The Student Room Group

help with physics graph please

Hi guys;

I have to make up a graph for a physics lab, and given that I have NEVER Studied physics I am badly struggling :frown:

I did a practical experiment to find the refractive index of light shone through perspex at varying angles.

I have made my table, got my results.

Problems I am having:

1) I am unsure as to how correctly add error bars. I need them with 0.5 +/-; but when I attempted that, they got very large and messy looking.
2) I need to provide a best fit line, a shallowest line and a steepest line. I am unsure how to do these either; because I am unsure as to the error bars :frown:

Can anyone please help
Original post by apronedsamurai
Hi guys;

I have to make up a graph for a physics lab, and given that I have NEVER Studied physics I am badly struggling :frown:

I did a practical experiment to find the refractive index of light shone through perspex at varying angles.

I have made my table, got my results.

Problems I am having:

1) I am unsure as to how correctly add error bars. I need them with 0.5 +/-; but when I attempted that, they got very large and messy looking.
2) I need to provide a best fit line, a shallowest line and a steepest line. I am unsure how to do these either; because I am unsure as to the error bars :frown:

Can anyone please help


You're quite likely to get large error bars for an experiment like this, as it's hard to measure angle correctly, etc which can give you large uncertainties.
The error bar should be a line 0.5 according to your scale on either side of each point (above and below or left and right, depending on which variable has the uncertainty)
Your lines of best fit are just the steepest and shallowest lines that can pass through the error bars.
If you're unsure, just plot the error bars as best you can, draw and calculate your gradients. Then do a some 'sensible checks' that can indicate whether you've done it correctly. For example:
Do you know what the gradient SHOULD be? is your average gradient roughly equal?
Should the graph pass through the origin? If so, does it? If not, do you know what the y-intercept should be? Is yours roughly equal to this?
These are the kinds of things you can use to deduce whether you've plotted correctly. Of course. if things don't tally up, it could be a problem with your data rather than the graph, but these can be a good indicator.
Hope this helps you slightly. If you still have trouble, let me know :smile:
Reply 2
My error bars go left, right up AND down; is this correct..........?

I am doing all this in excel,

"your lines of best fit are just the steepest and shallowest lines that can pass through the error bars."

I am unsure as to how I would calculate this using excel/ the graph that is produced in word? Would I do this manually, then specify the gradient for each?

Also, I have been told to provide 3 lines, one for line of best fit, one for shallowest and the last for the steepest. How would I specify what points?

Im sorry for the dumb questions, I am not lazy, I just have studied physics 12 years ago and through an adminstrative oversight, stuck doing a module 4 weeks late :frown:
Reply 3
Original post by halfhearted
You're quite likely to get large error bars for an experiment like this, as it's hard to measure angle correctly, etc which can give you large uncertainties.
The error bar should be a line 0.5 according to your scale on either side of each point (above and below or left and right, depending on which variable has the uncertainty)
Your lines of best fit are just the steepest and shallowest lines that can pass through the error bars.
If you're unsure, just plot the error bars as best you can, draw and calculate your gradients. Then do a some 'sensible checks' that can indicate whether you've done it correctly. For example:
Do you know what the gradient SHOULD be? is your average gradient roughly equal?
Should the graph pass through the origin? If so, does it? If not, do you know what the y-intercept should be? Is yours roughly equal to this?
These are the kinds of things you can use to deduce whether you've plotted correctly. Of course. if things don't tally up, it could be a problem with your data rather than the graph, but these can be a good indicator.
Hope this helps you slightly. If you still have trouble, let me know :smile:


Help me get this finished for real? And Ill actually send yu money via paypal :smile:
Original post by apronedsamurai
My error bars go left, right up AND down; is this correct..........?

I am doing all this in excel,

"your lines of best fit are just the steepest and shallowest lines that can pass through the error bars."

I am unsure as to how I would calculate this using excel/ the graph that is produced in word? Would I do this manually, then specify the gradient for each?

Also, I have been told to provide 3 lines, one for line of best fit, one for shallowest and the last for the steepest. How would I specify what points?

Im sorry for the dumb questions, I am not lazy, I just have studied physics 12 years ago and through an adminstrative oversight, stuck doing a module 4 weeks late :frown:

Your error bars should up, down, left and right if both variables have uncertainties, ie. you have uncertainties in both your independent and dependent variables.
Personally, I don't think excel works very well for this kind of thing, so I prefer to draw graphs by hand, it usually turns out to be quicker :')
The line of best fit is the line that passes through as many points as possible, with an equal number of points on either side.
In excel, you should get it to plot a scatter graph, and it has a line of best fit button that does it automatically.
As for max/min gradients and error bars, I don't really know how this works on excel - like I said, hand-drawing is usually easier, but i'm sure if you googled, there must be a way! (sorry, not much help there! :s-smilie:)
"How would I specify what points?" - I'm not really sure what you mean by this...
Reply 5
Original post by halfhearted
Your error bars should up, down, left and right if both variables have uncertainties, ie. you have uncertainties in both your independent and dependent variables.
Personally, I don't think excel works very well for this kind of thing, so I prefer to draw graphs by hand, it usually turns out to be quicker :')
The line of best fit is the line that passes through as many points as possible, with an equal number of points on either side.
In excel, you should get it to plot a scatter graph, and it has a line of best fit button that does it automatically.
As for max/min gradients and error bars, I don't really know how this works on excel - like I said, hand-drawing is usually easier, but i'm sure if you googled, there must be a way! (sorry, not much help there! :s-smilie:)
"How would I specify what points?" - I'm not really sure what you mean by this...


Ok; so......draw the points by hand; and the higher they are than the N value of perspex; the more inaccurate the results?

Do I draw three seperate graphs, or one graph and include three lines on it?
(edited 8 years ago)
I'd just say in general you get most of the marks for 'doing the right things' rather than getting the right answer... so a graph with bad error bars is much better than a graph with no error bars at all because you weren't sure about doing error bars correctly.

if it was me I'd plot all 3 lines on the same axes and either use different colours or styles
e.g. solid line, dashed line - - - - - , dotted line . . . . . . to differentiate them
which means you need a key (or legend)

FWIW the free stats package 'R' does nice 'academic' looking graphs with error bars etc. but it's quite an effort to get started with so if you're close to a deadline I'd just do it by hand for now to save time.
Original post by apronedsamurai
Ok; so......draw the points by hand; and the higher they are than the N value of perspex; the more inaccurate the results?

Do I draw three seperate graphs, or one graph and include three lines on it?


I would include all the lines on one graph.

Quick Reply

Latest