[rant] 1/ i don't want to get too Rumsfeldian before i have even had my tea this morning but... "we" (researchers, dev people, general) need to be more specific when we talk about "unintended" consequences. Sometimes we use it synonymously with "unanticipated"

2/ But there are both anticipated (or better: anticipate-able/foreseeable) unintended consequences as well as those that were unanticipated (unanticipate-able or unforeseeable).

3/ Of course there are plenty of things in the world that are hard to anticipate and we cannot plan programs, policies, and measurement strategies around *all* possible eventualities.

4/ But there are lots of things, based on common sense, the literature, theory, and bloody talking people before you get going that we *can* and therefore *should* anticipate.

5/ These may indeed be *unintended* but since they can be anticipated, program and policy design -- as well as measurement -- should do so.

There are several dimensions to why this is frustrating but one angle is that...

6/ We are often told that we need qual work to deal with unintended consequences. Of course, qualitative work can add a lot of texture and insight to *all* interactions with and consequences of program and policy change. This is a good thing.

7/ But when 'unintended' is used as a loose synonym for 'all bad stuff,' this is pretty lazy. Yes, qual work (or even just a few open-ended questions within quant work) is very useful for bringing out unexpected things.

8/ But to the extent that many bad/undesirable/unintended things (not that all unintended outcomes are bad, so that is also a poor synonym) are can be anticipated, they need not be left to "mop up" types of qual work. You can anticipate them! Do so!

9/ i could go on for much longer but bottom line is that we should be *much* more precise when throwing around this "unintended consequences" term -- or just abolish it. Rather, there are desirable & undesirable outcomes, which can be reasonably foreseen or not. Fin. For now.