Interpretation of Statistics in Clinical Laboratory Studies

Lisa W Kay; Mark A Gebert

doi:10.29074/ascls.19.1.18

This article requires a subscription to view the full text. If you have a subscription you may use the login form below to view the article. Access to this article can also be purchased.

Lisa W Kay, PhD⇑
1. is an Assistant Professor, Eastern Kentucky University, Richmond KY
Mark A Gebert, PhD
1. is an Assistant Professor, Eastern Kentucky University, Richmond KY

Address for correspondence: Lisa W Kay PhD, Department of Mathematics and Statistics, 313 Wallace Building, Eastern Kentucky University, 521 Lancaster Avenue, Richmond KY 40475. (859) 622-1621, (859) 622-3051 (fax). Lisa. Kay{at}eku.edu

Extract

Throughout the course of a research study that involves data, there are many opportunities for making errors. Poor survey construction, data collection methods that introduce bias, misuse of statistical analysis procedures, and misinterpretation of results can all lead to drawing incorrect conclusions. Researchers who are producing reports need to be able to collect and interpret data accurately, and clinical practitioners who are reading reports need to be able to assess the statistical content of such reports. While the technology of today makes it relatively easy to produce numerical summaries of data sets, the interpretation of the output still requires some careful thought. Correct interpretation of inferential procedures will be the focus of this article, and common misinterpretations will be described.

CONFIDENCE INTERVALS One of the most commonly used inferential tools is the confidence interval. A confidence interval is used to estimate a parameter of interest and is of the form: estimate ± margin of error.

An example of a fairly standard interpretation of a confidence interval might be the following: “We are 95% confident that the true mean age is between 25 and 35 years old.” Here 95% tells us how confident we are in the process of constructing the confidence interval. If we were to randomly select samples of a particular size over and over again and construct a confidence interval for each one in this same manner, in the long run approximately 95% of the intervals would contain the true value that we are trying to estimate.