Friday, April 09, 2010
Error term
In any regression analysis, there are always the 'observed' variables which could be used to be the explanatory variables.
On the other hand, there exist these 'unobserved' variables which could affect the explained variable as well, but since we can't really observe it, it's hard to control for them in our analysis.
Finally, there is this mystical creature called the 'error term'. The error term refers to random effects that are almost impossible to account for, such as luck, inspiration, opportunities, fate etc.
Assuming what you care is the log wage.
Don't ever pray for to achieve certain values for the 'observed' variables as you can hardly control them, e.g. age, race and gender.
Don't ever pray for the 'unobserved' variables as well. Even though other people can't see it, you know it yourself. It's all inside you, and your're the only one who can determine if these variables come to fruition. Don't ever shift to responsibility to some higher being. He has already planted them in you, and u should dig them out on your own accord. These variables include determination, charisma, ability to do things well, personal relationships etc.
Pray only for the 'error term', are these are the only things that are out of your control, e.g. luck, inspiration, opportunities, fate etc.
As a saying goes, God only help those who help themselves. If you can't even bother to work hard at the 'observed' and 'unobserved' variables, can you ever expect God to help you with the 'error term'?
Having said that, even though I've done my part in creating the former two types of variables, at this stage I would need the 'error term' to be positive. For that, I pray. I might or might not get it, but in the end, at least I know that I've done my part to my very best, and I can safely blame the error term more than anything else.
lowtide blogged @
11:45 am
