Okay, help a sister out here..... this whole thread has left me feeling dumb as a bag of hammers......
Quite a few people have given a POV as to what they believe patriarchy means, which boils down to a bias in many areas in favour of men - their biological form in design, their advantages in the capitalist system because they don't (generally) bear children, the default to male models for researching and treating diseases that affect women differently, and the fact that although rape is common, it is now under-reported (hence convictions going down) because women have experienced having to prove they really were victims to a higher standard than men proving they didn't do it (sorry, bit ham-fisted there but I'm sure most will get the gist).
Apparently none of these things down to any form of patriarchal bias, but are just one of those things due to natural effects of biological differences and we should be focussing on the "real" issues that women face, which aren't that many apparently because all the big changes have been made and what more do we really need?
I don't think anyone is saying that these kinds of things don't happen or exist, even if they might not agree about the nature of certain specific ones.
To me, the question is whether when we talk about them, is are we are remaining closely rooted in the material reality of the situation. Because if we are not, what we are talking about is something that is just a mirage.
So the very first thing that is important is to make sure your observations are correct and accurate, before you even try and speculate about causation or anything like that. This can be difficult even in the sciences at times, but even more so in sociological study where you have multiple complex systems and factors that may or may not be related, and you can't run experiements that control for variables and such. The best you can usually do is try to work with big data sets and control for variables in the analysis, but it's not perfect. One real problem is that no matter how you try and choose good categories of data to collect, or be open to seeing all kinds of links, whatever pre-conceptions you have about that will tend to assert themselves. What would the data show if you grouped it in a different way?
Once you have your data you look for patterns, for correlations, etc. So if your interest is in women, you will probably look for different patterns or disadvantages between women and men, if it is race between different racial groups, etc. This is tricky - if you narrow the focus too much you may miss that your smaller patterns are part of larger ones, maybe even that are outside the specific group you are interested in. But if you don't, it's too much to deal with, and you can't really make useful conclusions.
What you want to do is see the causation working through the patterns. But that causation isn't something abstract - it's material and specific. A chain of events, the way bodies are made, a real historical circumstance, an economic effect, a human psychological mechanism. This effect comes from this chain of causes.
For an abstract idea like patriarchy to be applied to all of those chains of events or real specific causes, or even to use it to describe all the relevant patterns and effects, you need to be really sure the things you are grouping together belong together - they are materially and specifically related. Otherwise it's rather like you decide that all things with wings are birds, and then go on to try and describe the actions, biology, care and nature of birds with reference to your grouping - your result won't be very useful or reflect reality.
This type of mistake (my husband the chemist is telling me - "that's junk science you just an't do that!") is really common in political ideology, economics, even psychology I think, because of the need to talk about abstractions. The abstractions can easily get far away from the reality, because the human doing the thinking begins to reverse the process - instead of looking to the material instantiation and carefully deriving a pattern and causation, they use the ideological tool to examine the material reality - and it will always give the data that the tool is meant to "see". You can see this effect in some of the more bizarre theological concepts of the middle ages, where an abstraction is reasoned out to the enth degree as if it was the thing it was meant to be talking about, but by the end the concept has been wholly unmoored from whatever that was.