Okay, so it's a common problem: given a set of measurments (in my case velocity and absolute magnitude) that each has a different error (different sources) i would like to find a linear regression to fit the data.

The final equation must have a separate error term and the reqgression fits must reflect this as well.

For example:

y

1.0±0.1
2.0±0.2
3.0±0.3

x

4.0±0.4
5.0±0.5
6.0±0.6

y=mx + c±error.

The question, then, is how do I do this in any of the available programs (hopefully MINITAB) ? I have scoured the internet for a solution but to no avail. I really need someone that has done this to help.

BTW, this is part of a project in an astronomy class relating logarithms of rotational velocities to the absolute magnitude (Tully-Fisher).