Take a look at Project Blue Book Special Report #14. This was a study commissioned by the U.S. Air Force in 1955 and carried out by the Battelle
Any of the big studies you look at come to about the same conclusions: Anywhere between 80% and 95% of reports are typically found to have
conventional or natural explanations. Somewhere between 5% and 20% are found to have no conventional explanation. Most of the studies I've looked at
or read about tend to agree that hoaxes and frauds constitute a very small percentage of UFO reports, usually 1-2%.
The Battelle study ended up with about 21% of reports categorized as "unknown" - meaning that the appearance and maneuvers of the object reported
could not be fitted to the pattern of any known object of phenomenon.
Another interesting fact about Special Report #14 is that UFO reports that were considered to be of "excellent" quality - meaning those from the
most reliable witnesses, containing the greatest amount of information and displaying highest degree of internal consistency - were significantly more
likely to be categorized as "unknown" than were reports deemed to be of "poor" quality or the total body of reports taken as a whole.
To put it another way, while 21% of all the reports studied were declared "unknown," if you look only at the reports in the "excellent" category,
you find that 33% of those reports are "unknown."
Additionally, reports that did not contain enough information to allow investigators to identify the object were not put into the "unknown"
category, but were instead relegated to a separate category labeled "insufficient information."
The study also utilized an analytical tool called the "chi square test" which determines "the independence of two criteria of classification of
qualitative data." I'm not a statistical expert, so I don't claim to understand the exact ins and outs of this test. But as I understand it, the
test determines the statistical probability that two categories are qualitatively similar; in this study, it was used essentially to determine whether
the objects in the "unknown" category, as a whole, were likely to simply be "unidentified knowns." The probability of this turned out to be less
than 1%; in other words, the unknowns - as a group - were qualititively different than the knowns.
I'm always glad to find people who know little about this topic beginning to take an interest. You are right about one thing, there is a lot of
garbage out there associated with UFOs. It's all about using common sense and logic to sift the wheat from the chaff.
Here's some stuff to get you started:
The Symposium on Unidentified Flying Objects, held in 1968 by the U.S. House of Representatives Committee on Science and Astronautics.
If you read nothing else from these procedings, read the tesimony of James McDonald.
The UFO Evidence, published in 1964 by the National Investigations Committee on Aerial Phenomena.
NICAP was one of the big civilian investigatory bodies of UFO reports during the 1950s and 1960s. This publication is essentially the "unknowns"
from their files.
And, if you are willing to read a book on the subject, read "The UFO Controversy in America" by David Jacobs. This is a great and relatively
objective history of the phenomenon written by a professor at Temple University.
Best of luck.