Adaptive Assessment Technology

Aspiring Minds' proprietary assessment platform provides high-end statistical assessment based on the Item Response Theory (IRT). IRT is a global standard used for construction of high-stake tests such as university admission exams, medical and accountancy certifications, etc.

Aspiring Minds’ platforms brings for the first time in India, Computer Adaptive Tests with dynamic content providing accurate skill measurement. All our assessments are powered by our robust Adaptive Testing engine which does a multi-dimensional assessment of the candidate abilities using advanced statistical and AI algorithms. The assessment platform is capable of automated content (phase in and out of questions) and test delivery management ensuring complete end-to-end integration of assessments and content. The test reliability and precision, its content validity, leak-proof and guess check algorithms ensure that assessments are comparable and precise.

Standardization and Performance

Comparability is the key to credible and useful assessment. Long term standardization allows the assessment results and its trends to be used over time to link long term outcomes to the results of the initial assessment. This sheer capability of being able to link assessment results to performance of the selected group be it for education, training or job goes a long way in achieving desired benefits.

Standardization maintains the scores of the assessment to be comparable for years making sure that the assessment continues to throw up similar scores for similar ability people over time.

Aspiring Minds' assessment engine is designed to provide year over year standardization for multiple years making sure learning and feedback from past experience can be translated to improvements in the future.

Aspiring Minds’ proof of delivery lies in demonstrating improving candidate intake year after year and by correctly predicting performance of each incoming candidate.

After an initial correlation in test construction phase, year-after-year AMAF allows various assessment parameters to be correlated with student performance to build regression and prediction models, These models enable fine-tuning the assessment instrument for better filtering and selection of candidates.

Reliability and Validity

The best test to measure ability to juggle balls is useless if you want to find the ability of the person to walk on a rope. Validity implies that the instrument measures the precise dimension for which the assessment is being performed. Aspiring Minds’ content planning coupled with several parallel-content balancing techniques maintain test validity despite test-form variation. Long-term correlation and prediction studies fine tune test and preserve validity for your intake requirements.

Reliability refers to the ability of the instrument to consistently do correct measurement. In other words, if a candidate gives the test multiple times, he/she should get the same score each time even though he/she may see different questions.

Using multiple large-scale statistical simulation studies, Aspiring Minds technology adjusts test length and uses effective question-delivery algorithm to keep variance under internally defined bounds. The test is robust and gives accurate ability measurement at each delivery instance.

Unbiased and Test Integrity

Assessment always faces the challenge of leaks and bias. While leaks to some extent are inevitable (some questions are always retained in memory and discussed), your Assessment results don't need to get affected by prior knowledge.

Aspiring Minds Assessment Framework ensures that effects like prior knowledge of a few questions, rote learning and guesses do not bias the assessment results. Test construction ensures correct ability measurement even if the candidate knows 20% of the questions. Guesses are automatically taken care of by using advanced statistical models for questions.

The technology works on a large and constantly evolving question bank to keep content diverse and un-masterable. Our Exposure control and test-overlap control techniques eliminate the same question being seen by more than a given threshold of candidates. Over-exposed and non-conforming questions are regularly weeded out from the test.

Precision

Complete Ability Spread

Meaningful assessment requires proper differentiation at all levels of ability. Assessment is required to classify the candidates into buckets of ability levels – granular enough to be able to use the assessment results for taking hiring and training decisions that are accurate and result oriented.

Traditional assessment is ridden with the task of putting together the right set of questions depending on the audience. While we lose any differentiation at lower ability levels by making a tough test, the reverse happens when we put together a simpler test.

Aspiring Minds Assessment Technology's Adaptive Framework ensures that every candidate is tested with precision at his or her ability level ensuring hair thin precision over the entire ability spread. Unlike traditional tests, our assessment is neither too simple nor too hard for any candidate.

Comparability

To make sure candidates don't have prior knowledge of questions, a large question bank is a must. But it is not enough! Candidates seeing different questions should be comparable.

Each question needs to be statistically rated, such that one can estimate the performance of a candidate on a benchmark question set, given his performance on the question set he/she saw.

Aspiring Minds has a very large question bank statistically rated on Indian candidates, the only such existent pool. This key ingredient that makes Aspiring Minds assessment comparable over years is continuously being expanded.

Any where any time testing

The Any where any time testing module has been designed to assess the knowledge of the candidate about the Any where any time testing environment, shell structure and programming, commands, file system and organization. This module can be used to assess candidates seeking "Administrator" profile in IT companies.

Assessment as a task can be fairly simple if we can collect everyone who needs to be assessed around the world on a single day, at a single time and conduct a test. Such testing by the very restrictions of availability we put on it becomes comparable and ensures some level of integrity.

Anywhere any time testing allows reliable and valid assessment to be delivered anywhere in the world, at a time convenient to the candidates.

Anywhere anytime assessment gives the assessor the capability to exponentially increase the reach to a large number of people across regions still keeping the logistics and infrastructure requirements for delivery in a controlled scale. For example, a computer based test of 3 hours could be delivered to 10,000 candidates in an year with just 10 dedicated computers.

Anywhere anytime testing also relieves the candidates’ anxiety quotient making their entire testing experience a lot more un-nerving.