Quality Of Measures

981 Words4 Pages
This paper will highlight the information epidemiologists utilize to evaluate the quality of measures used in research as well as provide reasons why the quality of measures are essential and finally explain how epidemiologists ensure the measures used in research are of good quality.
a. What type of information do epidemiologists use to judge/evaluate the quality of their measures?
Tests for validity and reliability are the criteria mostly used by epidemiologists to judge the quality of measures used in research. Reliability of the measure shows that when different observers use the measure at different occasions, they will obtain similar test results (1). For example, a measure that determines weight gain among male adults in Saskatoon
…show more content…
The standards of acceptable reliability which epidemiologists seeks to attain include internal consistency exceeding 0.8 and stability measures greater than 0.5 (3), nevertheless, even if the internal consistency of a measure is low, it is better the measure has high stability and content validity.
Validity tests involve ensuring that the instrument is measuring what the epidemiologist intended (4), validity has several components including content and construct validity. Validity of the measure at minimum can be assessed subjectively by expert opinion of one or more experts in the field of study, where experts review the measure to determine if the instrument appears to be assessing the desired qualities (Face Validity) or determining whether the instrument samples all the relevant content (Content Validity) (3).
Additionally, constructs are “mini theories” used to assist in developing a new or better instrument in order to explain a broader range of findings or make more accurate predictions about a person’s behavior. Hence, when no other measures exist, the epidemiologist can link an attribute measured to another attribute by a hypothesis or construct and test the instrument by applying it to a sample to determine if the expected relationship is obtained; this is regarded as construct validity. For example; if a test for blood sugar was devised because there is no other test
Get Access