Category Archives: Computer Adaptive Testing

Photographs As Items for Assessment – Free Example

Photographs of people in activity is a promising newer area for development of business-relevant assessments that has been in use for years in healthcare.  Originally developed in the Netherlands to help patients suffering from fear of pain when moving the body (kinesiophobia), the University of Maastricht’s website has details on citations and free compressed (zipped file) short version of the main test.

Clearly, this same approach could be used to develop more engaging employee and organizational assessments that may be difficult to fake, have better face validity, and more workplace fidelity than other types of items.  Further, with cheap and even free video sites, video items could also play a bigger role in future assessments.

Consider these possible fruitful examples.

a) Vocational Interest Assessment

Vocational interest tests help people identify career paths for which their interests, values, and aptitudes are particularly suited.  But most all are purely text-based.  What if each career alternative had photographs of the tasks in each job or job family, with video vignettes of major tasks?  Perhaps this could be a fun way to assess what activities and careers would ultimately help the person realize their goals.  Take another look at the picture at the top of this article.  It’s an actual picture from PHODA’s assessment, but couldn’t it represent the task of lifting articles out of a trunk for the job of a taxicab driver?

b) Employee Selection

Cognitive and knowledge-based tests are often used to select new employees, but not nearly as often or instead of the ubiquitous job interview.  What if good instruments could be developed, perhaps with a combination of item types, to include pictures?  People could rate pictures like this one on the degree to which it looks similar to their desk – would you expect highly conscientious people to endorse this picture?

I would guess that highly conscientious and prudent people would be unlikely to indicate that this picture reflects their own office.  Sales Convention pictures would be good for the high-end of extraversion; Police taking down violent offenders for low levels of agreeableness.  The potential for pre-hire selection, especially using to add to Computer-Adaptive Testing item banks is tremendous.

c) Culture & Climate

Static pictures may be difficult to identify that reflect various organizational cultural differences, but videos could certainly be used to assess these. 

Limitations
As optimistic as I am about the potential for picture-based items to take a larger role in organizational assessment, I recognize there are also downsides.  First, while digital cameras are cheap, actors may not be.  If you can find existing workplaces where you can take these pictures, it may help you avoid hiring actors for static pictures, but perhaps not for videos that could really suffer with amateur actors.

Second, one New Zeland user of the PHODA complains that if the photographs are context-specific, they can loose value in other contexts.  I remember once when I worked for AT&T Microelectronics, we hired Wally Borman to redo his 1970’s era rater training videos because while the content was good, the actors wore sideburns, bell-bottoms and leisure suits.  This was never going to be very persuasive as “cutting edge” to managers in a bleeding-edge semiconductor factory (computer chips).

Do you see the same potential for photograph-based items as The Scientific Leader?

, , , , , , , , , , , , , , , , , , , , , ,

Advertisements

Australian CAT for Kids

The Australian State of Victoria’s standard educational assessments include computer-adaptive tests (CATs), reports their new, free manual on report interpretation.  I was pleased to discover that the Victorian Curriculum and Assessment Authority uses the most modern form of human assessment to help children of all ages learn.

In particular, it is noteworthy that their easy-to-read manual includes an understanding of Rasch Measurement.  It notes the specific locations where there are items that are out of scope for a given assessment.  In these places, the child is mismatched with the test – the questions are either too hard or too easy to produce a trustworthy metric.

I’m hopeful that Australia’s educational leadership rubs off on more schools around the world.

, , , , , , , , , , ,

CAT for Senior Citizens

Professor Jette

Can Computer-Adaptive Testing Help Senior Citizens?  Research from Boston University suggests it can.  Professor Alan Jette, Director of the Boston University Health & Disability Research Institute published a recent paper examining disability assessments in traditional and computer forms.  With data from 671 older adults residing in care facilities, CAT compared favorably with fixed-form scales, even for a version of the CAT with only 10 questions.  In his study, each CAT was administered in less than three minutes, and were highly correlated with the original instrument.

His research strongly suggests that in situations where time is a scarce resource, and measurement fidelity is still important, that Computer-Adaptive Measurement approaches are often more useful than others.

, , , , , , ,

Madoff Destroys $50 Billion with “Giant Ponzie Scheme”

Bernie Madoff is the latest in the series of senior executives to destroy value, this time with an apparent $50 billion dollar fraud, according to the Financial Times.  Madoff, a former Chairman of the NASDAQ stock market, on thursday admitted to his employees including his two sons that his operations were “all just one big lie” and “basically, a giant Ponzi scheme”.  The alleged fraud is the largest ever investor fraud ever blamed on a single individual.

Previously, I had written about the “Three Stooges of Operational Risk“, where I detailed senior executive destruction from Key Lay of Enron, Bernie Evers of Worldcom and most recently, Dick Fuld‘s follies with Lehman Brothers.  In two of those three I noted the dishonesty and fraud that accounted for their downfall similar to Madoff.  But unlike Madoff, they were less candid about thair fraud.  After Madoff’s brazen alleged admission, is there any uncertainty that leadership due dilligence is a critical part of the selection process of hiring senior executives?  Could it be any more clear that the pre-hire assessment procedure is a non-trivial subset of Enterprise Risk Management?

In fairness, these Industrial Organizational Psychology methods have their limitations.  No forecast could ever be perfect, or and even the best assessment procedures only account for 30-60% of the variance in job performance.  But it’s relatively rare that factors such as conscientiousness are used to screen executives – and conscientiousness highly predicts dishonest, and imprudent behavior in the workplace like that of Madoff.  With new methods from Rasch Measurement, Computer-Adaptive Testing, and an innovation from the Scientific Leader, “Inverted Computer Adaptive Testing” using Virtual Realtity, it’s increasingly difficult for people to fake or misrepresent themselves on these assessments. 

How much risk are you accepting when you use standard interviews to hire your employees?

, , , , , , , , , , , , ,

Electronic Health Information Systems Natural For Rasch Computer Adaptive Testing

I was pleased to discover a new instrument in the Journal, Psychiatry that uses Rasch Measurement to better assess Catatonia.  Catatnonia is a psychiatric condition that groups a number of pathologies ranging from bipolar disorder to drug abuse and schizophrenia.  The authors use a partial-credit Rasch approach that allows for ratings along a behaviorally-anchored rating scale. An example of one of the scale’s partial-credit items appears here:

This research did not benefit from the Electronic Health Record (EHR) or Electronic Health Information Systems that are universally starting to be used in healthcare.  It seems inevitable that as EHRs become ubiquitous in healthcare, that computer-adaptive Rasch scales like the author’s new KANNER scale will be come commonplace.  Whether for patient pain measurement and consequent medication, or psychopathology, or occupational therapy/science, Rasch Measurement has a long history of successful use in these areas – but without the benefit of seamless integration in a CAT in your health record.  I foresee CAT-based assessments being combined with Statistical Process Control (SPC) and Fuzzy Logic to better track patient conditions over time.  In hospital settings, I suspect this will happen first as patients with acute and potentially fatal ailments already are tracked continuously.  But non-acute patients, through wireless devices, could easily benefit from these approaches.  This is particularly true if medicine were to use one of the patents I authored while at Motorola, still pending, that combines computer-adaptive assessment and wireless devices.

This is a good example of the power of interdisciplinary innovations combining for something significantly better.  Rasch Measurement comes from Psychometrics; Statistical Process Control was developed by Industrial/Systems Engineers and Statisticians; Wireless technology comes from Computer Scientists; and the content comes from the various health sciences including medicine, nursing, physical therapy and occupational therapy.

Utah Leadership Supports Computer-Adaptive Testing In Spite of “No Bureaucrat Left Behind” Act

Nine schools in Utah have found the benefits of Computer-Adaptive Testing to trump older methods.  Adaptive tests change to match a student’s skill level, avoiding wasted time and effort on questions that are far below or above their proficiency level.  They’re also at least 20% shorter.  This allows for periodic reassessment, and personalized focus on the specific curricular areas a learner needs to work on.  Each student is treated as a special, unique person.

But the US Federal Government’s Department of Education is behind the times, and making it difficult for Utah to use the modern psychometric methods, according to Utah’s Daily Herald.  The “No Child Left Behind Act” requires outdated, non-adaptive methods to be used in addition to the modern approaches.  While on the surface, the DoE’s request for peer review is something that is good, in applied settings, it’s rarely used.  The instruments I’ve developed would certainly pass the scrutiny of my peers, and the feedback they give is useful.  But these extra steps are typically unnecessary to ensure that instrumentation is useful, as long as professionals develop the Computer-Adaptive Tests.  It’s downright destructive to children for the federal government to force Utah to use outdated, longer, and less precise measures of learning.  While I presume those favored by Washington are “peer reviewed”, I suspect that the review committee is selected by those who are friends of politicians, and are likely unskilled in the recent developments in computer-adaptive measurement.

Fortunately, Utah appears to have visionary, contemporary leadership about steadfastly supporting good measurement to help children learn.  The Utah Legislature, the State School Board and Governor all approved the plan to continue to use it – and the Feds require the outdated assessments to be used as well.  This is both a hassle, unnecessary cost, and an opportunity cost – the children could have been spending the time they’ll take on the DoE tests on learning something new.  Are you a visionary leader like the folks in Utah?  More by The Scientific Leader on Computer-Adaptive Measurement, applied to organizations and business is free here.

, , , , , , , , , , , ,

Pandora’s Box Worth Opening: Example of Future Rasch Measurement

I’m really amazed at the combination of the Music Genome Project and my iPhone.  Pandora Radio is a free service that allows you to stream customized music “stations” to your computer or your iPhone.  Today, I was able to stream it to my iPhone and pipe it into my car through the Auxiliary plug.  By selecting a specific song or genre, it is able to use its’ information on similar other songs and create an unlimited series of songs similar to the one you like.  Better still, you can have it mix across channels. And any new channels you create on either your iPhone or computer are automatically available on all your devices.

As each song plays, you’re able to give it a “thumbs up” or “thumbs down” to indicate whether or not you like the song.  Favorable and unfavorable ratings help the system adjust future selections to be in the same general domain as the original song, but will be more similar to those you liked.  These binary or dichotomous selections can be considered the most basic form of data useful for productive Rasch Measurement.  It’s an excellent example of why the Rasch model has so much potential in the Web 2.0 world where people are constantly sharing and rating things on Facebook, LinkedIn, Amazon.com and other social websites.

The current version is useful, but appears to be uninformed by psychometrics.  It is based on ratings of over 400 attributes to describe songs, and an engineering algorithm to organize them.  Each song is represented by a list of attributes.  Wikipedia’s entry describes relatively atheoretical approach used to parse the attributes into “genes”, and then recommend similar alternatives to create a “channel”.  The firm has several patents, both for the algorithm, playlist technology and interface on the technology.

J. "Mike" Linacre

John Michael Linacre

But apparently 10% of songs are rated by more than one “technician” to try to improve reliability.  From the description on Wikipedia, it appears the firm isn’t using it’s data to the fullest.  Pandora has over 2 million users, as of December 2, 2008.  First, even if there were only 500 or so ratings of each song, or genre, the users “thumbs up” and “thumbs down” would be excellent fodder to analyze with Rasch Measurement software that could further improve both song and listener measurement, on the same “ruler” of interest.  Second, whether technicians or listeners, some people are likely to be easy thumbsters and others hard.  This leniency and severity bias in ratings can be adjusted with Mike Linacre’s Many Facet Rasch Model (MFRM) method.  Linacre also happens to write the two most popular Rasch analysis programs, Winsteps and Facets.  Highly robust to missing data, this approach could really help Pandora improve assessment of both the songs, Genres, and listeners simultaneously – especially with a form of Computer-Adaptive Testing.

While Pandora Radio is a fantastic overall offerring today, my only other suggestion for improvement is to allow it to be used outside the USA.  When I was in India a few months ago, I learned the hard way that Pandora’s geolocation software could detect that my IP address was abroad, and refused to send me my normal stations, until I returned to the states.  I didn’t try US-based proxies to try to still connect, but hopefully, both Rasch Measurement and Global Licensing will be future features they’ll be adding in the future.

, , , , , , , , , , , , , , ,