|
This site is being built out
of pure necessity. The most popular personality
tests on the Internet aren't accurate enough.
The most common reason
that people don't believe in personality patterns is that they've taken
a flawed test, were given inaccurate results, and decided
that psychological personality typing was no more than random, vague,
all-encompassing statements that didn't describe them well. We're currently evaluating
questions to solve that problem. This site
has several improvements over the past personality type indicators which
dramatically increase its accuracy. Here are just a few of the improvements:
Psychologically profiled control group matched self-perceptions with actual types:
We took hundreds of people who knew little or nothing about personality typing, profiled them
individually, then had them each answer hundreds of questions so we could match their
self perceptions with their actual personality types
Variable weighting based on predictive ability:
Unlike other tests
which assume that all questions have the exact same predictive ability, we measured
the predictive ability statistically to identify:
- The numerical relative predictive ability of each question
- Co-predictivity, we isolated which questions were actually measuring the same aspect of a trait
User question selection: Other
tests confound the results by pigeon-holing people into one choice or another when neither
or both might apply or when the user might not understand the question. On this test the user ranks each question
on a scale giving confusing or inapplicable questions less influence on the results than
questions which strongly apply to them.
Better questions: Everyone in this field knows people who answer the questions based on who they want
to be instead of who they actually are. By using a profiled control group we were able to find questions
which are more likely to fool these 'professional test takers' into producing an honest result. Basically,
if people chose an answer which didn't actually apply then the question was statistically deleted in favor
of ones where their answers matched their actual personality.
25,000 Alpha Testers: We owe a tremendous debt to the alpha testers who are taking test after test after test of
questions to verify the statistical distribution
for each question and verify the separation ability of the final models. This test is still not
in its final form, if you take this test today you will further increase the accuracy of this model.
We respect the contributions of our predecessors to the science of personality typing.
Costa and McCrae's Five Factor Model, The Keirsey Temperment Sorter, the Myers-Briggs Type Indicator and others all
contributed to public understanding of personality typing and brought it into social acceptance in the mainstream.
This test is the
next step in that effort, taking advantage of current computational abilities such as data
mining, high-speed computations and the Internet. We're reinventing personality testing in an
era of better technology. Once people's types can be accurately determined there are thousands of applications,
from job choice to predicting which movies each person will enjoy, which will enhance people's lives in the future.
The first application of this test (after giving people accurate profiles) will be
to help people find romantic chemistry.
We've studied dozens of true-love couples and a phenomenally promising pattern of traits in common and different
has been found.
The application of that pattern to the general population will only be possible if people can be accurately typed.
For more info on this project read the project history and description.
Thank you for taking a few minutes to contribute your opinions to further the science of personality typing.
| |
