Experimental Design B/C

[Blwrunner]

What kind of graph do you recommend I do for states? Also I was wondering if anyone has a good website that explains standard deviation.

[JustDroobles]

What kind of graph do you recommend I do for states? Also I was wondering if anyone has a good website that explains standard deviation.

A quantitative experiment is the best kind of experiment to try to easily hit all of the points on the rubric. This means your independent variable and dependent variable should be numbers. If you do a quantitative experiment, you should should make a scatter plot with the independent variable on the x-axis and the dependent variable on the y-axis. Plot each of your trials. Then, for Division B, you draw in an approximate "line of best fit". This is one smooth line that shows the average trend of the data. This line does NOT "connect the dots". There should be about an equal number of points on each side of the line. Example: http://hotmath.com/hotmath_help/topics/ ... it-e-2.gif

(By the way, a common mistake I see for scatter plots is the use of trial number as an axis. You should not use the trials at all to determine the location of the points. However, it can be helpful to make each trials a different shape or color on the graph with a key to indicate which point is which trial.)

If you end up doing a qualitative experiment, which means your independent variables are either general qualities or different objects (examples: apple, a banana, and an orange OR red, blue, green, basically not numbers) then you should use a bar graph. The x-axis should be the independent variable and the y-axis should be the dependent variable. Your bars can be either each separate trial, or the average of all trials for each separate independent variable with error bars to show your standard deviation.

The following is a pretty good guide to standard deviation.
http://www.mathsisfun.com/data/standard-deviation.html
Something important to note is whether you use N or N-1 when dividing - you probably should be using N-1, because I doubt you will ever take a sample for an entire population in Experimental Design.

[nejanimb]

I wrote this set of annotations for a friend who would be running Experimental Design and what to do with scoring. I thought it might be helpful for some of you, since it explains (more or less) how to get all of the points on the rubric. Which is totally doable.

http://cl.ly/text/0z1B0f2r0A3i

[sturmde]

I usually don't include a mode, unless there actually is one. Usually there isnt

When I score this event, I would always credit someone for saying "there is no mode" whereas if you say nothing, you get nothing. Always note whether a mode exists or not if the data has multiple points.

David Sturm
UMaine

[hc1220]

Ok thank you. I'll definitely do that at Nats!

[Cjkowalcz]

Hey guys! I have this soon and I think I'm pretty well prepared. Wish me luck!

[caseyotis]

Good luck! I hope you fare better than I did!

[mrburrito]

Hey guys! I have this event soon and I think I'm pretty well prepared. Wish me luck!

So, how did you do?

[caseyotis]

I'm guessing it's going to be on Saturday, but... Maybe not. *shrug*

[fantasyfan]

nope, PA went today