Quote:
Originally Posted by someone123
You can see there how the numbers bounce up and down. Some of this is due to labour market volatility and some of it is error.
The monthly employment numbers in Halifax are a small random sampling based on a survey. I don't think there is a way to know exactly how many people were working in the city during the past month (taxes would give you better data but I don't think businesses report monthly).
The sampling error is amplified when you focus just on the change in the employment total the year-over-year monthly change. Imagine if there were a good probability of being just 1% off in the estimate of the number of employed people in the city; you might go 1% over one month and 1% under during the same month the next year. This is an apparent large decrease in the number of jobs (about 4,000 less in the Halifax case) that is just sampling error.
|
You can see the sampling error if you "Add/Remove Data" here:
http://www5.statcan.gc.ca/cansim/a26...ataTable&csid=
The monthly point values from Feb 2015 to Feb 2016 are
6.3 6.5 6.8 7.0 6.6 6.2 5.8 5.9 5.9 6.1 6.2 6.5 6.8
The standard error for the Feb 2016 sample is 0.6 (it's 0.5 to 0.6 most months), so the confidence interval is 6.2 - 7.4.
The standard error for month-to-month changes is consistently 0.3, so the change of 0.3 from January - 6.5 (confidence interval 5.9 to 7.1) to February - 6.8 is entirely within the standard error, so not statistically significant.
The standard error for year-over-year changes is even greater at 0.7 to 0.8.
Thus the difference between February 2016 - 6.8 (or 6.2 - 7.4) and February 2015 - 6.3 (or 5.8 to 6.8) is 0.5, less than the error of 0.7 and therefore again not statistically significant.
In essence, we really don't know if the rate has gone up at all, though it appears to be trending upward since last August.