Reliable and valid
Reliability and validity are the basic properties which any research should have or at least should be aiming at achieving as closely as possible (approximate results are not always a bad thing). Reliability and validity of research results come from reliable and valid procedures (Johnstone, 2000) which I would like to clearly define in this post.
A reliable procedure is the one which yields the same results every time it is employed (e.g. each time I use the formula E = mc2, it always works well to calculate energy).
A valid procedure is the one which yields the results that are correct and true (e.g. if I calculate energy and my calculi are correct, then the results of my application of the formula E = mc2 are going to be valid).
Johnstone, B. (2000). Qualitative methods in sociolinguistics. United States of America: Oxford Univ. Press.