M. A.

Part I
Semester I

By
Balaji Niwlikar

https://www.careershodh.com/
 Measures of Central tendencies:
 Mean
 Average
 Median
 Middle value
 Mode
 Most repeated score
 Measures of Variability:
 Standard deviation
 Quartile deviation
 Range
 Z Scores
https://www.careershodh.com/
 Standard/Average performance .
 Methodology – to understand psy tests and
proper interpretation of scores.
 Norms, Reliability , Validity, Item Analysis and
Test Design.
 Raw Scores - 35 in English. 20 in math ????
 Expressed in different units –kg, hour , no .of
correct/incorrect responses, no. of trails,
 So we cant just directly compare it.
 It can only interpreted in clearly defined and
uniform frame of references .
https://www.careershodh.com/
 Definition:
“Norms may be defined as the average performance on a particular
test made by a standardization sample.”
Standardization sample-true representation population, cross
cultural,
 Empirically established .
 To discover where S/he falls in that situation - we convert
Raw scores into Derived Score-
 Derived Score-
 Difficulty level – all score low 1,5,10 /1000. we can compare D scores
 2 objectives /purpose/ goals of Derived Score .
1. To indicate the individual’s relative standing in the normative
sample and thus permit an evaluation of her/his performance
in reference to other persons.
2. To provide comparable measures that permits a direct
comparison of the individuals performance on different tests.
https://www.careershodh.com/
 Types if evaluation
1. Formative /concurrent Evaluation
 To evaluate learning
 Not standard and Informal way
2. Summative Evaluation
 To find out which area is strong / weak
 Useful in Training program, language ,math
3. Diagnostic Evaluation
 At end of program
4. Follow up Evaluation
 In Corporate areas
 Attitude changing Program

https://www.careershodh.com/
 Steps in developing Norms:
1. Defining the target population
 Normative group -based on intention of test .
2. Selecting the sample from the target
population
 True representative sample.
 Cross sectional
 Large sample
 Random sampling
3. Standardizing the conditions
 Test administration must be standard, valid
https://www.careershodh.com/
 Derived Score-
 Expressed in two major ways i.e types

1st Developmental Level Attended or

2nd relative position within a specific group

https://www.careershodh.com/
 Norms

Developmental Within Group

Norms Norms

Percentile Standard Scores Sten Stenine Deviation IQ

Age Grade Ordinal scale

https://www.careershodh.com/
 A way to attach meaning to scores.
 To indicate how far an individual has progressed the
normal developmental path.
 Ex – children smile after certain age.
 Binet proposed early developmental age norms and
gave concept of Mental Age
 Mental Age-
 Binet –Simon
 Items passed by the majority of 8 years old in
standardised sample were grouped together and placed
in the 8 year level. = Mental Age of 8.
 8 year old Sheldon Cupper scored well on intelligence
test of 80 year old –means he has Mental Age of 80 year
old https://www.careershodh.com/
1. Age equivalent norms:
 Criteria – Ave. performance of standard sample at
certain age level.
 Most suitable for trait or ability which increases
systematically.
 e.g. Height, weight, Cog. Abilities, intelligence etc.
 Limitations –
1. It is not fully standard and uniform unit for
measurement for over all period.
2. Some of the traits can not be explained by age norms
though they are related to age.
Ex- maze learning will not develop after adolescent, IQ will
not increased after 16 but vocabulary can.

https://www.careershodh.com/
2. Grade Equivalent norms
 Like age equivalent norms – criterion –Grade/Standard
 In field of educations . Achievement test & educational test.
 Ex -4th grade performance in math ,language skill.
 The average no. of problems solved correctly on a math test
by the 4th Grader in a standardization sample is 23,them
raw score of 23 corresponds to grade equivalent of 4
 It can be expressed in decimal (4.5); If we considered
months .
 Limitations –
 Same students in different subjects not comparables (math
with social sciences).
 Not suitable for higher grades level ( 1 subject for 2 years)
 Not suitable for subjects which occurs rapidly growth in
lower grades ;will be same in higher grade
https://www.careershodh.com/
3. Ordinal scales –
 Not like statistics (providing rank order to individual without
knowledge about amount of differences between them )
 Designed to identify the stage reached by the child in the
development of specific behavior functions
 Originated from research of child psychology
 Based on Model of Guttman Scale or simplex(1944)- successful
performance at one level implies success at all lower level.
 Success in functions of locomotion , concept formation, etc.
 Gesell Developmental Schedule (1947)–child has attained a
certain level in –motor, adaptive ,language ,& personal-social.
 Development theory of Piaget – schema ,object permanence

https://www.careershodh.com/
 Almost all psychological test provides it.
 Used most near comparable standard group ex
- same chronological age /same school grade.
 Within group scores / norms have uniform
and clearly defined quantitative meaning .
 Used in most of statistical analysis.

https://www.careershodh.com/
 Nominal scales –
 involve classification or categorization based on one or more
distinguishing characteristics,
 Ex - “men,” “1,” “B,” or “women,” “2,” or “A.”
 Ordinal scales-
 permit classification and rank ordering on some characteristic
 Ex- merit list of SP College.
 Interval scales -
 contain equal intervals between numbers.
 But like ordinal scales, interval scales contain no absolute zero
point
 Ex -IQs of 80 and 100
 Ratio Scales
 In addition to all the properties of nominal, ordinal, and interval
measurement,
 It scale has a true zero point.

https://www.careershodh.com/
1. Percentiles:
 Most common and popular
 Percentile -% of persons (standard sample) fall below a given point.
 Percentile and Percentile Rank are two different concepts.
 ex – if the 30% of the person obtain fewer than 18 problems correct
on math then raw score of 18 corresponds to 30th percentile (P30) i.e.
percentile rank is 30 and percentile score is 18
 Lower the percentile the poorer the persons standing.
 PR 50 –median .PR 25 and PR 75 are called 1st n 3rd quartile points ..
 Different from percentage ( %) – raw score where percentile is
derived score.
 PR 0 & PR 100?
 a raw score lower/more than obtained in in the standard sample.

https://www.careershodh.com/
 Simple to understand.
 Familiar to population.
 simple for computation.
 Percentiles are placed on an Ordinal Scale
means it regarded as rank in group of 100
 Limitation – inequality of unites
 The distance between the extreme PRs is larger
than the PRs in the middle of the NDC.
 Percentiles can be converted into large number
of other norms.

https://www.careershodh.com/
https://www.careershodh.com/
 Standard scores :
 Increasing trend . most satisfactory derived score.
 a SS is a raw score that has been converted from one
scale to another scale, where the latter scale has some
arbitrarily set Mean and SD.
 Raw scores may be converted to standard scores to
easily interpret.
 With a standard score, the position of a test taker’s
performance relative to other test takers is readily
apparent.
 SS can obtained by Linearly & Non linearly
transformed
 Ex -z scores, T scores, stanines, and some other standard
scores.

https://www.careershodh.com/
A. Linearly transformed score –
 They retain exact numerical relation of the
original raw score.
 Standard score duplicate all the properties of
raw score thus all results are distortion less .
 Units of the scale are equal so that they
convey the same meaning throughout the
whole range of the scale.
 They removes the problem of inequality.

 Simply known z scores

https://www.careershodh.com/
1. Standard score/ z scores –
 It express the persons distance from the mean in
the terms of SD of the distribution.
 zero plus or minus one scale. This is so because it has
a mean set at 0 and a SD set at 1.
 z=(X-M)/SD
Limitation of Linearly transformed standard score
If one distribution is skewed and other is normally
distributed then two standard scores cant be compared
Lay people may uncomfortable with z-scores.
 don't like negative numbers
 uncomfortable with a z-score of 0 being average.
 Ex- Swapnil got z-score of 0.

https://www.careershodh.com/
A. Non Linearly transformed score –
 when the data under consideration are not normally
distributed yet need compare with normal
distributions .
 Here ,the resulting SS does not necessarily have a
direct numerical relationship to the original, raw
score.
 Examples
1. mental age,
2. percentile score,
3. Normalized standard score

https://www.careershodh.com/
 Normalized standard score
 SS which are expressed in the terms of normal
distribution
 Maeshall & Hales (1972) ‘’Normalized standard score
which have been adjusted to produce a normal
frequency distribution and convert to a standard base
with pre assign Mean & SD’’.
 NSS can expressed in same form of linearly transformed SS
i.e. with Mean= 0 and SD =1.
 Examples
 T scores ,stanines, sten ,C scores , Deviation IQ

https://www.careershodh.com/
1. T Scores
 called a fifty plus or minus ten scale; i.e, a scale
with a mean set at 50 and a standard deviation
set at 10.
 Devised by W. A. McCall (1922, 1939) and
named a T score in honor of his professor E. L.
Thorndike,
 This system is composed of a scale that ranges
from 5 SD below the mean to 5 SD above the
mean.
 T score = 50+/-10z
https://www.careershodh.com/
3. Stanines Stenines= 5+1.96Zn
 Standard nine.
 Distribute entire scores into 9 units
 It has mean @ 5 and SD at 1.96
 If researcher knows PR scores corresponding Stanines value
can be calculated.
 Stanines= 5+1.96 Zn
 Zn –Normalized z scores (we already calculated PR.)
 +v –reasonably easy to understand .
 Useful to counselor, educational psychologist , selection &
recruitment process.ad

https://www.careershodh.com/
https://www.careershodh.com/
 Sten scores:
 They are also called as Standard Ten.
 After proposing 16PF Raymond Cattle
proposed the concept of Sten scores.
 It distributes entire score range into 10 units.
 It has mean of 5.5 and SD of 2.
 If researcher knows PR scores corresponding
Stanines value can be calculated.

Sten scores= 5.5+2Zn

https://www.careershodh.com/
https://www.careershodh.com/
5. C Scores
 G.P. Guilford
 11 standard units
 Ranger from 0 to 10

https://www.careershodh.com/
 Deviation IQ
 IQ
 Not comparable for different age group IQ of
115 @ Age 10 and IQ of 125 @Age of 12
 Deviation IQ is a Normalized standard score
has M=100 & SD 16 for Stanford Binet Scale
 Deviation IQ is a Normalized standard score
has M=100 & SD 15 for Wheschlers
intelligence test .

https://www.careershodh.com/
 The theory talks about how norms are
interchangeable.
 It refers to the concept that if researcher knows one
type of norm he can predict about the other.
e.g. If researcher knows about percentile score of a
subject that score can be converted into a Stanine
or Sten score.
 But in the case of linear norms relativity
experiences limitations.
 For conversion of the score one should know the
shape of the distribution too.

https://www.careershodh.com/
 Three principle reasons of test score variation –
1. Content –verbal/ numerical /spatial
2. Scale unit-different SDs-16/15
3. Standardization of samples- slow/Ave/better will
matter
 Normative Sample- large, representative, selective factors,
defined population
 National Anchor Norms –solution for the lack of comparability
– equipercentile method –scores are considered equivalent
when they have equal percentiles for different test .
 Specific Norms –standardize test on more narrowly defined
population( ex 1st FYBA students ) – local norms
 Fixed reference group- college board SAT -1st
 Item Response Theory- used for difficulty. To established
uniform ‘sample free’ scale of measurement ie applicable to
person/group
https://www.careershodh.com/
 Computers play an important role in generating
data analysis.
 It helps in conduction of experiments.
 It influences the process of test construction.
 Calculation of item total correlation, item analysis
is possible with the help of computers.
 It is useful in the method of factor analysis too.
 Following calculations became popular as well as
possible due to computers.
https://www.careershodh.com/
 Exploratory
factor
analysis

Online test Confirmatory

conduction factory analysis

Computers
helps in

Cross cultural
studies New methods
of reliability

Easier validity
computation

https://www.careershodh.com/
 Computer scoring
 Interactive computer system
 System for Interactive Guideline Information (SIGI)
 Major concern
1. To score comparability
2. Narration interpretation scoring

https://www.careershodh.com/
 For more study material visit
https://www.careershodh.com/

https://www.careershodh.com/