Let me give you a short answer as a place-holder for a possible edit into a longer answer.
1) "Validity" here means hiring of better rather than worse workers, as all workers can be rank-ordered as to work performance, as compared to doing nothing to select workers.
2) Usually in hiring, it's enough to outrank someone else in a hiring criterion, for example general mental ability, without necessarily reaching any particular score level. The general idea is to rank higher than the other applicants when applying for a job, and to choose higher-ranked rather than lower-ranked applicants when hiring other people.
I'll see if I have time to add specifics later after some activities with my family.
Sorry - I see now that my original questions were pretty vague.
I guess what I'm trying to do is turn your interesting information into an actionable strategy.
Let's try a scenario: Let's say I have a job opening for a senior developer. My goal is to recruit the best possible candidate. I put out the job description, and I get 20 responses. I give them all the same structured interview (.53 validity). I also ask for a work sample (.54 validity). So, in terms of these validity scores, what do the numbers represent? Does a .53 validity say that the applicant who scores better will have a 53% chance of performing better in practice?
Anyway, lets say I somehow qualitatively interpret their interviews and samples, and using some rubric, rank them. (But, how do I know that my qualitative review is even valid?) Then, I give the top candidate an offer.
How do I determine how much better the top candidate is than the 2nd top (to give an appropriate offer)?
Also, how do I know if need to continue interviewing another 20 candidates?
Also, how good of a GMA score is needed? That is, how good of a score would be needed for me to stop looking at other candidates?