Percent Error vs Percent Difference Explained
Learn when to use percent error against a known value and when percent difference between two measurements is the honest choice, with worked math.
Percent error: measuring against a known value
Percent error answers a specific question: how close did my measurement come to a value I already accept as correct? That accepted value might be a textbook constant, a certified reference standard, or a manufacturer's specification. Because one side of the comparison is trusted, the formula divides by the accepted value alone rather than an average. The result tells you the size of your measurement error relative to the truth, which is exactly what a physics or chemistry lab report is usually asking for.
Percent difference: comparing two peers
Sometimes there is no accepted answer, only two measurements you want to reconcile: two instruments, two lab groups, or the same sample run twice. Percent difference treats both values as equally valid by dividing the gap between them by their average magnitude. That makes the calculation symmetric, so it does not matter which value you call A and which you call B. It measures disagreement between peers rather than error against a reference, and reporting it as percent error would falsely imply one value is correct.
A worked comparison
Say two thermometers read 100 and 120. As a percent difference the calculation is |100 - 120| divided by the average of 100 and 120, which is 110, giving 20 / 110 x 100, or 18.1818%. Now suppose 100 was actually the calibrated reference: the percent error of the 120 reading is |120 - 100| / 100 x 100, or 20%. The two numbers differ because percent error divides by 100 while percent difference divides by the larger average of 110, a reminder that the denominator choice is not cosmetic.
Reporting error cleanly in a lab
State which quantity you treated as the accepted value and cite where it came from, so a reader can judge the comparison. Keep a sensible number of significant figures; this tool shows up to six and trims trailing zeros, but your reported precision should match your instrument. A percent error near zero means your method and equipment agree with the reference, while a large one points you toward calibration drift, systematic bias, or a procedural mistake worth investigating before you rerun the experiment.