Triadic Enneagram Test Statistics

Statistics are taken for curiosity's sake and for improving the test. Records of old trials are kept for the sake of measuring progress.

Current Statistics

Here are the statistics on the current trial of the Triadic Enneagram Test. These statistics are on people who have reported their type while taking the test. They are divided into four groups, based on how certain people claim to be about their type.

If you haven't yet taken the test, please finish it before viewing the current statistics. These will reveal what each question is measuring for, which may influence how you answer the questions.

Old Triadic Enneagram Test Statistics

General Accuracy is the percentage of questions people answered according to the answers intended for their reported type. Pure chance should have a general accuracy of 33.33%. Bullseye Accuracy is the percentage of people who scored as their reported type. Pure chance should have a bullseye accuracy of 11.11%. Since these measures could be contaminated by the preponderance of one type over another, the Weighted measures avoid this kind of contamination by counting all types equally. No matter how many members of any type answer questions, all members of that type count equally with all members of any other type. Because of this, the weighted measures of accuracy should be regarded as the most accurate.

Trial # Began Ended General Accuracy Bullseye Accuracy Sample Link Weighted General Accuracy Weighted Bullseye Accuracy
4 Jan. 9, 2001 Jan. 12, 2001 50.65% 50.65% 75 stats4.html
5 Jan. 12, 2001 Jan. 19, 2001 50.51% 54.08% 98 stats5.html
6 Jan. 19, 2001 Jan. 30, 2001 48.58% 44.79% 163 stats6.html
7 Jan. 30, 2001 Feb. 4, 2001 47.94% 41.38% 116 stats7.html
8 Feb. 4, 2001 Feb. 12, 2001 47.71% 43.12% 109 stats8.html
9 Feb. 12, 2001 Feb. 20, 2001 48.77% 45.38% 119 stats9.html
Bug fix: Scores from 2s are now recorded
10 Feb. 20, 2001 Mar. 3, 2001 45.39% 37.85% 214 stats10.html
11 Mar. 3, 2001 Apr. 12, 2001 46.65% 41.50% 682 stats11.html 43.23% 32.85%
12 Apr. 12, 2001 Oct. 6, 2001 45.76% 40.94% 171 stats12.html 42.76% 33.36%
13 Oct. 6, 2001 Feb. 4, 2002 45.82% 42.94% 1495 stats13.html 43.02% 34.35%
14 Feb. 4, 2002 Feb. 18, 2002 48.56% 40.93% 193 stats14.html 46.64% 35.06%

Beginning with trial #15, data is taken on how certain each person can be about what his/her type is. This data is used to break the statistics into four reports. Each report is on everyone who reported their type with at least a specified minimum degree of certainty. The first report is on everyone, and the other reports are on smaller and smaller samples.

Trial #15
Began Feb. 18, 2002 Ended Mar. 13, 2002
Minimum Certainty Sample Size General Accuracy Bullseye Accuracy Weighted General Accuracy Weighted Bullseye Accuracy Link
Flimsy 353 47.80% 44.19% 45.23% 37.76% stats15-0.html
Weak 310 49.78% 45.48% 46.81% 39.24% stats15-1.html
Moderate 234 50.69% 45.73% 47.22% 37.94% stats15-2.html
Strong 141 51.27% 44.68% 48.65% 39.64% stats15-3.html
Trial #16
Began Mar. 13, 2002 Ended Apr. 2, 2002
Minimum Certainty Sample Size General Accuracy Bullseye Accuracy Weighted General Accuracy Weighted Bullseye Accuracy Link
Flimsy 370 47.95% 43.24% 44.66% 33.99% stats16-0.html
Weak 321 49.15% 44.86% 45.69% 35.12% stats16-1.html
Moderate 244 50.47% 47.54% 46.92% 38.31% stats16-2.html
Strong 130 51.15% 52.31% 46.27% 40.66% stats16-3.html

Toward the end of trial 16, I updated the statistics page to reveal information on which set of triads each question is a best match for measuring. It is based on how many types have their top scores in the answers that fit a pattern for a particular triad. This information is useful for extracting useful information from questions that were a bad fit for what they were measuring. It can be used to find a better fit for a bad question, as well as to get ideas for new questions. In replacing some of the questions from trial 16 for trial 17, I eliminated questions whose best match was not what was being measured. This was different than my usual method of focusing on accuracy percentages, and it resulted in a slightly different selection of questions to replace. This resulted in dislodging some questions that had remained for a relatively long time. I also used best fit information from bad questions to write some of the new questions.

Duniho Enneagram Test Statistics

Trial # Began Ended Accuracy Sample Link
1 Dec. 29, 2000 Jan. 1, 2001 56.41% 26 stats1.html
2 Jan. 1, 2001 Jan. 6, 2001 41.27% 109 stast2.html
3 Jan. 5, 2001 Jan. 9, 2001 49.24% 51 stats3.html

Life of Questions

This table shows how long individual questions last.

12345678910111213141516171819202122232425262728
41111111111111111111111111111
51121212122212211112112211111
61221222232323311112213212121
71221223242424412112314322121
81231223352524513113415423121
91331233462635613113416524121
101431243573646613113516625121
111541253674746613124516645121
121641363684847713134517656121
131642363794958813144518666121
14164336381051069813154519666221
1517433639115117108141655110666221
16184436410125128108141756111666221
17194546411135139118141856111676321