[time-nuts] Anybody have suggestions for time related science fair projects?

[jimlux]

On 5/11/18 7:07 AM, Philip Gladstone wrote:
> On 11/05/2018 07:23, jimlux wrote:
>> On 5/10/18 9:55 PM, Hal Murray wrote:
>>>
>>> A few months ago, I was a judge for the county level middle school 
>>> science
>>> fair.  (I'm not very good at what they wanted, but that's a different
>>> problem.)
>>>
>>> What sort of interesting time related experiments can a middle school 
>>> geek do?
>>>
>>> Borrowing serious gear may not be off scale as long as a youngster 
>>> can run it.
>>>
>>
>> The whole area of celestial nav is time related and uses very simple 
>> equipment -
>>
>> Telling time by measuring the sun in some way.  Occultation of stars 
>> by the moon.  Positions of jupiter's big 4 moons.
>>
>> Pendulum experiments.  If the student has a way to change their 
>> altitude, can they measure changes in g.  Driving a pendulum.
>>
>> Coupled resonators  (spring/mass, pendulum, vibrating rods)
>>
>> Measuring the speed of light (Fizeau or Michelson method? Other ways)
>>
>> Water clock, sand hour glass, etc.  Measuring performance variation 
>> over environmental variations.
>>
>>
>> the trick with good science projects is finding something that's not 
>> just a "lab demo" - where there's some engineering component to 
>> figuring out how to execute the demo with unusual or improvised 
>> equipment, or where you're measuring something that's not been done 
>> before.
> The advice that we got when doing a middle school science project was 
> that you wanted an experiment with only one variable (altitude or 
> temperature etc) and a  measurement of a single variable (maybe over time).
> 

and with multiple measurements possible - most middle school projects 
tend to be a "one and done" - you get big kudos if you show even basic 
statistical analysis - a simple significance test is a big deal, 
assuming you're not doing it with some cookbook calculator.  You'd need 
to be able to explain what it means to the judges.

And something where you show an appreciation of measurement precision 
and any curve fit.  I used to mark down projects where they used Excel 
to do a regression curve, and then reported the coefficients with 5 
digits of precision, on measurements with at most 2 digits. (and no, not 
thousands of measurements to get a sqrt(N) improvement)

AVAR is a kind of sophisticated concept - I think it would be hard for a 
middle schooler to adequately explain what it is (heck, there's enough 
trouble for people who do it for a living).