Chapter 7 Notes

Just the facts!

Standardized tests

n Same

n Administration

n Scoring

n Interpretation

n Generalizable across different classrooms

Do teachers use standardized tests?

n 1995 study of Ark. Elem. Tchrs.

n 77.5% - Yes

Norm vs. Criterion Referenced

n Norm – compare to another group

n Criterion – pre-set standards

n One not better than the other—different purposes

Study: How are norm referenced useful?

n Diagnostics – 28%

n Comparisons – 20%

n NOT USEFUL – 14%

n Drawbacks?

n Invalid measure – 22%

n Too much on one test -12%

Study: How are criterion referenced useful?

n Diagnostics – 37% (28)

n NOT USEFUL – 15% (14)

n Drawbacks?

n Invalid measure – 6% (22)

n Too much on 1 test – 6% (12)

Study: What standard scores do you use?

n Z-scores – 12%

n T-scores – 11%

n Stanine – 54%

n Percentile – 83%

n Grade-equivalency – 80%

n Standard deviation – 51%

Why do we use standardized tests?

n To compare students & schools

n Development – progress over time

n Rate schools – quality indicator

n Publish in the newspaper

n Identify student strengths and weaknesses

n WE HAVE TOO!

What is the difference commercial and state mandated tests?

n Norm – comparison of a pupil (group) to a larger general group

Vs.

n Criterion – minimum performance, competency tests (rate schools, teachers, students)

Role of classroom teacher

n Very important

n Must have standard administration

n Teacher’s attitude will affect the students

Terms

n Item difficulty - % who got it right

n Usually about 50%

n Item discrimination – compare those who did well against those who did poorly

n High as possible

n Norms

n Comparison group (fixed)

n Representative group

n Criteria-Size, Representativeness, Recentness, Procedures

n Local norms – comparison group more like your school

n Raw score

n Actual number right

n Not useful – # of questions

n Percentile rank

n Percent of the NORM group that this student scored higher than

n NOT than took the test

n Confused with % grades

n Stanines – standard nines

n Divided into 9 areas

n 4, 5, & 6 are average

n Less misinterpretation

n Grade equivalent score

n Based on year and month

n Compare to norm group

n Most misintermpreted

n NCE – normal curve equivalents

n Spreads out the middle

n Look at high an low, moves to the middle

n Percentile bands

n Accounts for error

n More reliable

n 95% sure of the real score

Interpreting scores

n Main problem is over interpretation

n Use to show

n strengths and weaknesses

n Trends over time (year to year)

n Use other indicators before making a decision

Should teachers teach to the test?

n No but—

n Greater the stakes - more likely

n It does take away from planned curriculum

n Best preparation is good teaching/learning and a good curriculum

Validity

n Coverage of objectives

n Not covered in class?

n Not part of curriculum

n Norm group

n Test administration

n Interpretation

State-Mandated

n ACTAAP - Arkansas Comprehensive Testing, Assessment and Accountability Program

n Varies from state to state

n Not new – Canada

n Cut scores – minimum performance