Error Analysis Help: Line of Best Fit with Error Bars

[$id]

Homework Statement

I am trying to work out if there is an elegant way to do a line of best fit(straight) through a set of points (3) while taking into consideration their error bars? Clearly I'm looking for a piece of software that can do this. Obviously excel, etc can do a standard straight line + equation and add errors later on but the fit will be different if errors are considered in the place on each point individually.

The reason I'm doing this is because I want to know the error in the gradient, taking account of the fact that each point has an error ( every point has different errors)

The Attempt at a Solution

The box method noted here will sort of work but it is quite tedious

http://wwwchem.uwimona.edu.jm:1104/lab_manuals/c10appendix7.html

But I'm looking for alternative ways of doing this. I only have 3 points of data as well.

Regards,
sid

 

[tyco05]

You only have three points of data?!??!

There's your first problem.

In the past I have used an error box method, but not the same as the one on the linked website.

You have x error bars and y error bars - these can be made into error boxes around each data point.

The line of maximum gradient then goes from the bottom right hand corner of your smallest x value to the top left corner of your highest x value.

The line of minimum gradient goes from the top left of the smallest x error box to the bottom right of the highest x error box.

Make sure that these lines go through ALL error boxes - if they don't, you have to make some adjustments.

The points just mentioned and gradients are pretty easy to spit out of excel.

Good luck.

 

[$id]

I can't take any more points

While i understand the box method, I was interested in a some sort of program being able to do the hard work for me lol

sid