@puzzledperson,
Incidentally, we have to be careful when talking about conviction rates. Does this refer to convictions as a percentage of:
(1) Total rapes (including unreported) assumed from crlme victimization surveys?
(2) Rapes reported to the police?
(3) Cases where an arrest is made?
(4) Cases where a prosecutor decides that evidence is sufficient to move forward?
(5) Cases which make it into the court system (including plea deals without jury trial)?
(6) Cases which go to trial by jury?
In Britain, for example, the rape conviction rate in case (6) is 58 percent, which is actually high compared to the same statistic for some other violent crimes:
"A report in to the effectiveness of juries, published by the Ministry of Justice last month, confirmed that rapists are more likely to be convicted than acquitted, with higher success rates than those for rates for grievous bodily harm, threats to kill, manslaughter and attempted murder."
http://www.telegraph.co.uk/news/uknews/crime/7442785/Rape-conviction-rate-figures-misleading.html
Now, that may itself be misleading, to the extent that prosecutors cherry-pick cases they consider winnable; but on the other hand prosecutors commonly do that with all cases, regardless of the charge.
So the question comes down to why a case might be winnable. These days DNA evidence is common in cases that make it to trial; but there is still the issue of consensuality. If someone has traces of a rape drug in their system or bodily injuries consistent with forced penetration and/or defensive injuries, that's likely to make a better case. On the other hand, if two individuals get really drunk and the woman wakes up the next day after passing out, finds evidence of sexual activity, but doesn't recall giving consent, that may make for a poor case for prosecution, even if she was in fact raped.