![]() 04/23/2018 at 20:00 • Filed to: None | ![]() | ![]() |
In statistics, at least, your answer will never round exactly as it did for your professor. No matter if you round after every step or never round until the end. It’ll always be off somewhere after the tenth place.
![]() 04/23/2018 at 20:09 |
|
Sig figs FTW!
![]() 04/23/2018 at 20:13 |
|
significant digits maybe?......
![]() 04/23/2018 at 20:16 |
|
Algebra all the way to the end, don’t round ever, then sig figs?
Mind you this is why my students hate webassign.
![]() 04/23/2018 at 20:17 |
|
I had a physics lab TA who marked us all wrong on something because he was using a rounding scheme that no one in the class had ever been taught (one that sometimes rounds .5 up and other times rounds it down).
![]() 04/23/2018 at 20:24 |
|
The Econ class I had to take sophomore year that was really a Stats class, was like this. I’m actually kinda having this problem with an International Finance class, but the professor actually recognizes this.
![]() 04/23/2018 at 20:29 |
|
Does the 10th decimal really matter tho? If it doesn’t then neither should the prof care about it (I know that is not always the case).
I’m a grader for a subject (not statistics) and I don’t even look at the 1st decimal.
![]() 04/23/2018 at 20:35 |
|
.4 rounds down, .6 rounds up, .5 rounds even. It’s arbitary but statistically more valid. I threaten my students with it on occassion.
![]() 04/23/2018 at 20:49 |
|
It was many years ago, but I thought I hated Statistics with a passion. Then I got to Econometrics and really found out what hate was.
![]() 04/23/2018 at 23:31 |
|
I’ve never even heard of that, so I’ll assume Hitler himself is the only thing worse.
![]() 04/23/2018 at 23:33 |
|
Depends. His tests are multiple choice and the fifth answer is always “none of the above” so sometimes you get 0.2700 as your answer and the closet one is 0.2581 and you have know idea if you just messed up on the rounding or the answer is “none of the above”
![]() 04/23/2018 at 23:45 |
|
My rule of hand is that max error allowable is 5%