Performance Capabilities
The
Adaptiqs platform is designed with performance in mind. The technology
is designed to be more than just a static survey engine. It continuously
adapts to each individual respondent based on the information they provide.
The ACORN instrument was recently used in a test
to determine the capabilities of the technology as a viable method for
survey development. ACORN was used because it represents one of the
largest and most complex surveys used today. ACORN was developed from
questions in the O*NET framework (version 98). It contains 447 questions
designed to obtain information about respondent's occupation according
to five areas: the knowledge, skills, and abilities required for performance;
the generalized work activities (GWAs or tasks performed); and work
context (or environment in which it is performed). The five areas can
be answered using either a 5 or 7 point scale related to the importance,
frequency, and level that is required in the respondent's job. The dataset
from this survey contains information for 1197 different occupations.
ACORN uses this information to create a probabilistic
network that predicts how individuals in certain occupations
tend to respond to each question. When respondents are asked about their
job, the network continuously updates a list of jobs that most likely
match their responses.
The results of the test provide an encouraging outlook for the use of
the technology in survey administration in a number of specific applications
in medicine, human resources, marketing, and product development.*
ACORN is capable of achieving remarkable efficiency
using adaptive administration. Only a fraction of the original survey
length is required to reach a decision. Most often as few as 10% to 15% of the items in the
survey are necessary to reach the same results. Preliminary beta-tests of the prototype show that in as few
as 15 to 20 questions, our technology can determine a person's job and identify
other knowledge, skills, and abilities used on the job.
How does it determine which questions to ask? In
short, the moment a person submits a response to a question, ACORN performs
3,303,720 calculations to determine the next best question. It uses
the concept of entropy to ask "based on what I know now, what piece
of information would provide the most information about this person's
job?" Visit our Artificial Intelligence Tutorial
to learn more.
