Why is it important to analyze data and not blindly trust software output?

  • Thread starter Thread starter Evo
  • Start date Start date
AI Thread Summary
Analyzing data critically is essential because software outputs can be misleading if the underlying data does not fit the expected model. In a quiz scenario, a student attempted to determine resistance using a linear fit on non-linear data, leading to an inaccurate conclusion. The discussion emphasized that while data itself is valid, the analysis must align with the data's characteristics to yield meaningful results. Participants noted the importance of understanding the context and limitations of measurements, such as temperature effects on resistance. Ultimately, the conversation highlighted the need for careful examination of data before trusting software-generated outputs.
Evo
Staff Emeritus
Messages
24,029
Reaction score
3,323
Awesome quiz!

Thank you ZapperZ for putting this together.

We had an awesome group of participants, and this was a HARD quiz! Questions carried a different number of points depending on the difficulty.

The winner with a total of 12 points was Ryuzaki! Way to go!

Our newest member, phys_12 answered the most questions correctly, with 7.

Points awarded

Ryuzaki - 12
phys_12 - 9
trollcast - 6
Jonathan Scott - 4
Darth Frodo - 2
HS-Science - 1

Congratulations to all!
 
Last edited:
Physics news on Phys.org
Congrats everyone! Nice work ZapperZ!

6 months Gold membership for Ryuzaki!
 
Evo said:
and this was a HARD quiz!
Is it possible to see the questions somewhere?
 
Zz promised to post them with solutions later.
 
mfb said:
Is it possible to see the questions somewhere?

I saved the last one, as I should know the answer.

It's making me a bit crazy.

(I'm blaming my flu for being able to answer only one question, and those darn Neutrino kids for welching me out of my one point by answering it before ZZ even posted the question... Curses to this new FTL generation! )

ZapperZ said:
A student is given a circuit containing a variable power supply, a resistor, and an ammeter all in series, and a voltmeter attached across the resistor. The student is then asked to determine, as accurately as possible, the resistance of the resistor.

Knowing that Ohm's Law stated that V=IR, the student made a series of measurement of potential difference (V) versus current (I) on the resistor. The student end up with this data set:

table1_zps43c6082b.jpg


Using Ohm's Law, he knows that the slope or gradent of PD versus I will give him the resistance of the resistor. He then enters these numbers into his data analysis program, runs a linear line fit routine, and it spits out the value of the slope that he wants, which is ~0.4 Ohm.

Sounds alright, doesn't it? However, there is one MAJOR problem with this. Can you describe what it is? You may assume that the voltmeter, ammeter, and power supply are "ideal", meaning they do not contribute to the accuracy or inaccuracy of the measured values. (2 points)
 
Hint: plot it.
 
Vanadium 50 said:
Hint: plot it.

it's non-linear. The plot is supposed to be linear.

Therefore the data is bogus.

I believe that was what Zapper was trying to tell us.

[ZapperZ] 2:42 pm: You guys went way too deep into this by giving the reason on why the curve isn't a straight line.
[ZapperZ] 2:42 pm: You don't have such info... all you have is the data.
[ZapperZ] 2:42 pm: The problem here is that the student was trying to fit a STRAIGHT LINE to a curve that isn't straight!
[ZapperZ] 2:43 pm: The straight line fit does not represent what the data looks like.
[ZapperZ] 2:43 pm: The numerical software will fit anything it is given and spits an answer, but it doesn't mean that the answer has any meaning.
 
The data isn't bogus. Data is data. It is the analysis that isn't valid. It is trying to make the data to conform to something which it isn't, which is a straight line.

Zz.
 
I enjoyed the quiz. I got caught out by the one about the motor boat and the raft. If I'd been properly awake, I'd have applied Galilean relativity and written the answer down immediately.
 
  • #10
Jonathan Scott said:
I got caught out by the one about the motor boat and the raft.

I had problems because of the wording (not Zz's fault, but limitations of my English). Up to now for me "after a while" meant "after a SHORT period of time" (something like a moment). Apparently it means "any period of time of unspecified length".

I never know when, where and why my English will blow.
 
  • #11
Thanks, Evo and Greg! It was an awesome quiz. Had me on the edge throughout. Absolutely loved it! :D Can't wait for the next one, ZapperZ! ;)
 
  • #12
OmCheeto said:
it's non-linear. The plot is supposed to be linear.

You're making an assumption that isn't necessarily true.

Question to ponder: would you expect the temperature of the resistor to change, as the current and voltage are increased?
 
Last edited:
  • #13
Redbelly98 said:
You're making an assumption that isn't necessarily true.

Question to ponder: would you expect the temperature of the resistor to change, as the current and voltage are increased?

I don't think I have a 40 watt resistor in the house. Unless of course you count my 40 watt antique warming globe.

But yes, of course the temperature will rise, with real world consequences.

And I think the question caught me off guard, as in my 40 years of electrical experience, I've never measured resistance in such a manner. I always use an ohmmeter.

But I think Zappers ultimate point was that you shouldn't blindly plug data into a piece of software without looking at the data, to see if the software's output really makes sense.

-----------------------------
And now I have to go figure out why two of the four cylinders on my outboard are not firing. There are two coils, and two sets of points. The two cylinders that are not firing are fed by one coil and one set of points. Time to take off the flywheel!
It's going to be 70+ degrees today, and I have a date with a lady friend and the river. :smile:
 

Hot Threads

Recent Insights

Back
Top