How to draw line of best fit? (1 Viewer)

[QZP]

I'm unsure how long the line is meant to extend to

E.g. if the domain of my points takes place from x=1 to x=3, should my line only extend from x=1 to x=3 or go past it?

Look at Q22 http://www.boardofstudies.nsw.edu.au/hsc_exams/2013/pdf_doc/2013-marking-guide-physics.pdf
They extend past it :S

 

[Crisium]

I doubt you'd lose marks for going beyond the domain let alone staying only within it

But other than that when drawing the line of best fit just make sure that you draw a straight line that is at LEAST touching the plotted dot and covers as many dots as possible (you may miss certain dots if there were outliers in results)

Hope this helps

 

[acronical]

I would extrapolate anything a bit too far past it. Its ok to have a little bit extended, but if you need more than say 3mm extra, extrapolate it. Otherwise, I think you have to remember to have equal dots above and below the line of best fit.

 

[QZP]

I would extrapolate anything a bit too far past it. Its ok to have a little bit extended, but if you need more than say 3mm extra, extrapolate it. Otherwise, I think you have to remember to have equal dots above and below the line of best fit.

Sorry what do you mean?

 

[acronical]

Like when you have a broken dotted line after the solid line. The solid line is your line of best fit and the broken dotted line after that is the extrapolation.

 

[Hi there444]

I dont exactly get what you mean but since there are no points after the domain, I wouldn't draw any lines past the domain. But that's my word so dont take my word for it.

 

[Kurosaki]

I'm unsure how long the line is meant to extend to

E.g. if the domain of my points takes place from x=1 to x=3, should my line only extend from x=1 to x=3 or go past it?

Look at Q22 http://www.boardofstudies.nsw.edu.au/hsc_exams/2013/pdf_doc/2013-marking-guide-physics.pdf
They extend past it :S

Doesn't matter too much, I personally keep it in the domain though.

 

[Fizzy_Cyst]

Keep it within the domain. Your line of best fit is only valid for the data which you have, there is no way of telling if the LOBF is valid for any points outside those given, thus it would not be valid to use it outside the points given

 

[tomnomnom696]

shouldnt your line of best fit intersect the origin for things such as plotting a pendulum experiment? otherwise you assume that your experiment is already invalid?

 

[acronical]

Tomnomnom I would extrapolate in those instances to zero or to whatever values you need. I'm not 100% sure though if you should just extend the line to include zero in the domain or keep it within the domain of the dots.

However, I would still go about extrapolating the line of best fit wherever necessary rather than extending its domain to within outside of the points given/used.

 

[tomnomnom696]

Tomnomnom I would extrapolate in those instances to zero or to whatever values you need. I'm not 100% sure though if you should just extend the line to include zero in the domain or keep it within the domain of the dots.

However, I would still go about extrapolating the line of best fit wherever necessary rather than extending its domain to within outside of the points given/used.

alright thanks acronical

 

[Fizzy_Cyst]

shouldnt your line of best fit intersect the origin for things such as plotting a pendulum experiment? otherwise you assume that your experiment is already invalid?

It should (according to the equation), but it won't (as your experiment will have systematic error)

 

[theodore0307]

Make sure there's the same number of points on the bottom and top

 

[panda15]

Make sure there's the same number of points on the bottom and top

Not necessarily.

 

[tomnomnom696]

Make sure there's the same number of points on the bottom and top

untrue. the line of best fit is advantageous as it gives different weighting to different data depending on deviation from the trend. this minimises the error produced and thus there is not always the same number above as below

 

[acronical]

Ideally I would try my best to make sure there are equal points above and below the line and maybe one or two points touching the line but that doesn't always work unfortunately. However there are cases when its quite difficult to do so, which may be due to outliers involved, or if there is a massive spacing between the points, meaning they would have rather big errors in results in that case.