The most direct way to do this is to note that thelargestthe length and width could be is 17.6 cm and 9.5 cm, respectively. That means the largest possible area is (17.6)(9.5)= 167.2 sq. cm. Thesmallestpossible length and width will be 17.2 cm and 9.3 cm. That means the smallest possible area is (17.2)(9.3)= 159.96 sq. cm. (17.4)(9.4)= 163.56 which is 167.2- 163.56= 3.65 smaller than the largest possible value and 163.56- 159.96= 3.6 sq. cm. larger than the smallest value. So we are certain the true area is sq. cm. Rounding that "error" to one decimal place would give your "4" but I don't see why you would do that.

There is an engineer's "rule of thumb" that "when quantities are added, errors are add; when quantities are multiplied,relativeerrors are added". So you are correct that since the "relative errors" are (or 1.15%) and (or 1.06%), the relative error in there product is 1.15+ 1.06= 2.21% so the error would be (.0221)(163.56)= 3.62. I can see rounding that to 3.6 (the error I have, above) but not to "4".