Assess or Guess?
10 ways you might be fooling yourself about assessment
David Singer is a visiting
scholar at MIT, in the Department
of Brain and Cognitive
Science. He is a specialist in
assessment related to the
science in education.
to his visiting
post at MIT, he
was the director of Education
of the former American Nobel
Committee. Singer received
his doctorate in education
from Boston University (MA).
Here, the scholar and assessment
authority takes a hard
look at how we could be fooling
ourselves, even with our
best attempts at assessment.
EDITOR’S NOTE: At the
Campus Technology 2006 conference in Boston
(July 31-August 3), Singer
will moderate a panel from
MIT about the power of
assessment, and its proper
Want to be considered for Campus Technology’s Top 10? Send your countdown and a brief background/bio summary to firstname.lastname@example.org
Surveys do not accurately describe students’ real behaviors, attitudes,
or what they have learned.
- Validity—and especially the reliability—of student surveys are often low.
Approximately 10 percent of research on the effect of education
technologies focuses on student learning or performance.
- 90 percent of such research bases its conclusions on student opinion surveys.
Assessment usually only deals with the mean change in student
attitude or performance.
- Assessments often assume that the shift in distribution is uniform for all students
when, in fact, students who perform worse than or better than the mean
may be differentially affected by any new educational strategy.
Once the methodologies and technologies are in place, it gets harder
to apply them consistently.
- It is a common problem that instructors fall back on old habits.
Even if the number of students in a research effort is large enough, the
results cannot necessarily be generalized to all populations of students.
- A study of 1,000 students in Biloxi, Mississippi may have little relevance to
1,000 students in Bangor, Alaska or Beverly Hills, California.
Student motivation to try a new educational methodology can
decrease the validity of the findings.
- The so-called “novelty effect” increases student motivation and d'es not address
the real research questions regarding the pedagogy of the educational methods.
Instructors’ attentiveness and interest in research in education often
detracts from the validity of results.
- An instructor’s enthusiasm to prove the benefits of a new approach to teaching
can, in itself, improve students’ motivation, interest, and therefore performance.
There is a not a generally well-accepted definition of what constitutes
learning in education.
- The basis for what a student must do to receive a given grade varies dramatically
across schools and campuses.
Collaborative, interactive, and hands-on learning methods are not
effective for all students and may be counterproductive for some.
- Although these are popular and very effective methods for many students and
some courses of study, some students may not benefit due to different learning
and personality styles. Certain students may, in fact, do worse because of them.
Improvements in learning often cannot be attributed
to the methodologies of technologies that are being investigated.
- Research in education has historically been very poor because of the cost,
difficulty, or because of investigators not knowing how to set