 |
|
Webxaminer offers both 'static testing' and the recently more popular 'adaptive testing'.
Adaptive testing means that the computer automatically determines which
questions are presented based on responses to prior questions.
When questions are answered correctly, the estimate of ability is raised
and a more difficult question is presented and vice versa, giving the
potential to test a wide range of ability very concisely.
Static testing is the more traditional type of test containing a pre-configured question set, i.e. each student is asked exactly the same questions.
Adaptive Engines
We offer a range of adaptive engines as described below. Please feel free to contact us to discuss the solution that best suits your needs.
IRT-3 Engine.
The IRT-3 engine gives fast and precise results. It works by selecting
each next item based on the examinee's scored responses to all previous
items, according to the algorithms and mathematical rules of the three
parameter Item Response Theory model. Here the probability of a student
answering an item correctly is given by an expression that takes account
of its difficulty, discrimination and guess factor.
The IRT-3 engine is thus driven by a question pool containing question
items that have the above set of IRT related attributes - each item having
its own unique statistically evaluated set of values.
The classic IRT-3 engine has some pre-requisites and constraints.
The information now provided here will help you to choose the most
appropriate testing engine for your business.
IRT-3 Engine Pre-Requisites
- Question Pool
The IRT-3 engine runs against a set of questions of varying difficulty.
The engine continuously evaluates examinee ability and always asks the
most appropriate question in order to get the maximum information from
the 'item response'. It does this by selecting from the question pool the
question
which is closest in terms of difficulty to the 'theoretically required'
level of difficulty as evaluated by the engine.
The closer to the actual question difficulty, the more information gained
from the item response.
Thus the larger the question pool to draw upon, the more precise the
engine will be - simply because of the variety of actual question
difficulties.
It is recommended the question pool is several times larger than the
typical number questions actually to be asked in the test, together with
the widest possible variety in question difficulty.
- Item Pre-Calibration
Each test item must be pre-calibrated in order to evaluate the question
difficulty and its guess probability. It is only possible to accurately
calibrate an item (test question) from a
relatively large result set of previous results (even pen and paper
historical sources can be acceptable).
IRT-3 Engine Constraints
The IRT-3 engine always asks a minimum required set of questions to
evaluate examinee ability with given precision. The main goal of the
result is to get a fast and precise overall ability evaluation without any
in-depth analysis, for example sub-topic ability analysis etc.
For customers who either cannot meet the pre-requisites of the IRT-3
engine or who require sub-topic analysis we offer a range of alternative
engines.
Limited_Pool Adaptive Engine.
The Limited_Pool Adaptive engine is an alternative adaptive solution if
you do not have a question pool of sufficient size to meet the constraints
of the IRT-3 engine.
The Limited_Pool Adaptive engine will measure a candidate's ability with precise accuracy.
Typically suited to tests where:
- a thorough examination of the candidate's ability is required;
- a summary analysis of the candidate's performance is sufficient.
Your candidate's score in a Limited_Pool Adaptive exam is not analysed by sub-topic because potentially too few items are presented to the candidate to accurately identify strengths and weaknesses within individual sub-topics and here you may prefer our Adaptive Plus Engine or Placement Tester Engine.
Adaptive_Lite Engine.
The Adaptive_Lite engine is a faster, lighter version of our Limited_Pool
Adaptive engine. Here the candidate's ability will be determined in the
least number of questions possible.
Typically suited to tests where:
- a rough assessment of the candidate's overall ability is sufficient;
- an analysis of the candidate's overall performance is sufficient;
- the test time is required to be kept to an absolute minimum (e.g. a pre-interview 20..30 minute test).
Adaptive Plus.
This is a hybrid engine. Hybrid in that it contains a fixed number of questions whilst at the same time continually adapting the degree of difficulty of the questions presented to the candidate depending upon their previous answers. The advantage of an Adaptive Plus engine is that the candidate's sub-topic scores can be analysed, thus detecting the candidate's strong/weak areas.
Typically suited to tests where:
- the level of question difficulty needs to adapt whilst still analysing strengths and weaknesses within sub topics
- a precise assessment of the candidate's ability is required
Placement Tester.
An adaptive engine suited to student placement tests. The Placement tester will accurately assess a student's skill level to help determine which grade the student should be placed.
Typically suited to tests where:
- the test comprises a set of questions over several levels (e.g. grades);
- the candidate needs to achieve a pass score at one level in order to progress to the next level;
- the test needs to adapt to the next level (up or down) as soon as ability at the current level is determined;
- the test needs to stop immediately the skill level has been determined;
- the candidate's level needs to be accurately determined;
- an analysis of the candidate's performance at each level undertaken is required;
- an analysis of strengths and weaknesses by sub topic within level is required.
Classic Static Exams.
Alternatively we also offer any test to be built as a more traditional classic static test. Here the questions are fixed in number with the sequence and difficulty of all questions configured in advance. Many professional examination bodies still elect to set these classic static tests.
|
|
