Page 1 of 13

Experimental Design B/C

Posted: August 4th, 2018, 11:37 pm
by Adi1008
Experimental Design B/C: This event will determine a participant's ability to design, conduct and report the findings of an experiment conducted entirely on site.

Experimental Design Wiki

Experimental Design Test Exchange 2018

Past Threads: 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018

Re: Experimental Design B/C

Posted: September 4th, 2018, 5:08 am
by dxu46
Big rules change...thoughts?
EDIT: also, we can't make up data anymore :/

Re: Experimental Design B/C

Posted: September 4th, 2018, 12:35 pm
by Anomaly
dxu46 wrote:Big rules change...thoughts?
EDIT: also, we can't make up data anymore :/
Smh you shouldn’t have been doing that in the first place

Re: Experimental Design B/C

Posted: September 4th, 2018, 1:29 pm
by nicholasmaurer
dxu46 wrote:Big rules change...thoughts?
EDIT: also, we can't make up data anymore :/
I think it is a step in the right direction. Under the old system, teams were incentivized to begin writing the analysis sections etc. before all the data had been collected, which defeats the point. I also like the new checklist format better - it will reduce the number of math and totaling errors.

However, I am still skeptical of the event in general. One primary concern is there is often little to distinguish between the top teams. At the Solon HS invitational last year, the best fifteen teams were separated by less than 5 points. Top teams memorize a framework that covers all of the rubric requirements and simply modify this to fit a different simple experiment each tournament. If you read their XPD write-ups from different tournaments, they are remarkably similar and formulaic.

Therefore, I fail to see the intrinsic value of XPD with the current format; real scientific articles (and academic lab reports) vary much more widely to meet the needs of the experiment/topic/field. Because teams simply regurgitate similar write-ups each time, the current structure of this event does little to assess students' understanding of even the most basic concepts of experimental design: study types, biases, statistical analysis, statistical power/sample size, research ethics, etc.

Re: Experimental Design B/C

Posted: September 4th, 2018, 3:08 pm
by TheChiScientist
^COMPLETELY AGREE
The fact is that competitions will always come down to 1 or 2 points. Not assuming for ties which are extremely annoying. As for the new rule change, I think it will shake up the operations of many teams as the final portion can be one of the most time-consuming.

Re: Experimental Design B/C

Posted: September 4th, 2018, 4:52 pm
by Unome
nicholasmaurer wrote:
dxu46 wrote:Big rules change...thoughts?
EDIT: also, we can't make up data anymore :/
I think it is a step in the right direction. Under the old system, teams were incentivized to begin writing the analysis sections etc. before all the data had been collected, which defeats the point. I also like the new checklist format better - it will reduce the number of math and totaling errors.

However, I am still skeptical of the event in general. One primary concern is there is often little to distinguish between the top teams. At the Solon HS invitational last year, the best fifteen teams were separated by less than 5 points. Top teams memorize a framework that covers all of the rubric requirements and simply modify this to fit a different simple experiment each tournament. If you read their XPD write-ups from different tournaments, they are remarkably similar and formulaic.

Therefore, I fail to see the intrinsic value of XPD with the current format; real scientific articles (and academic lab reports) vary much more widely to meet the needs of the experiment/topic/field. Because teams simply regurgitate similar write-ups each time, the current structure of this event does little to assess students' understanding of even the most basic concepts of experimental design: study types, biases, statistical analysis, statistical power/sample size, research ethics, etc.
I have heard that MIT in one of the more recent years was different. But in general, I agree. This sort of criticism of the event seems to have gotten progressively more common over the last few years.

(then again, do we really want whatever else Inquiry would give us instead?)

Re: Experimental Design B/C

Posted: September 4th, 2018, 5:17 pm
by nicholasmaurer
Unome wrote: I have heard that MIT in one of the more recent years was different. But in general, I agree. This sort of criticism of the event seems to have gotten progressively more common over the last few years.

(then again, do we really want whatever else Inquiry would give us instead?)
I wouldn't remove XPD (which would require Inquiry to find a replacement) - I would reformat it. Rather than have teams perform an experiment, I would suggest breaking it up into sections.

For example:

One portion could be a written test covering topics such as biases, study design, research ethics etc. Have another section where sample data is provided and teams need to perform the appropriate statistical analysis. Another section could present a question/topic and ask teams to design a suitable protocol for researching it.

Just a thought...

Re: Experimental Design B/C

Posted: September 9th, 2018, 8:34 pm
by dxu46
nicholasmaurer wrote: I think it is a step in the right direction. Under the old system, teams were incentivized to begin writing the analysis sections etc. before all the data had been collected, which defeats the point. I also like the new checklist format better - it will reduce the number of math and totaling errors.
I agree that this is a good idea, and also according to soinc.org, the total amount of points is 110 instead of 106 like on the rules.
nicholasmaurer wrote: However, I am still skeptical of the event in general. One primary concern is there is often little to distinguish between the top teams. At the Solon HS invitational last year, the best fifteen teams were separated by less than 5 points. Top teams memorize a framework that covers all of the rubric requirements and simply modify this to fit a different simple experiment each tournament. If you read their XPD write-ups from different tournaments, they are remarkably similar and formulaic.
I completely agree with this point, although to me, it seems like the rules include the procedure for tiebreaks purposely vague so that competitors have to actually try to go above and beyond instead of doing a "fill-in-the-blank" style write-up and judges have to be able to tell which teams are obviously more skilled and have put more effort. I understand what you mean, and this could be remedied with more sections, although idk how teams are going to get more than what there already is done.
nicholasmaurer wrote: Therefore, I fail to see the intrinsic value of XPD with the current format; real scientific articles (and academic lab reports) vary much more widely to meet the needs of the experiment/topic/field. Because teams simply regurgitate similar write-ups each time, the current structure of this event does little to assess students' understanding of even the most basic concepts of experimental design: study types, biases, statistical analysis, statistical power/sample size, research ethics, etc.
IMO, this event is more about partnership, cooperation, and scientific procedure than actually experimenting. Like, don't we already know that if you drop a ball from a tall height, it will rebound? Also, teams are given only 50 minutes to write so much, they would need extreme teamwork abilities to finish the experiment and write-up. The new format for the event seems to emphasize this part, as you have to use your time wisely.

These are only my thoughts, and they shouldn't be taken too seriously.

Re: Experimental Design B/C

Posted: September 15th, 2018, 11:05 am
by Jacobi
I just realized that you can load the rubric into your programmable graphing calculator.

No more rubric memorization!

Re: Experimental Design B/C

Posted: September 15th, 2018, 11:22 am
by dxu46
Jacobi wrote:I just realized that you can load the rubric into your programmable graphing calculator.

No more rubric memorization!
Only division C though :(