Uncertainties and LINEST in Excel....

[elevenb]

I don't think this qualifies as a homework question, as I can work around it; please move if necessary.

I am currently doing a lab report on the Magnetic Susceptibility of different samples.

In one part of this report I have to plot Force vs B^2, which is a linear fit. I have errors which have been propogated through on both the Force and B^2. I need a value for the gradient so I use the LINEST tool on Excel to do a linear regression and find a value for my slope. However the errors bars for Force and B^2 are not incorporated into this linear fit and the error on the slope from the linear regression seems to be separate from these. Is there any way to incorporate the errors from the values of force and b^2 into my linear regression fit?

 

[Jhero]

One way to incorporate the errors from the values of force and B^2 into your linear regression fit is to use a weighted linear regression. This approach takes into account the uncertainties associated with each data point and assigns a weight to each respective value. The weights are determined by the uncertainty associated with each data point, and the process of fitting the line then uses these weights rather than simply the raw data points. In Excel, you can access this weighted linear regression by going to the Data tab, click on Data Analysis, then choose Regression. Select the Y Range (Force) and X Range (B^2) and then check the box for 'Set Intercept'. Then, select the Weighted option and enter the Range of values that correspond to the errors associated with Force and B^2. Finally, click OK. This will generate the results of a weighted linear regression that incorporates the errors from the values of force and B^2 into your linear regression fit.