Many Pies

Many Pies

Tuesday, October 10, 2006

Data analysis - always more complicated than you think

My first proper job (apart from working on a pig farm) was as "Computing Support Officer" for what was then York Health Authority and is now probably North Yorkshire and York Primary Care Trust.

One of the things we did was a survey of mental health patients. Once the data was in a database I had two Community Mental Health nurses coming back to me again and again with questions on the data, and analyses that they wanted me to do. Twenty years on and I'm doing the same thing. As before, it's always more complicated than you think, and it amuses me that it is.

So I'm asked, "can you tell me who gives regularly?"
So I ask, "by standing order?"
"Do you want to know those who have just filled one in, or those who are actually giving by one."
"The latter."
"Do you want me to spot those who give quarterly and annually too?"
"Yes, of course."
"And presumably you want me to not include people who missed their last payment?"
"Yes, of course."
"So I need to check that in all the times when they should have given that we have actually received a gift."
"Yes, it's more complicated than I thought isn't it?"
"Yes. OK, moving on from that. What about those who regularly given cheques."
"Yes, we we want to include those too."
"What do you mean by regularly."
"Well, every month or so."
"So say 10 times out of the past 12 months?"
"And those who have started giving that way in the past year, you want me to reduce the amount proportionally - so 5 times in the last 6 months, say?"
"Yes, anything else?"
"Direct debits...."

I don't mind having conversations like this. Just so long as I manage to tease out what they really want.


No comments: