Beyond The Commons

Aaron Wherry covers all the goings-on in and around Parliament Hill. Follow Aaron on Twitter: @aaronwherry

Categories: Beyond The Commons, Capital Read

And now, algebra

by Aaron Wherry on Monday, August 23, 2010 3:32pm - 0 Comments

In keeping with today’s theme, here is Mike Moffatt sorting out the time cost of the new census.

We haven’t considered the x variable, however. It could be that the long-form census will only be filled out by people who are particularly efficient at filling out the form. This would mean that the value of x under the old system is higher than the value of x under the new system. But of course, this would also mean that we’re admitting that the new system provides a biased sample, and is therefore unreliable!

In other words, the changes to the census have either increased the administrative burden on Canadians or provide near worthless data or both. Increasing the time cost of the census to Canadians by over 20% should not be ignored.

Tags: census, Mike Moffatt

More by Aaron Wherry

WDM

Algebra? Wherry, it's a Monday, give a guy a break.
http://onelinecritic.wordpress.com/ DirtyOldTown

Okay – I give up! I'll tell you whatever you want to know! Anything! Just stop the math – for the love of God stop the math!
LC Bennett

Not to worry, this isn't math, algebra or any other objective discipline. It is opinion, conjecture and guesstimation; the equivalent of reading sheep entrails.
Cats

Mike Moffat misses something.

We can define "administrative burden" and "opportunity costs" in a meaningful way by phrasing a question like so:
Did you have something better to do with your time ?

With a voluntary survey the answer is always NO. (If they did, they would not do the survey.)
With an involuntary survey the answer is frequently YES. (They have no choice.)

The counter argument that comes to mind is why don't people just do the form when they have nothing better to do ? (Making the assumption that spare time with the kids has zero value).

Well, like taxes, this will probably be done at the last minute when Canadians have a million other things on the go.

Cats!

Cats

Hey Cats ?

Yes! (I'm Paul Wells again!)

Do you think Mike Moffat was just being disingenuous when he raised the point about that nasty Mr. Harper wasting the time of Canadians ?

Of course I do!

He was just find a round about way to restate an old objection – that this survey will have a low response rate.

But hey! It gave Aaron Wherry an excuse to dredge up this long forgotten, no longer controversial, nobody cares about it issue didn't it ??

Hey Cats? How long can a dead horse be flogged for ?

As long as the media likes Cats. As long as they like.

CATS! (AND CATS!) AWAYyyyyy.
Emily

Warning: Cons are now about to equate filling out the census with bad parenting.

Gaunilon

There's a flaw in Moffat's logic, right here:
" But would anyone anywhere consider a 57% response rate to the voluntary census to provide useful data?"

Answer: yes, because 57% for y=0.33 would represent roughly the same number of responses as the 95% response rate for the y=0.2 case. In fact that's why it works out to the same time burden assuming x remains unchanged.

For the same sample size the time burden will be the same. If the time burden increases that will only be because the sample size has increased (assuming x remains unchanged) which will provide greater confidence – a benefit.

Of course, if the voluntary census becomes biased, that's a real problem that has been addressed quite thoroughly elsewhere.

LynnTO

Same number of responses, perhaps. But would the distribution be constant? That, to me, is the greater issue – not the number of responses a mandatory versus voluntary survey would receive, but the distribution of those responses among the population.

Fake edit: and a Census isn't really a Census if it doesn't receive a critical mass of responses. It's just an attempted Census.

Gaunilon

If the distribution is weighted, then that's the sampling bias issue everyone has already discussed ad nauseam. Yes, that would be a problem, but that's not the problem Moffat is addressing here.

"…a Census isn't really a Census if it doesn't receive a critical mass of responses. It's just an attempted Census."

It doesn't matter what the response rate is; all that matters is the total number of responses compared to the overall population. If you send it out to every household in Canada and get a 20% response rate, you've got the same efficacy as if you send it to 20% of households and get a 100% response rate (again, neglecting sample bias).

Olaf

Ok, I'm not fantastic with algebra, statistics, numbers, concepts, arguments, or thinking*, but if anything, wouldn't the fact that it was mandatory, assuming that the total number of responses are approximately equivalent (under the x=.57 scenario), and assuming a relatively constant value for x, actually lead to less opportunity cost? Follow with me, friends, on an epic journey of adventure and authorial incompetence.

Wouldn't having a voluntary survey mean that those who voluntarily choose to fill out the census consider it the greatest use of their time? So, the opportunity costs would be less for them than for others forced to fill it out? In other words, wouldn't the people filling out the voluntary census (sorry, the "voluntary useless survey") either value their contribution more (civic duty!) thus making their time commitment worthwhile to them (and thus seeing an increase in overall utility) or have nothing better to do, making those opportunity costs marginal (or, I guess, at least equivalent)?

So, shouldn't there be another variable that measures what the actual opportunity costs are for each individual? The opportunity cost for a CEO forced to fill out a survey when she gets home at 11:30, forcing her to ignore her homework, stay up late and be less productive the next day, be massive compared the opportunity costs for someone like, say, me or anyone else on Macleans.ca on a regular basis, who literally would have nothing better to do? Why are "time costs" held constant when a CEO's time is worth $50 a minute and mine actually costs my organization money (HR, b*itches!). And if the actual cumulative opportunity costs are significantly less under a voluntary useless survey than under a massively important, country-altering mandatory crucial information providing census, wouldn't the whole premise of Moffatt's (I assume, tongue-in-cheek) argument go to pot?

*Note: I've made at least 100 factual, analytical, mathematical, and spelling errors in the above ultra-amateur analysis. Can you spot them?

madeyoulook

Why are "time costs" held constant when a CEO's time is worth $50 a minute and mine actually costs my organization money (HR, b*itches!).

FLOOR MANAGER: Well, it's about time you owned up to it. Here is your termination notice, a cheque for two weeks' salary, and the three DVD-ROMs we needed to burn your "personal" files for you as we preserved your actual work on this 1.44MB floppy for ourselves. Joe here (motions to big burly security guard) has the empty box for your personal effects, and will see to it that you leave behind your keys and that you are off the premises within a half hour. If you really want a reference letter, you know where to — ah, you and I both know you won't want one from me. So long, Olaf.

Anon 001

And now, Alghabra!
http://thevillagevoice.blog.com Jeff Casselman

There is some merit to the algebraic argument.

Say for example you poll 18-25 year olds in 2011 with a voluntary census. Next time, in 2015, you happen to poll more 18-25 year olds voluntarily, thus biasing the result as it could be compared to it's last, fixed parameter, result. I mean to say, if the pool which you take from is not exactly the same size as the last pool (or at least measured proportionally with any population increase) it is not possible to get a proper measure.

FOR EXAMPLE: in 2011 the census measures how many people 25 and under smoke. They get their result. But the next Census, more 18 to 25's respond, and the incidence of smoking under 25 shoots up, even if it hasn't.
Tony

One other thing Mr. Wherry, if science is your argument you might want to cease the appeal to authority (as seen in some of your other BLOG entries)- that doesn't work
madeyoulook

Hey! I posted a comment with some algebra of my own, and it's gone poof. What gives?