Professional Documents
Culture Documents
Rationale
Prof. Jonathan V. Macayan
MIT-Department of Psychology
Causes of Fatal Accidents by Decade (percentage)
Cause 1950s 1960s 1970s 1980s 1990s 2000s All
Pilot Error 40 32 24 25 27 26 29
Mechanical Failure 21 20 23 21 21 28 22
Sabotage 5 5 11 13 10 9 9
Other Cause 0 2 2 1 0 1 1
The table above is compiled from the PlaneCrashInfo.com accident database and represents
1,300 fatal accidents involving commercial aircraft, world-wide, from 1950 thru 2009 for which
a specific cause is known. Aircraft with 10 or less people aboard, military aircraft , private
aircraft and helicopters are not included.
Participants
Prof. Jonathan V. Macayan MIT-
Department of Psychology
SAD-MAM_P was administered to a randomly selected
participants (N=144) in a private flying school in Manila.
Examinees were given a limited time (35 minutes) to
answer the entire test. Time pressure is incorporated,
since decision-making in actual flights especially in
emergency situations require prompt execution. After
giving complete instructions on the test (including the time
limit), the examinees were requested to answer the test to
the best of their knowledge. Debriefing proceeded after the
completion of the test.
Item Characteristic
Item Difficulty Item Analysis
Curve (ICC)
(pH + pL)/ 2
Reliability
FIT Indices
Analysis
INFIT 1.00; OUTFIT was 1.10
(Internal Consistency)
◦ Cronbach’s alpha (α=.457)
Reliability Indices
◦ Split-half reliability ( =.398) Person = .43; RMSE of .57
Item = .34; RMSE of .21
Item Analysis
Item Difficulty
◦ Item Difficulty (24 items = Easy items- item 2, 3, 6, 7, 9,
10,11,15,18,19,20,22,24
Average)
Difficult items- item 1, 4, 5, 8,
◦ Item Discrimination(5, 7,12,
12, 13, 14, 16, 17, 21, 23
13, and 18 are Reasonably
Good )
Results
Prof. Jonathan V. Macayan MIT-
Department of Psychology
Prof. Jonathan V. Macayan MIT-
Department of Psychology
Prof. Jonathan V. Macayan MIT-
Department of Psychology
Prof. Jonathan V. Macayan MIT-
Department of Psychology
22 out of 24 items passed the criteria of fit in the
1PL-Rasch Model, which indicates that the items
in the test represent the expected ability of the
target examinees (student pilot).
Discussion
Prof. Jonathan V. Macayan MIT-
Department of Psychology
CTT and IRT both provided a marginal level of
reliability estimate of the instrument when data
was tested for item consistency (CTT =.478; IRT
=.34); such findings can be attributed to the
quantity and quality of observations (scores) in
the test, where all the participants (N=144)
performed well in answering the majority (at
least 50%) of items correctly (lowest score = 12/
highest = 24; total mean score= 18.50/ standard
deviation = 2.70/ sample variance =7.49).
Discussion
Prof. Jonathan V. Macayan MIT-
Department of Psychology
The results of item difficulty vary between CTT
and IRT, where all items are classified as average
in the traditional method while the Rasch model
delineated easy and difficult items in the test. For
example, IRT was able to detect the easiest
(Item no. 2) and the most difficult (Item no.5)
items in the test, such findings were not detected
in CTT.
Discussion
Prof. Jonathan V. Macayan MIT-
Department of Psychology
The analysis of data also provided an important
aspect of the instrument’s validity in terms of fit
(INFIT =1.00; OUTFIT = 1.10). This indicates
that the matching of item difficulty to person’s
ability (or vice-versa) was well-defined and
corresponding.
Discussion
Prof. Jonathan V. Macayan MIT-
Department of Psychology
Other methods of reliability testing aside from internal
consistency can be recommended (i.e alternate form,
inter-rater, test-retest) to further verify the reliability of
the measure.
Recommendations
Prof. Jonathan V. Macayan MIT-
Department of Psychology
THANKS FOR
LISTENING!!!
Prof. Jonathan V. Macayan
Department of Psychology
Mapua Institute of Technology