You don't explain how you do your test prints. If they are, say: increments of 2.5 seconds each, in a strip one after the other, adding up to say: 15 seconds in total, then the time the enlarger lamp takes to get to full illumination for each mini exposure will usually mean your test print will be slightly lighter than the single exposure of your full sized print.
My take on test prints is to do increments in stops or partial stops, using my Jobo Variomat easel. Initially I would use full stops for the four exposures on a single sheet of 20cm x 25cm paper. This would invariably be 2.5 sec, 5 sec, 10 sec and 20 sec. I would pick the closest density to what I want and run a test in half or one quarter f/stops around that chosen test print. This way the enlarger lamp is on for the same time a full print will be on and any difference between a whole print as opposed to a test print, will see minimal to almost no change in density.
As an additional difference, as a colour print is made darker (more density) it will become more red. In colour printing you may have noted you only change filtration for two of the colours, Cyan is not used. Cyan/red is controlled (generally) by print density. The more exposure or density given, the more red, the less exposure the print will be more cyan.
If your filtration is completely on the money, then you will certainly see this cyan/red change easily. If your filtration is close, then you will still see the difference, but other colours may creep in.
As mentioned, greater accuracy will be obtained if your test paper and your final print are from the same batch, better still, from the same box or roll.
Mick.