Global fitting works by minimizing the sum (for all data sets) of the sum (for all data points) of squares of distances of the data points from the curve. This approach only makes sense when all the data are expressed in the same units.

If different data sets are expressed in different units, be very cautious about using global fitting. The problem is that your decision about which units to use can change the results.

For example, imagine what happens if you change one data set from expressing weight in grams to expressing weight in milligrams. All the values are now increased by a factor of one thousand, and the sum-of-squares for that data set is increased by a factor of one million (one thousand squared). Compared to other data sets, expressed in different units, this data set now has a much greater impact on the fit. If you really need to do global fit to data sets using different units, consider first normalizing the data so they are comparable.