[mr_crankypants]

A preceding B does not imply that A causes B, it only implies that B doesn't cause A. Which doesn't really reduce the world of possible causal relationships by all that much, since there may be any number of other factors that weren't considered.

For example, it could just as easily be the case that adjustments to dysfunctional aspects of the company's internal politics improve both morale and productivity, but the morale change becomes apparent more quickly.

[crimsonalucard]

Correlation does not imply causation.

Little known trick to establish that A causes B is to trigger A at will and see if a B is observed.

If you trigger A 100 times and it precedes a B 98 times then you have established causation to a degree. Similar to the correlation coefficient there must be some causation coefficient.

[rightbyte]

You might imply it, but not triggering A should give less B than triggering A, too.

"Turning on my TV at 5 a'clock causes channel 4 to broadcast The Simpsons."

[crimsonalucard]

Good point. If B was just happening all the time, then a B would always happen after A. Have to make sure the inverse is true.

Though if B happens all the time, but also A causes B. Then there's really no way to establish causation.

"The Channel four studio has electrical sensors that will trip and broadcast the simpsons either at 5pm or when I turn on my TV at 5pm."

[clhodapp]

Sure. There is a subtle difference between "things that make employees happier also improve business performance" and "making employees happier improves business performance", but in both cases you generally want to do things that make employees happy if you want business success (you just might want to be a bit more selective about which happiness-boosting things).

[robocat]

Yeah, confounding* variable C causes happiness and profitability (A and B).

* https://en.wikipedia.org/wiki/Confounding