Interpretation of confidence interval [Article in Bosnian]

Abstract

In hypothesis testing, study results expressed through p-values provide information whether to accept or reject a null hypothesis. However, expressed this way they do not provide suffi cient information for a reader to assess the true difference in a measured effect. On the other hand, confi dence interval is informative enough to assess both the difference in a measured effect and its strength and reliability. Confi dence interval provides a reader with a span of values that contains the true value with a 95-percent probability and also tells the reader whether the sample is big enough for reliable results.