There are many sources of patent office and examiner statistics. See the Examiner Ninja as a great free example, obviating the need for paying for this data which the USPTO should make freely available in the first place. Patent practitioners can use this information to tailor the response strategy, possibly tweaking the approach depending on many factors. However, one must be careful in interpreting this data.

Consider the situation where an Examiner has a 75% allowance rate without appeal, and a 75% allowance rate with appeal. Does this mean there is no benefit to an appeal? Well, it of course depends because statistics are a complicated business and easily misunderstood. If you have had no action from the Examiner, then yes it makes sense that whether or not you appeal, your chance of success is the same. The trick is that when you are in prosecution (e.g., facing your second final rejection, as an example), you have much more data than before the case was started. Examiner decisions are not random, like coin flipping. In the case of a coin flip, even if you flipped a coin 3 times and had heads every time, your chance of heads on the next flip is still 50%. The same is not true with examiners making decisions about whether an invention is patentable. Knowing you had three rejections in a row from an examiner that typically allows 75% of the cases within three actions can mean that your chance of another rejection on the next action is higher than a decision in a random case by that Examiner. Another way to look at the situation is that your case is falling into the bin with the 25% of cases the Examiner will not allow. Now, the appeal statistics can be viewed in the proper light - you still have a 75% chance of obtaining an allowance through an appeal versus sticking with the Examiner and being stuck in the 25% of cases that will not be allowed. In other words, it is more likely than any case being appealed is one that would have fallen into the 25% bin of cases the Examiner will not allow. From this viewpoint, an appeal is better since of those cases, a large majority of appeals win. Taking the stats to the extreme illustrates the point further - equal rates with and without appeal do not mean you are no better off with an appeal.

Of course the above analysis is a simplified, non-rigorous example; but, even if you work through the detailed statistic analysis, you can see that just because the allowance rate with and without appeal are the same does not mean that you are no better off on appeal.