This doesn't really make much sense.
Subjectively, i.e. from my viewpoint, everything listed is small. Not that anyone else would be interested in my subjective impression, so why ask?
Objectively, anything coming in below the population's average is small. That average is, as we know, a matter of continuing research, none of it particularly trustworthy. Methodology is always suspect. I saw one graph which superficially resembled the familiar Gaussian distribution - the one you'd expect to see were penis size a truly random variable (although there's no reason to think that it is) - except that there was a distinct bimodal factor. In other words, the peak was split; it was really two peaks rather than the expected single one. The problem seemed to be that some data points were measured to a precision (not an accuracy!) of 1/2 inch. Others were measured to 1/4 inch. Both data sets were combined on the same graph. This methodology would tend to depress 1/4 inch increments, and emphasize 1/2 inch increments. In other words, the twin peak was a sampling error artifact, and the graph was crap.