Measurement Error

Perhaps the most significant barrier to discovery in science is measurement error. A single measure of any construct – it really does not matter what you are measuring – is tainted with measurement error.

The idea of measurement error is relatively easy to understand if you are trying to measure psychological states like greed or guilt or depression. Everyone can agree that the measurement of such vague constructs is difficult and bound to miss the mark by a wide margin.

Measurement of even physical constructs has measurement error. If you and I measure the length of a football field, I may come up with 102 yards. You may come up with 99 yards.

Once you acknowledge the huge role measurement error plays in all of the sciences, you will honor the importance of using more than one variable to measure whatever construct you have decided to measure.  When you us more than one measure of your construct – you can actually correct for the attenuating effects of measurement error. The role of error which is uncorrected masks the true relationship.

I suggest that whenever possible, design you dissertation so that each construct has more than one measure. And incidentally, if you are interested in publishing the results of your dissertation, using multiple measures of the same construct is expected by most reviewers.

Robert Rodgers, Ph.D.

Leave a Reply

Your email address will not be published.