a thoughtful web.
Good ideas and conversation. No ads, no tracking.   Login or Take a Tour!
comment by user-inactivated
user-inactivated  ·  3761 days ago  ·  link  ·    ·  parent  ·  post: Ethics and the cult I'm being forced to join

    significant time make sure our data looks good as opposed to the outcomes.

I never would have thought lying through your teeth was such a big part of being a nonprofit before I started writing software for them.





user-inactivated  ·  3758 days ago  ·  link  ·  

We haven't directly lied through our teeth (as far as I'm aware of). We just stretch the truth and manipulate our data.

An example of this is a questionnaire we have to fill out when a client is exiting the program. One of the required questions is if they have AIDS or have tested positive for HIV. If so, a secondary required question is if they are currently receiving treatment for it. It is not a required question if they have not tested positive for HIV/AIDS.

The issue with this is the data summaries. There's a big chart that shows the answers to these questions for every client within a time constraint. If an answer is not filled in, the unanswered cell is filled red and bold. So, quickly scrolling through this chart you see a bunch of big red boxes that look like mistakes or missing data. IF questions make up about half of all of the total amount of questions.

The report that summarizes all questions by completion percentage makes it appear just as bad, because it's just straight percentages of all the questions answered. So if 95% aren't receiving treatment for HIV/AIDS, there's a bar sitting at 5% in between 100% bars, which makes our data look bad so somebody apparently.

This way of handling data came from a county-wide meeting of service providers held by the county organization that is in charge of regional database. I don't like harping on degrees or qualifications because I lack them as well, but their data analysts have no background in tech or data at all. They don't seem to understand how fucked or meaningless our data is on the whole, because I cannot leave any questions blank or answered with "Don't Know" or "Refused." If they do, they don't care, because the goal must be to inflate the numbers.

This is all in the name of keeping HUD grants coming to my county, because HUD wants to do away with transitional housing and focus on rapid rehousing (Housing First model).

It's dumb. Really dumb. Nobody will be able to draw truly meaningful or accurate data from a dataset that's half populated with garbage. It is disingenuous and above all else, a huge waste of time I could be spending face-to-face with my clients.

But I guess that's just the way it is.

user-inactivated  ·  3758 days ago  ·  link  ·  

I maintained an HMIS for three years (not the one you're using, it doesn't exist anymore because none of us wanted to deal with that shit anymore), I know what you're talking about. I think you might have read "lying through your teeth" as more disapproving than I meant it. Telling the people with the money what they need to hear is what you've got to do, so it's what you do. It would amount to the same thing and waste far less time if grants came down to the local level with very loose restrictions along the lines of "use this to do stuff for homeless people" and reporting requirements amounting to letting an accountant check that you did.