Transcript Powerpoint

Reliability, Validity and Fit
Functional Independence Measure
(FIM): Example 17s

In Example 17, 35 arthritis patients have been
through rehabilitation therapy. Their admission
to therapy and discharge from therapy measures
are to be compared. They have been rated on 18
items of the Functional Independence Measure
(FIM™). Each item has seven levels. A prime
concern is to measure the change of patient
disability from admission to discharge. First, we
need to check validity and reliability.

Here, admission and discharge measures are
placed together (“stacked”)
Looking at the Diagnostics



What information does this table provide?
How do we read it?
What do we learn about these data?








Item Fit
Person Keyforms
Summary Statistics
Variable Map
Probability Curves
Dimensionality Map
(Table 7.1- Person Responses)
(Table 2:2- General Keyform)
The Structure and Stability of the Functional
Independence Measure (FIM)
http://www.rasch.org/memo50.htm
 Cites sources regarding deficiencies of raw
score analysis of scales
http://www.rasch.org/rmt/rmt103e.htm
 Defends use of Rasch to analyze these
data for a critic
Another option for checking
comparability
In Winsteps
 Exam12.txt
 Start by copying exam12.txt information
and pasting it in exam12lo.txt control file.
Save to desktop.
 Do the same for exam12hi.txt
 Run each of the analyses in Winsteps.

Export Measures to Excel







Output Files  Item File
Display with Excel
File format: Text: tab-delimited (best for Excel)
Column Headings: Include
Insert  Chart  Scatter
Put cursor in “Data Range” box, then highlight
columns in worksheet.
OR you could create a scatterplot in Minitab
This concept can be applied elsewhere
Comparing people over time (how much
people improve after intervention)
 Comparing items over time (how much
easier items become after intervention)
 “Rack and Stack”

Example: www.raschsig.org/rmt171.pdf
To look at person change: Stack


Items are modeled to
maintain their difficulties
across the two time points,
and the persons are entered
twice, once at admission to
rehabilitation and once at
discharge from it.
Changes in patient
independence can be
identified by cross-plotting
the admission and discharge
measures for each patient, as
in example 12 or looking at
the difference in measures
Items
Persons
Time 1
Persons
(again)
Time 2
To look at item change: Rack


To look at WHAT has changed instead of WHO
has changed.
Items are entered twice, once at admission to
rehabilitation and once at discharge from it.
Items
Persons
Time 1
Items (again)
Time 2