The latter more than the former. The former as a state of mind perhaps up to a point, rather than of action. For what it is worth, I don't know of any lawyer with whom I have dealt, who would hold the views expressed by some on this thread.
A fundamental problem with attributing social phenomena to education levels is that education is closely correlated with both wealth and urbanization, among other things. It can be difficult to distinguish which factor is influencing what, and for that matter it's an open question to what degree they should be distinguished.
I would say that's the reason why class is such a useful defining factor, because class merges multiple factors under one unifying conceptual rubric, although it's primarily meant to signify economic status (needless to mention).
That all said, I think there's no doubt that education levels are highly correlated with tolerance of liberated sexuality (not just with regard to homosex, but other types as well), but in a somewhat paradoxical manner. Very generally speaking, the highest degrees of both tolerance and non-normative behavior may be found at the highest and lowest extremes of class (and education). The highest degrees of conformity and intolerance may be found within the middle ranges, especially amongst the upwardly mobile.
In my view the reason for that, to put it as simply as possible, is because the lowest class has little to lose and the highest class has little to fear.
Now, I want to backtrack a bit and clarify just one point, with regard to the correlation between the status of women and the expression of homosexuality. From a historical standpoint, there are three general systems of male homoerotic behavior: age-structured, transgenderal, and androphilia. The most prevalent system in the modern West is (egalitarian) androphilia - and that's what is commonly referred to and conceptualized as homosexuality.
Now, by and large (although hardly exclusively) androphilia is a modern phenomenon, or at least it's a modern phenomenon as the cardinal system. My comments above with regard to women's lib and gay lib are intended for that context. In premodern societies, the correlations are different.
More precisely, societies in which age-structured homoeroticism is normative or prevalent (and this includes the subset of class-defined homoeroticism) the opposite correlation appears to be true: A higher prevalence of age-structured homoeroticism tends to correlate with lower status for women. This may perhaps be the case because age-structured homoeroticism also tends to correlate very strongly with martial organization (e.g., Greek hoplites and Japanese samurai).
On the other hand, societies in which transgendered homoeroticism is more prevalent tend to correlate with a higher status of women. This may perhaps be the case because societies that feature normative 'third genders' tend to be less advanced (e.g., the Amerindian berdache or the Oceanic mahu). More to the point, they tend to be less urbanized, and before the modern era the status of women tended to be lower the more urbanized a society.
These are not the only factors, nor is this a complete overview, but it's enough for an idle internet post!
PS. I realize two of you may not care, but no one's forcing you to read this! I'm just pinging you because you were part of the earlier discussion.