CRITERIA OF MEASUREMENT QUALITY
    Precision and Accuracy
    Reliability
    Validity
    Who Decides Whatís Valid?
    Tension Between Reliability and Validity

    We've already looked at the choices regarding the precision of measurements: you can say someone lives in Los Angeles or in Southern California.  Do not confuse precision with accuracy, however.  If the person in question lives in Pasadena, then the less precise label is the more accurate.

    Reliability is a matter of consistency or repeatability in measurement.  If you could make the same measurement over and over, would you get the same answer each time?  It's easy to see this in a physical example like a bathroom scale.  If you step on the scale repeatedly, you should get the same weight each time.  If you get different weights, your scale has a problem of reliability.  It's harder to imagine comparable situations in social research.  If I asked you the same question over and over, you'd probably give me the same answer (before turning and running away from the crazy man).  We can imagine, however, that if people for very precise information that they will be required to estimate (e.g., "How many miles have you driven during your lifetime?") there's a good chance they would come up with a different estimate if asked again (and didn't remember what they said the first time).

    Validity is a matter of whether you are measuring what you say you are.  If you want to know how hard people work, asking them how much they make is not the best strategy.  It reflects many things in addition to hard work.

    As you will have imagined, it's best to make measures that are high in both reliability and validity.  Unfortunately, as you'll see, these two qualities are often at odds with each other.  Measures with undeniably high reliability are often weak in validity and vice versa.