothesized parameter while CI focuses on the estimate of the sample parameters.

References

Hespanhol, L., Vallio, C. S., Costa, L. M., & Saragiotto, B. T. (2019). Understanding and interpreting confidence and credible intervals around effect estimates. Brazilian journal of physical therapy, 23(4), 290-301. https://www.sciencedirect.com/science/article/pii/S141335551831058X Shreffler, J., & Huecker, M. R. (2023). Hypothesis testing, P values, confidence intervals, and significance. In StatPearls [Internet]. StatPearls Publishing. https://www.ncbi.nlm.nih.gov/books/NBK557421/