COMPASS - South Seattle Community College

Download Report

Transcript COMPASS - South Seattle Community College

COMPASS
National and Local Norming
Sandra Bolt, M.S., Director
Student Assessment Services
South Seattle Community College
February 2010
TABLE OF CONTENTS
• What Is COMPASS?
• Is Placement Testing Necessary?
• How Are Cut-scores Developed And
Evaluated?
• Answers To The Most Common Questions.
• Concluding Notes.
WHAT IS COMPASS
• It is a placement test. It is curriculum based with a diagnostic component.
• COMPASS ensures the examinee is ready for the rigor of the course. The
cut-scores signal that the course will not be too hard or too easy. It is used
for post-secondary level placement.
• It is not an entrance exam; it is not an aptitude test. COMPASS does not
have the criteria required by EEOC to be a workplace test instrument.
• The instrument is correlated with the institution’s academic courses.
Reliability is maintained by correlating test scores to course grade.
• Reliability:
Grammar .85; ESL-Reading .86; Listening .85
Writing .85; Reading .85; Math .85 - .86
(The above national scores are based on the standard test length. Reliability
increases with the extended or maximum test lengths. South’s test is
based on the standard length. We can adjust the length with data.)
WHAT IS COMPASS
• The instrument is progressive.
– The ESL COMPASS is linked to the Standard COMPASS.
With the appropriate skill set, for example, an ESL
student can start in ESL COMPASS and test into
Standard COMPASS and into ENGL 101.
– Examinees are given six questions. If answered
correctly, the system presents easier or harder
questions within a domain. Examples of domains are:
pre-algebra, algebra, college algebra and
trigonometry. With data, we set the routing scale,
what is tested, and the diagnostic measures.
IS A PLACEMENT TEST NECESSARY?
• When the COMPASS test is correlated to the
course, the instructor is able to teach students
who are ready for the course content at the
level of the course outline and course
description.
INSTRUCTOR COMMENTS FROM AN
UNDERUTILIZED COMPASS TEST
• “I don’t use COMPASS, it doesn’t work for my class.”
• “I review the whole chapter in class; I find it necessary
to teach to the test.”
• “I rarely use supplemental material; my time is spent
reviewing, over-and-over, the basic text material.”
• “I was hired to teach in my field; I’m doing less of that
and more instruction in English and math skills.”
• “Over the years, I have had to ‘dumb down’ my
course.”
HOW ARE CUT SCORES DEVELOPED
AND EVALUATED?
• Cut scores are developed in two ways:
– The Normed Reference Process
– The Criterion Reference Process
HOW ARE CUT-SCORES
DEVELOPED AND EVALUATED?
• Each measurement helps to identify:
Success rate (% above median who get a “B” or better.)
When 85% of the class receives a “C” or better, “B”
is used.
Accuracy rate (sum of true positives & true negatives).
True positives: those that scored above the median
and passed with a “B”.
True negatives: those that scored below the
median and failed course.
Percent placed into lower classes
HOW ARE CUT SCORES DEVELOPED
AND EVALUATED?
• The Normed Reference Process
– All examines are tested. Scatter graphs of test scores
are sent with course descriptions to ACT. The data are
analyzed against national data and returned with
preliminary cut-scores.
• ACT reviews, for example, 100 colleges that offer the same
English courses and test for the median. The median is the
minimum score for which we estimate that a student has
50% chance of earning a ‘B’ or better in the course.
• The percent of correct or incorrect placement can be based
on the course grade of each cut score above the median.
HOW ARE CUT-SCORES DEVELOPED
AND EVALUATED?
• The Criterion Referenced Process
– Instructors review the COMPASS test categories and determine what
percent of the battery a student should answer correctly based on the
needs of their class. Answers to this process are added to the
national data we receive.
– Accuracy of placement is conducted by assigning each student a
rating based on ACT’s ‘course readiness’ definition. The rating is
conducted within the first week of class. This score is then matched
to the course final grade. The cut-score is determined by those
receiving a ‘B’ grade and a rating of 3.0+ in the course.
HOW ARE CUT SCORES DEVELOPED
AND EVALUATED?
• The Student Assessment Office reviews the
national and local data with faculty. The initial
scores are set.
• The process must not end at that point. For
cut scores to be reliable, scores must be
measured against grades every few years. This
is not difficult or time consuming process. It is
necessary.
QUESTIONS TYPICALLY ASKED
• “When cuts are normed, is it purely South’s data or is data
comprised from all three campuses?”
– All three campuses must show a rationale for placement. The test
developer, ACT, details the procedure. Some institutions choose other
methods.
– District aligned scores have only been set for college level English at
this time. Developmental courses will not likely be set district wide.
• “Do we see different levels of classroom success among different
groups of students?”
– I do not have data on that question. If, however, students are not
placed into the courses by established cut scores, simple reliability
studies are not conducted, and teaching and grading are not
consistent, then it is likely that the instructor must teach outside the
course outline.
CONCLUDING NOTES: South’s Story
• English and math were normed in 1989 for
ASSET. When COMPASS was established in
1995, we correlated scores and faculty
conducted reliability studies. Some academic
and professional technical courses went
through the 1989 rigor to determine cuts.
CONCLUDING NOTES
• Proper prerequisites are very important in
sequential courses. Scores need to be upheld.
Instruction and grades need to be consistent.
• The institution is receiving dollars based on
the State’s criteria of student retention and
completion. COMPASS can ensure course
readiness, the foundation for success and
retention, but grades and program entry need
to be no less reliable than the COMPASS
scores.