What does error percentage mean?

I think it’s the deviation of your data from the given data. A high error percentage would then mean that your data are very different from the given data, a low percentage means they resemble them very closely.

But I’m not very knowledgeable about the intricacies of how mathematics and statistics work, so don’t take my word for it.

Error percent = (estimated – actual)/ actual * 100%

@Fyrius You’re totally right. Error percentage is a measure of how different the two pieces of data are. The higher the percentage, the worse the equation, essentially. Usually, >3% would be good, but since this seems like it’s for school, you can probably claim a much higher percentage as being okay.

@BhacSsylan
Yay! :D
It’s nice to be right about something occasionally.

In quantitative research, the error is the probability that the actual value lies outside the given range. If you took a sample of 100 people and took their average height, you could assume their average height is the same as that of the population (if you sampled correctly). This assumption would be correct to a degree of accuracy. The standard deviation of your statistics gives you a clue. For a standard bell curve, 96% of the population lies within two standard deviations. You can then state in a published paper that the average population height is “x” +/- 2SDs, with p<0.05. 0.05 is your error value. 0.05, or 5%, is usually the accepted value in medical research.

NB. This only refers to type 1 errors. For type 2 errors the calculation method is different, and the accepted margin of error is 20%.

Giving an absolute number such as “3 percent error is ok” or “p=0.05 is ok” or whatever for a general case does not sound reasonable to me. How much is ok depends on 1) what data you have, and 2) what you are gonna use the results for.

@Janka This is true, but that’s why i said “more is probably fine”. In general, for biochemistry and apparently medical research, ~<3–5% is considered a reliable result. It does depend heavily on the purpose and data set, but it seems we’re both coming from research backgrounds, where it is more of a set value.

That said, I also mentioned that as this seems to be for school, a much higher range is applicable here.

@BhacSsylan I come from a research background where a set value is often defined, too. That does not mean I consider it a good or a sensible practice.

But as this is for school, simply demonstrating you are able to calculate said range might be enough, regardless of the result (depending on the assignment). :)