Have you really not met someone who could do maths until it came out of their ears, history until they could recite the exact details of the 100 year war including perfect quotes, learn science in half the time it took you... and yet fail completely to learn how to make tea or pack a suitcase without some basic error?
If I give strength higher priority then I'm looking to build a fighter or someone who uses strength. This runs true for every characteristic and skill in the game. I've played with high and low in all of them except CON, for obvious reasons in our games.
So there is no weighting involved other than my preference and the design I'm hoping to achieve. There is no ultimate stat. INT is not king and neither is any other stat. No amount of dexterity will make you better at walkin around in full plate, bar some limited circumstances, and no amount of strength will help you hit someone with a spell, unless it happens to be a spell placed on a heavy rock or something.
So to bring the analogy back to the test itself, if you are applying weighting to the various categories as the person who's set up the test then the test is based on biases. It's then essential to ensure that those biases are universal (objective) and not specific (subjective).