You are correct that adding too much boost will result in clipping. There was some discussion here that it'd be best to center your EQ around 0dB because applying too much gain (either negative or positive) would result in some bands overlapping one another. For example, giving a -10dB cut in band 5 would result in bands 4 and 6 also getting (which wouldn't be the case if band 5 were only cut -5dB). This is assuming that the Q value determines the "shape" of the curve.

Because of this, the tutorial was written with keeping the gain or cut as close to 0dB as possible.

To best honest, I'm not sure I agree with this theory 100%. I always thought that as gain was increased (either by a positive or negative value) the shape of the curve became more "pointy". My assumption was that the Q value determined the high and low end of the range of frequencies that were to be modified (in absolute terms). My view was based on an interactive flash animation at one of the major car-audio sites. This animation let you adjust the gain, center frequency and Q value for a virtual parametric EQ. As gain as increased, the curve became more pointed. As Q was increased, the beginning and ending frequencies were moved "away" from the center frequency. For the life of me, I can't find this web page that shows this.

I guess you could change steps 10 and 11 to try to apply only cuts in the EQ. Because the empeg stores multiple presets, you can do EQs for both. The process of using the SPL meter doesn't change, so you can use your findings for both.

Maybe you can give it a shot and see which sounds better?