kristers Posted April 2, 2009 Report Share Posted April 2, 2009 Alright, so,I graphed some data from an experiment and the points resembled a rational relationship between the variables, i.e., a curve fit would produce a curve resembling a rational curve.I needed to find a line of best fit in order to calculate the gradient so I changed one axis to "1/x" when originally it was just "x".This produced what I needed; the data now lined up linearly and the program fit a linear fit perfectly. My problem is, I need to include error bars on my graph, but I am unsure of how the uncertainty would be converted when I did this with my graph..Any help would be greatly appreciated! Link to post Share on other sites More sharing options...
Sandwich Posted April 2, 2009 Report Share Posted April 2, 2009 (edited) I'm not 100% on this, but if you had your errors for X and you changed X to 1/X, should your error bars not be indicative of 1/Errors for X? O: Edited April 2, 2009 by sandwich Link to post Share on other sites More sharing options...
kristers Posted April 2, 2009 Author Report Share Posted April 2, 2009 I'm not 100% on this, but if you had your errors for X and you changed X to 1/X, should your error bars not be indicative of 1/Errors for X? O:no, that can't be right, my error is less than one so 1/error gives an error bigger than the actual quantity!maybe I have to do something with Percentage uncertainty? idk.. Link to post Share on other sites More sharing options...
Sandwich Posted April 2, 2009 Report Share Posted April 2, 2009 I mean error for x, rather than error. So you work out what your maximum possibly x value could be and your minimum possible x value, given the error, and then put 1 over those...? Then have them as your error bars.Just an idea, I'm not sure if it works Sorry if that's what you'd done already. Link to post Share on other sites More sharing options...
kristers Posted April 3, 2009 Author Report Share Posted April 3, 2009 Okay, I have another question for anyone,Is it even necessary for me to linearize data in a physics lab? Can I just perform a curve of best fit and leave it while still fulfilling DCP 2 and 3? Link to post Share on other sites More sharing options...
SharkSpider Posted April 3, 2009 Report Share Posted April 3, 2009 Yes, it's necessary to do so. Also, when using any multiplicative operation on error, convert the error to a percentage, then convert it back afterwards. Link to post Share on other sites More sharing options...
kristers Posted April 4, 2009 Author Report Share Posted April 4, 2009 Yes, it's necessary to do so. Also, when using any multiplicative operation on error, convert the error to a percentage, then convert it back afterwards.Okay Thanks a million!!!! =D Link to post Share on other sites More sharing options...
Recommended Posts