In February, as reported by Statistics Canada?
No-one really knows, eh? For several reasons, amongst them the following: It would cost too much to poll all Canadians for information about their employment statuses, and each person's employment status is more or less fluid. Consequently, StatsCan draws a sample of Canadians, asks them about their employment situations at some point in time and then applies statistical methods to estimate the overall population figures. Since a sample can never exactly represent a whole population (even allowing for the care that StatsCan exercises) there is inevitably some statistical error in these estimates.
If you peer carefully at the StatsCan report for February you will find that they say that there is, in some sense, a probability of 0.95 that the actual employment rise in February was between -6,400 and 108,400. In other words, it's possible that there could have been a small loss in employment or a more substantial gain.
I read this here first.