I’m conducting a study of television audience behavior.  I have descriptions of program titles and a list of program categories.  I have two graduate assistants helping me. I ask them to each code the same sample of 20 programs.  The two graduate assistants code 85 percent of the programs the same–meaning they put them in the same categories. From this I can say the measure is:

Reliable

Valid

Neither reliable nor valid

Because I didn’t give the coders enough to work with, I can’t tell whether the measure is reliable or valid.

Statistically significant