The Off Topic Topic

[crazy canuck]

That's one monthly movement.  Even in US there are times when both employment and unemployment numbers go down in the same month.  Given stated statistical uncertainty, that's guaranteed to happen occasionally. 

I don't know the details of how it's measured in Canada, but I imagine statistical noise must be even greater than in US due to lower population (unless they keep the sample size at US levels), if it's measured in roughly the same way.  Whatever the reason, drawing conclusions based on one single monthly report in unwise.

The conclusion was that the unemployment rate dropped that month largely because young people had stopped looking for work that month.

No one is attempting to predict what might happen next month, year, or decade.  But for now we know that the participation rate among young people dropped.

[DGuller]

That's one monthly movement.  Even in US there are times when both employment and unemployment numbers go down in the same month.  Given stated statistical uncertainty, that's guaranteed to happen occasionally. 

I don't know the details of how it's measured in Canada, but I imagine statistical noise must be even greater than in US due to lower population (unless they keep the sample size at US levels), if it's measured in roughly the same way.  Whatever the reason, drawing conclusions based on one single monthly report in unwise.

The conclusion was that the unemployment rate dropped that month largely because young people had stopped looking for work that month.

No one is attempting to predict what might happen next month, year, or decade.  But for now we know that the participation rate among young people dropped.

We don't really know what happened in one month.  All the numbers are sampled, so statistical noise in certain situation is bound to pop up, and drawing conclusions is silly (even if people who should've known better did).  And your your original statement that started this discussion was a bit more general than that, your tense implied a continuous pattern rather than a meaningless thing that happened in one month.

[CountDeMoney]

We don't really knows what happened in one month.

There are certain things that occur during some months that are known, such as jumps and cuts due to seasonal hiring (Christmas) or industry-specific contraction and expansion (energy).

[DGuller]

We don't really knows what happened in one month.

There are certain things that occur during some months that are known, such as jumps and cuts due to seasonal hiring (Christmas) or industry-specific contraction and expansion (energy).

That's why such numbers are seasonally adjusted, at least in US.  You want the numbers to inform you of what you don't already know, not of what you know will happen.

[CountDeMoney]

I did notice in the unemployment analysis there's a shortage of construction workers in some regions.  Go fig.

[alfred russel]

That's one monthly movement.  Even in US there are times when both employment and unemployment numbers go down in the same month.  Given stated statistical uncertainty, that's guaranteed to happen occasionally. 

I don't know the details of how it's measured in Canada, but I imagine statistical noise must be even greater than in US due to lower population (unless they keep the sample size at US levels), if it's measured in roughly the same way.  Whatever the reason, drawing conclusions based on one single monthly report in unwise.

The conclusion was that the unemployment rate dropped that month largely because young people had stopped looking for work that month.

No one is attempting to predict what might happen next month, year, or decade.  But for now we know that the participation rate among young people dropped.

What DGuller is saying is a known issue.

Basically, if you are running a voter preference poll, you ask a bunch of people their opinions, and then present them within an uncertainty band/confidence interval. For example, a poll might be that Macron would beat LePen 62-38 with a margin of error of plus or minus 4 pts.

Regarding economic statistics, the process of gathering the information is similar--you ask a bunch of people their opinions and then extrapolate to the larger population. But economic statistics are typically not presented with a margin of error--but rather as fact. So the unemployment rate fell from 6.5% in month one to 6.3% in month two. That could easily be within a margin of error if it was reported. It almost certainly would be.

But since the statistics are reported as fact without margins of error, as are the more detailed demographic data of the people responding to the survey, it is possible to report the causes of the movement in the statistic (it is as easy as looking into the survey data and seeing which demographic group caused the change month to month--regardless of the statistical significance of the change).

The process makes sense from a practical point of view--otherwise almost every month the report would read, "Unemployment last month was x.x% with a margin of error of x.x%. This month it was x.x% with a margin of error of x.x%. The changes were not statistically significant. See you next month."

[Josquius]

True. The media pretty much never mentions margin of error. I'm sure they used to write in small letters +/- x next to their figures but no longer.

[Razgovory]

My dad had another heart attack.  He's in the hospital right now.

[Eddie Teach]

Sorry, Raz. 

[Josquius]

Just learned today my uncle died of a heart attack.
'tis the season

[mongers]

My dad had another heart attack.  He's in the hospital right now.

He's hoping it turns out ok.

[The Brain]

I'm sorry guys.

[Josquius]

Odd thing. Apparently 3 weeks before my uncle died he gave a sealed letter to his daughter, to be delivered to his wife in case he ever died.
Amongst the expected it mentioned his funeral wishes (in tune with mine actually. Bar the suing people part)

Also a  scary realisation...
My great grandad died in his late 80s.
My grandad died in his late 70s.
My uncle died in his late 60s.

1: my dad is nearly 60
2. I'll die in my 50s?

[Eddie Teach]

But won't you be relieved when you do?

[Grey Fox]

Odd thing. Apparently 3 weeks before my uncle died he gave a sealed letter to his daughter, to be delivered to his wife in case he ever died.
Amongst the expected it mentioned his funeral wishes (in tune with mine actually. Bar the suing people part)

Also a  scary realisation...
My great grandad died in his late 80s.
My grandad died in his late 70s.
My uncle died in his late 60s.

1: my dad is nearly 60
2. I'll die in my 50s?

If you keep that life style, yes undoubtedly.