When is an FOIA response not reponsive? Depends on who you ask, I suppose.
In theory, there are basically seven possible outcomes to a FOIA request - your request will be denied in full, your request will be granted in full, your request will be granted in part, no records will be "found" responsive to your request, your request will be deemed insufficiently detailed to permit a reasonable search, the associated fees will exceed what you're willing to pay, or your request will be referred to another agency That's the theory, anyway.
In reality, there's a lot more to it - a dozen reasons and sub-reasons for not releasing all or part of a document, just for starters. But, as far as the annual FOIA reports each government agency produce (or are supposed to produce), everything basically gets broken down into four possible outcomes - no records found, granted in full, granted in part, denied in whole. Simple, straightforward, and easy enough for even government employees to understand.
Under the United Kingdom's FOI Act, you can request copies of records - but you can also essentially ask questions, and the government has to compile a response. The U.S. Freedom of Information Act, on the other hand, only allows a person to request records. A lot of requests aren't super-specific; if you don't know what records a given agency has on a subject, who's to know if they only send you two of the three they find? Certainly not you, and under mounting pressure to clear backlogs of FOIA requests, I wouldn't be too terribly surprised if a lot of FOI/PA sections were at least occasionally interpreting "all records" as "some records" to save themselves time. Something's better than nothing, after all. Right?
That's probably not too bad. I mean, it definately contravenes the spirit, and probably the letter, of the FOIA itself, but it's better than, say, the CIA's canned "we neither confirm nor deny the existance of records responsive to your request" line. Everyone cuts corners at their job, after all, and being a FOI/PA officer - knowing that virtually none of your "customers" would ever be able to call your bluff - has to tempt the weak-willed into trying to make their job easier, at least once in a while.
What is bad, though, is allowing personal opinion to color your judgement. What's bad is outright lieing to a requestor. What's really bad is telling a lie that you can easily be called on.
That's what happened to me.
I requested some training materiel from the United States Marine Corps. It wasn't secret, classified, or heavily-restricted stuff; quite literally any Marine has access to it. It is, however, about what some might consider a "sensitive" subject - chemical and biological agent protection. That, I believe, is why I was lied to.
To be fair, the USMC has always been pretty responsive to FOIA requests - for a given definition of "responsive", anyway. And therein lies the kicker.
I was sent a response by email, about a week after I'd mailed the request off. Attached to the email was a PDF file of a Marine Corps Institute textbook on "Nuclear, Biological, and Chemical Reconnaissance and Avoidance". This, I should point out, is not the course I'd requested, nor does it contain any of the theoretically "sensitive" details on chemical and biological agents that the "Nuclear, Biological, and Chemical Personal Defense" text does. It's a response to my request, but is it "responsive?"
The email explained that the class for the materials sent to me had "replaced" the class whose text I'd requested. On the surface, that seems perfectly reasonable, but it proved to be pretty problematic.
First of all, on a purely technical level, I didn't write the Marine Corps Institute and ask that, out of the kindness of their hearts as freedom-loving defenders of democracy, they send me something. Rather, I made a Freedom of Information Act request. Why is this important? Because, theoretically, under the FOIA, you're supposed to get what you ask for. You can request copies of 1930's field manuals, if you're willing to pay for the photocopies, and you will get them, not a more recent edition. If you ask for copies of annual reports from years past, you won't get last year's report, because it's "more current", you will (or should) get the one you ask for. So, on a purely technical level, releasing something other than what was requested is definately not "responsive".
On a more subjective level, the explanation didn't make a whole lot of sense. The "personal defense" course was still listed on the Institute website, and it appeared still possible to enroll in. A little bit of searching discovered that the course in question was created in 2004 - while the textbook I was emailed was from 1999. Suspicious? A bit, I think.
I contacted the MCI officer who'd emailed me with a few questions, and he suddenly was telling a very different story. The course in question was still being offered, but the coursework doesn't exist as a PDF file, and they were out of the CD-ROMs, so he thought I'd be happy with a competely different text - one which conveniently exists as a freely-copied PDF. If I really wanted, though, he was willing to mail me a copy of the CD whenever they might again become available.
Now, I don't know for certain why I was lied to, but it's inarguable that the original response was not completely honest nor truly responsive, and it highlights a major problem with the FOIA in this country. As my experience demonstrates, even when you know exactly what you're requesting, there's a chance that you won't be given it, even when it exists. I made a relatively "perfected" request - for the coursework for a specific, named class, which I knew existed - and still didn't get it.
Given that a lot of FOIA requests are for things that the requestor isn't positive exists, or doesn't know the specifics of, there's no sure way to know that even "responsive" requests are actually responsive. How are you to know how many records might exist that are responsive to your request? If the government sends you three, odds are you're going to be satisfied, and more importantly (for them), you're probably never going to know that a fourth, fifth, and sixth report exists.
Even more worrying are imprecise requests, where it's clear that the requestor doesn't know the exact nature of the records in question, but can infer their existance. Several years ago - in fact, in August 2001 - I came across a reference in an annual government report to a joint intelligence handbook prepared by the USMC and the CIA, and requested a copy from the Marine Corps. I forget the exact wording of my request, but I cited the report where I'd seen the reference, and asked for a copy of that handbook. A week after September 11th, I received in the mail a small book - "Kosovo : A Field-Ready Reference Manual". There was no return address - heck, no postage! - just, I kid you not, a piece of paper with my address on it, wrapped around the book and held on with a rubber band. Since I never received any other correspondence about my request, and DoD manuals don't regularly just show up in my mailbox, I assumed the volume came from the USMC; given what had happened the week before, and that I'd addressed my request to the Navy at the Pentagon, I didn't really feel like bugging them.
But I have to wonder - is that really the book I'd requested? Was that really a responsive request, or were the FOI/PA staff just betting on my not being able to call their bluff?
I'll probably never know, nor will the hundreds of thousands of others who file FOIA requests with the government every year. It's perilously difficult to prove a negative, and FOIA appeals take years. For all that the government is supposed to be on the citizen's side, journalists, researchers, and everyone else who make FOIA requests are at the mercy of the government. This is why it's especially important to do your homework when making FOIA requests; while even the most precise requests - as shown above - are no guarantee of getting what you want, or rather what you ask for, knowing as much as you can about what you're requesting helps immeasurably in assessing the responsiveness of what, if anything, gets released to you.