The document discusses the essential ingredients of effective psychometric testing. It outlines several criteria that should be followed: 1) Choosing the right test for the intended purpose; 2) Ensuring the test has established reliability, validity, and standardization; and 3) Obtaining the technical manual which details the test development and psychometric properties. Key aspects that determine a test's reliability and validity include the standardization process, item analysis, and operational definitions being provided.
1. ESSENTIALS OF PSYCHOMETRIC:
HOW TO IDENTIFY
Prof. Dr. N.K.Chadha
Dean,International Relations & Chairperson, Council for Doctoral Program,MRIIRS
Fr. Senior Consultant, UNDP, Chief Psychometrcian,MINDLER
Former Prof. and Head,Psychology Deptt, Delhi University
nkcdu11@gmail.com, nkc_du@yahoo.co.uk
2. Essential Psychometric Ingredients
■One of the common drawbacks of
psychometric testing in the corporate
environment is the
■Belief that anyone can administer
them
■Anyone can debrief or interpret the
results
3. Essential Psychometric Ingredients
■ To have effective outcomes, there are a few criteria that should be
followed:
■ Choose the right test for the right purpose from the thousands
of tests out there in the market.
■ Make sure to check for reliability, validity and standardization
process.
■ Ask the test developer to provide the technical manual of the
test which contains the details about the test development
process and psychometric properties.
■ Make sure to look for item level properties as well.
4.
5.
6.
7.
8. 1. 256x2 = ?
2. (25+30) x 5 = ?
3. (3x4 + 5x6) (6+5) = ?
4. under root of 225 = ?
5. Eighth root of 34521670 = ?
6. (8+10-5) x 20 = ?
10. Day Room 1
(actual/corr
ected)
Room 2
(actual/corr
ected)
Room 3
(actual/corr
ected )
FIRST 9.00pm/9.00
pm
9.05pm/9.00
pm
9.07pm/9.00
pm
SECOND 9.00pm/9.0
0pm
9.05pm/9.0
0pm
9.09pm/9.0
0pm
THIRD 9.00pm/9.0
0pm
9.05pm/9.0
0pm
9.06pm/9.0
0pm
11. ■ WHEN A WATCH IS VALID , IT HAS TO BE
RELIABLE
THIS LEAD TO
WHEN ANY PSYCHOMETRIC INSTRUMENT IS
VALID THEN IT HAS TO BE RELIABLE
12. ■ WHEN A WATCH IS RELIABLE, IT MAY OR
NOT BE VALID
■ THIS LEADS TO
■ WHEN ANY PSYCHOMETRIC INSTRUMENT
IS RELIABLE, IT MAY OR MAY NOT BE
VALID
13. ■ WHEN A WATCH IS NOT RELIABLE , IT CAN
NEVER BE VALID
■ THIS LEADS TO
■ WHEN A PSYCHOMETRIC INSTRUMENT IS
NOT RELIABLE, IT CAN NEVER BE VALID
14. METHODS OF RELIABILITY
■ 1. SPLIT HALF RELIABILITY
■ 2. TEST-RETEST RELIABILITY
■ 3. PARALLEL FOR RELIABILITY
■ 4. METHOD OF RATIONAL EQUIVALENCE
■ 5. CRONBACK ALPHA
15. You are a coach and while you are couching, client
said that it seems you have not done the home work
and you are wasting my and your time. How to your
respond
■ A. Ignore the client
■ B. Promise to recheck your style/assessment
■ C. Cancel the session
■ D. Try to find what bothers the client
16. When things don't go my way, I...
a) Try to force it around
b) Speak my mind and complain
c) Pretend it's not happening
d) Handle the situation at hand
17. When facing a challenging situation, I...
a) Face it with an open mind
b) Put a smile on my face
c) Turn and run
d) Pretend it's not happening and hope it will
go away
18. PLANNING THE TEST
PLANNING THE TEST
DEFINE THE TESTING
UNIVERSE
DEFINE THE TARTGET
AUDIENCE
PURPOSE OF TEST
19. Thus, the three anchors of a good psychometric test:
§ Standardization Process
§ Reliability
§ Validity
20. Standardization Process
■ Check the characteristics and the size of the sample of the norm group.
■ The norm group should truly represent your target audience.
■ Test should be standardized on a similar sample (in terms of language,
education, occupation, culture etc.) as for whom the test has been made.
21. Standardization Process (Ctd..)
■ The norm group should include at least more than thousand tested
individuals from the target group.
■ Changing the characteristics of the sample group (language of the test,
geography etc.) requires re-standardization.
§ Example: If the test was originally developed and standardized for
an English-language population, it is not suitable as such for Hindi-
speaking population.
22. Reliability- The Consistency and Repeat-
Ability Of Test Result
■ Check for test reliability. The test should produce consistent results every time it is
administered and would be free from outside factors such as stress, temperature
etc.
■ Administer the test to a group, and re-administer it on the same group after a week.
If the test is stable and reliable, the correlation between both the scores will be
positive.
■ An unreliable test would produce totally different scores on retesting.
■ Usually a correlation of 0.8 and higher is considered to be a good reliability.
23. Validity: The Accuracy And Predictability
Of The Test
■ Extent to which the test measures what it intends to measure.
■ Read the items carefully to see whether the items ‘look like’ what they say are measuring.
■ Try to analyze whether the test is measuring the construct it is claiming to measure rather
than something else. For example: if a test claims to measure numerical ability and is only
available in ‘Hindi” then it will not only measure numerical ability but also the ability to
read and understand Hindi for non-Hindi speaking group.
■ See for the degree of correlation between the test and other well established test of the
same characteristic or ability.
24. ■ In case of any doubt, ask the test developer what is the purpose of the test
and what constructs does it measure.
■ If the test developer has adopted fair measures to develop and validate the
test, he or she will be able to tell you what the test is supposed to
measure.
■ In the case of Validity the accepted degree of correlation is 0.7 or higher.
26. Objectivity
■ The test and its results should be objective and free from subjective elements. It should
measure individuals objectively, free from personal biases and perspective.
■ Try to analyze the test as well as its results. For a test to be fair and unbiased, no individuals
from any particular group should be disadvantaged when completing the test.
■ If your target audience has some additional needs, make sure to check and make adjustment
for them in the test.
27. Operational Definition
■ The test should have a proper operational definition
■ Ask the test developer for the Operational Definition of the construct and individual
dimensions.
■ Constructs are simply the manifestation of ideas.
■ An operational definition will enable you to have a clarity of the construct and of the
dimensions the scale will measure.
28. Item-Analysis
■ It is a process through which only those items that are valid and are suited to the purpose of the test
are retained while the rest are either modified or eliminated.
■ Ask the test developer to share the item-analysis charts with you. This is usually included in the
technical manual of the test.
■ These charts would reveal the item-difficulty index (only in case of test which have right or wrong
answers), item-discrimination value, item reliability, and item-test correlation.
■ It will reveal individual item as well as total test quality.
■ It will also enable you to know how many items have actually been removed or modified from the
original version after the analysis.
29. Item-Difficulty
■ Refers to the percentage of respondents who responded to an item correctly.
■ Item-Difficulty will let you know the difficulty level of each of the item.
■ By looking at the Item-Difficulty table you will get an idea regarding each of the item
and the ones (too easy and too difficult) that were eliminated.
30. Item-Discrimination
■ Separates or differentiate among respondents on the basis of how well they
know the material being tested.
■ By looking at this table, you will get an idea regarding the extreme scores.
31. Item-Reliability
■ Item reliability is obtained by calculating correlations between all the items of a test.
■ Also called as Inter-item Correlation.
■ Item reliability chart will give you a clear idea about the reliability of each of the items.
■ For a particular item if more than 50% correlations are found to be significant then that
item would be retained
32. Item-Total Correlation
■ Item-total correlation will assess the contribution of a single item to overall consistency
of the test.
■ This chart will give you an idea regarding the strength and direction of relation
between an individual item and the overall test.
■ It will let you know the total contribution of each item with the overall consistency of the
etst.
33. Test Manual
■ It will give you the information of the entire process involved in test construction along
with various indices calculated for standardization of the test.
■ It will be able to answer in-depth questions about the test, how it operates, how it was
created, what it is designed to do, how the results are calculated, and what to do if you
do not agree with the results.
■ It will reveal the psychometric properties of the test.