This page describes human error rates in a variety of contexts.
Most of the error rates are for mechanical errors. A good general figure for mechanical error rates appears to be about 0.5%.
Of course the denominator differs across studies. However only fairly simple actions are used in the denominator.
The Klemmer and Snyder study shows that much lower error rates are possible--in this case for people whose job consisted almost entirely of data entry.
The error rate for more complex logic errors is about 5%, based primarily on data on other pages, especially the program development page.
|
Study |
Detail |
Error Rate |
|
Baddeley & Longman [1973] |
Entering mail codes. Errors after correction. Per mail code. |
0.5% |
|
Chedru & Geschwind [1972] |
Grammatical errors per word |
1.1% |
|
Dhillon [1986] |
Reading a gauge incorrectly. Per read. |
0.5% |
|
Dremen and Berry [1995] |
Percentage error in security analysts' earnings forecasts for reporting earnings. 1980 / 1985 / 1990. That is, size of error rather than frequency of error. |
30% |
| 52% | ||
| 65% | ||
|
Edmondson [1996] |
Errors per medication in hospital, based on data presented in the paper. Per dose. |
1.6% |
|
Grudin [1983] |
Error rate per keystroke for six expert typists. Told not to correct errors, although some did. Per keystroke. |
1% |
|
Hotopf [1980] |
S sample (speech errors). Per word |
0.2% |
|
Hotopf [1980] |
W sample (written exam). Per word |
0.9% |
|
Hotopf [1980] |
10 undergraduates write for 30 minutes, grammatical and spelling errors per word |
1.6% |
|
Klemmer [1962] |
Keypunch machine operators, errors per character |
0.02% to 0.06% |
|
Klemmer [1962] |
Bank machine operators, errors per check |
0.03% |
|
Kukich [1992] |
Nonword spelling errors in uses of telecommunication devices for the deaf. 40,000 words (strings). Per string. |
6% |
|
Mathias, MacKenzie & Buxton [1996] |
10 touch typists averaging 58 words per minute. No error correction. In last session. Per keystroke. |
4% |
|
Mattson & Baars [1992] |
Typing study with secretaries and clerks. Nonsense words. Per nonsense word. |
7.4% |
|
Melchers & Harrington [1982] |
Students performing calculator tasks and table lookup tasks. Per multipart calculation. Per table lookup. Etc. |
1%-2% |
|
Mitton [1987] |
Study of 170,016 errors in high-school essays, spelling errors. Per word. |
2.4% |
|
Potter [1995] |
Errors in making entries in an aircraft flight management system. Per keystroke. Higher if heavy workload. |
10.0% |
|
Rabbit [1990] |
Flash one of two letters on display screen. Subject hits one of two keys in response. After correction. Per choice. |
0.6% |
|
Schoonard & Boies [1975] |
Line-oriented text editor. Error rate per word. Without correction / with error correction. |
3.4% / 0.52% |
|
Shaffer & Hardwick [1968] |
Residual typing errors per character. Subjects with error rates higher than 2.5% were excluded. All qualified touch typists, including excluded. 20 subjects finally used. |
0.63% |
|
Swain & Guttman [1983] |
Interpreting indicator on an indicator lamp. Per interpretation. |
0.1% |
|
Swain & Guttman [1983] |
Error reading an analog meter. Per read. |
0.3% |
|
Swain & Guttman [1983] |
Choosing an incorrect panel control control from a number of similar controls. Per choice. |
0.3% |
|
Swain & Guttman [1983] |
Error reading chart recorder. Per read. |
0.6% |
|
Swain & Guttman [1983] |
Error reading a graph. Per read. |
1% |
|
Swain & Guttman [1983] |
Turning control in wrong direction under extreme stress. Per turn. |
50% |
|
Tsao [1990] |
Nonword spelling errors in uses of TDD (telecommunication devices for the deaf) 130,000 strings. Per string. |
5% |
|
Van Nes [1971] |
Hand-written text, grammatical errors. Per word assuming 5 letters per word |
0.3%-3.0% |
|
Wing & Baddeley [1980] |
Grammatical errors in examination at Cambridge. Per word. |
0.5% |