The Reliability and Validity of Measurements

Measurement error is easily identified using measures of reliability. In the abstract sense, reliability is described as a matter of whether a particular technique, used repeatedly towards a similar subject, would supply the same result for every measurement event. The measure of reliability is the measure of the random errors in a measurement. Consequently, the more reliable the measurement is, the less would be the random error in it (Rubin and Babbie, 2004, p. 182). Rubin and Babbie (2004) mentioned that reliability does not guarantee accuracy.For example, if two measurement apparatus are not synchronized or standardized to each other, they might still reflect the same measurement, but the other apparatus might have been adjusted to some extent in order to confirm with the result of the measurement of the other apparatus (). Validity on the other hand is defined as the extent to which an empirical measure satisfactorily represents the real meaning of the variable being measured. T here is actually no basis to determine the adequacy of the measure. Comparison and agreement to pre-existing terms or concepts is done to accept the validity or not (Rubin and Babbie, 2004, p.186).Reliability is not a qualification to conclude that the measurement done is valid. We may say that the measured values, especially in quantitative research, are precise, but not accurate. There are also cases that the resulting measurements are accurate to some level, but are imprecise. In the worst case would be that the measurements are imprecise and inaccurate. Because of this, a certain tension exists, whether to choose reliability over validity, or the other way around (Rubin and Babbie, 2004, p. 198).Of course, the target would be to collect both valid and reliable data, and this can be obtained by being strict to the implementation of the procedures for measurement: being thoughtful and less sloppy. The difference between the two is that reliability requires fewer variables, and the refore less validity, vice versa. This kind of relationship is still in conflict (Martin, 2005). We could say therefore, that a measurement is valid even if it is not a reliable measurement, since the way to prove the reliability is the repeatability and reproducibility of the results.Also, the requirement to strengthen the validity is to consider many relevant variables and to get near to the exact value, while on reliability, there is no requirement for such (Martin, 2005). References Martin, Rogel L. (2005). Reliability vs. Validity. Retrieved August 29, 2008, from http://www. businessweek. com/innovate/content/sep2005/id20050929_872877. htm Rubin, Allen and Earl R. Babbie. (2004). Research Methods for Social Work. Retrieved August 29, 2008, from http://books. google. com. ph/books? id=eAdbEn-yZbcC