p=q²/r=5.46²/17.74=1.6805 to 5 sig figs.
Now we can replace q by 5.455 and 5.465, and p=17.735 and 17.745.
These produce the following values for p:
q=5.455 and r=17.735: p=1.67787
q=5.455 and r=17.745: p=1.67692
q=5.465 and r=17.735: p=1.68403
q=5.465 and r=17.745: p=1.68308
This gives us a range for p of {1.67692,1.68403}. The centre of this range is 1.680475.
The error is ±0.003555. To summarise this in a suitable degree of accuracy we should limit the accuracy to 2 decimal places: p=1.68, because this is the result of rounding the upper and lower limits of the range.