Percentage uncertainty for variable measurement but fixed absolute uncertainty

Whilst calculating the Young's Modulus for a wire, I have decided on an absolute uncertainty for the extension to be ±1 mm (resolution of the ruler). But as the extension of the wire is a variable measurement over the range of force chosen (shown in the table below) the percentage uncertainty varies (also calculated in the table below).

I originally thought the overall percentage uncertainty should be chosen from the greatest percentage uncertainty of 50%, but this feels wrong because it does not take the other percentage uncertainties into account and would give me an overall percentage uncertainty for the experiment of over 60%. (The calculated value is out by 40% from the accepted value for context).

So my question is: Should I be choosing the overall uncertainty as described above or is there another method which would give me a percentage uncertainty in the extension measurement which takes all values into account?

$$\begin{array} {|c|c|} \hline \text{extension}/m(\pm0.001m) & \%\text{uncertainty} \\ \hline 0.002 & 50 \\ \hline 0.004 & 25 \\ \hline 0.004 & 25 \\ \hline 0.005 & 20 \\ \hline 0.005 & 20 \\ \hline 0.005 & 20 \\ \hline 0.006 & 17 \\ \hline 0.007 & 14 \\ \hline 0.007 & 14 \\ \hline \end{array}$$