Many believe the most important piece of a lab experiment is the result. While of course important, the most important part of experimentation is often not taught in high school - how to do an error analysis! That is, how believable and reliable are your data? If there is some huge spread in measurements with all sorts of uncertainty, then how good can results and conclusions possibly be?
It reminds me of polling results before big elections. Suppose A is up on B in polling 51% - 49%. If this is all your given in a newscast, and it's the day before, you might cheer that A is going to win!! However, how does this conclusion change when you are informed the uncertainty of the polling is plus or minus 3%? Suddenly, we cannot reach any real conclusion about the result...it is statistically a tie, a tossup, and we'll have a nail-biter waiting for the results to come in!
The standard way to handle uncertainty for measurements where we do multiple trials is the standard deviation. Check out how to calculate it, and how to interpret it.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.