Introduction
Epidemiologists carry out research studies and compile clinical results for various health issues affecting people in different communities. During analysis of these research findings, confounding and random errors occur. This makes epidemiologists cautious when developing research studies to minimize their occurrences (Loewenson, 2004).
The role of P-values and Confidence Intervals
In epidemiological research study, new test results may differ from the old ones by a large margin. P-values help the researchers to assess whether the effect of new test results is real or it is a result of chance. These values provide enough information about the difference between the actual epidemiological research findings and the experimental ones. Furthermore, they quantify the degree to which collected data differ from the expected change (Gardner & Altman, 2006).
When p-values are small, high chances are that the research findings did not happen by chance. This allows the researchers to reject the null hypothesis. On the other hand, when p-values are large, it implies that the research findings happened by chance. This confirms that there are no differences between the old and new findings.
Confidence intervals are important when comparing the difference between real and experimental values. They give a range of values in which the true value reflects. They provide a guideline for the interpretation of the research results by looking at the effects of chance on results. This helps in determining whether the research findings are statistically significant or non-significant. The upper and lower extremes of confidence intervals depict how large or small the actual effect is. This helps in the interpretation of both significant and non-significant research findings (Aschengrau & Seage III, 2012).
The effect Sample Size has on P-values and Confidence Intervals
The sample size has significant effects on p-values and confidence intervals. When the sample size is large, small standard errors are realized, which result into more significant p-values. On the other hand, a small sample size makes statistical comparisons to show that statistical significance does not exist between the test groups. This occurs despite the informal inspection revealing that there are differences.
The sample size affects the width of the confidence intervals. Smaller sample sizes give low precise estimate effects than larger samples do. This means that confidence intervals are wider for smaller sample sizes, but narrower for large study samples (Altman et.al, 2002).
Validation of an Idea with Own Experience
Individuals in the medical field encounter situations where epidemiological methods and statistics play an important role in answering certain research questions. Questions about the effectiveness of new drugs on certain health issues or the effects of medical intervention on patient satisfaction can be addressed by using p-values and confidence intervals. Data collected for study undergo several tests to determine its significance to the study (ERIC Notebook, 2001). For instance, in statistical analysis, hypothesis is used to test hypothesized epidemiological conditions. During these tests, the main interest is put on testing the null hypothesis. First, the null hypothesis is assumed to be correct, and efforts are put to determine the probability of obtaining the values that were achieved under null conditions. When the values are significantly small, the null hypothesis is rejected, and the alternative hypothesis is adopted as true. Thus, two types of errors emerge, which include type1 and type 2. Type 1 error represents the probability of rejecting null hypothesis. On the other hand, type 2 error represents the probability of not rejecting the null hypothesis when the alternative hypothesis is true (ERIC Notebook, 2000).
Conclusion
In epidemiological study, confounders that bias the study results exist. To control the amount of bias, statistical analysis, or randomization methods are used. Random errors that occur during the analysis of research findings can be reduced by handling the results carefully (ERIC Notebook, 2001).
References
Altman, D.G., Machin, D., Bryant, T., & Gardner, M. (2002). Confidence Intervals in Practice. New York: BMJ Books.
Aschengrau, A., & SeageIII, G. R. (2012). Essentials of Epidemiology in Public Health. New York: Jones and Bartlett Learning.
Epidemiologic Research and Information Center (ERIC) Notebook. (2000).Confounding Bias, Part I. Durham: North Carolina. Web.
Epidemiologic Research and Information Center (ERIC) Notebook. (2000). Confounding Bias, Part II. Durham: North Carolina. Web.
Epidemiologic Research and Information Center (ERIC) Notebook. (2001). Common Statistical Tests and Applications in Epidemiological Literature. Durham: North Carolina. Web.
Gardner, M., & Altman, D. (2006). Confidence Intervals rather than P-values: Estimation rather than Hypothesis Testing. Washington, D.C: Br Med J.
Loewenson, R. (2004). Epidemiology in the Era of Globalization: Skills Transfer or New Skills? Int. J. Epidemiol., 33(5), 1144-1150.