Data meeting

Using Basic Statistics
to Look at Your Data

{

Data Coach Meeting: January 2014

Mean or Median?
If you want a “helicopter view” of how
your class did on a test:
• Use “median” in eschool if you are
looking at less than 30 data points.
• Use “mean” in eschool if you have 30
data points or more.

Using the Mean If You
Have More Than 1 Class
You can combine means using this
simple formula:
(Class 1 Mean) (# of class 1 students) + (Class 2 Mean)(# of class 2 students)
(# of students in both classes)

Fact: All Tests Have Errors
• When you give your test, eschool will
give you the standard deviation.
• Most teacher-made tests have a
reliability coefficient of .50-.60, so
assume .55.

• Only well-tested tests (WISC, WJ-III)
have upward of .90 reliability.

Find the Standard Error of
a Test
• Standard Error is the distance from a
student obtained score and their actual
score based on test error.
• You can figure out the standard error.
Let’s assume the standard deviation was
7, and the reliability was .55 for a
teacher-made test.

Standard Error of
Measurement (SEM)

SEM = 11 x the square root of 1-.55

SEM = 11 x the square root of .45
SEM = 11 x .67
SEM = 7.37

*So, if a child scores a 90% on a test, they could have a true score in the
range of 83 to 97. You simple add and subtract the SEM from the score
they receive.

Now that you know
SEM…
In the previous section, we saw the
standard error of measurement (SEM) give
us a glimpse of how reliable a fictitious test
is.
It is not unreasonable to consider the SEM
when determining how to use multiple
measures in the classroom.
The larger the SEM, the more of a need for
multiple measures

Data meeting

More Related Content

What's hot (19)

Viewers also liked (17)

Similar to Data meeting (20)

Recently uploaded (20)

Data meeting