Unfortunately the nutritional values given on packages aren't accurate enough for it to make a difference whether you eat chicken to the bone. Compliance testing in the United States and Canada only requires that an average of a certain number of servings be within 20% of the stated number. Individual servings can vary even more and still be in compliance, so long as the average of the random samples tested remains within 20%.
With something like a frozen fried chicken dinner there probably isn't going to be a wide variance in the actual numbers, but some variance is unavoidable. Chickens aren't all the same, some will be bigger, smaller, fatter or leaner. There's only so much any food processor can do to ensure that every chicken part that gets used in an product has the same weight and nutritional composition. There would also be some variation in how much batter gets coated on each part, and how much oil ends being absorbed when cooking.
On the other hand, the scraps you're leaving on the bone are probably only have tiny fraction of the total calories of the fried chicken part(s). I can't see it being more than 5% and it could easily be less than 1% of the total meal, depending on what else is included in the frozen dinner.
I should also say I think you're doing yourself a disservice by trying to fudge the numbers like this, for two reasons. The first is that when I eat fried chicken I love going over the bones trying to get the last of the meat and that tasty, tasty batter stuck to the ends. It's almost like getting to eat another piece of chicken, but one with practically no additional calories. The second reason is that you're heading down a slippery slope. Sure the nutrition information isn't very accurate, but they're the best numbers you're going to get. You don't have the information or tools to make a better assessment yourself. Trying to make up your own values will just lead you to more and more inaccurate estimates of the number calories you're consuming.