Page 160 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Educational Measurement and Evaluation

Notes T-scores are also like t-scores, with the only difference that in T-scores, the raw scores are
normalized. The basis of all standard scores is normalized distribution curve. If distribution
is not equal, then equivalent standard scores will not be correct too. Therefore, the distributor
is made equal, and then raw scores are changed into standard scores. All other characteristics
of T-scores are similar to t-scores. By Stanines is meant ‘standard nine’, that is, the scale
which has nine units. Stanine 9 is the largest unit, and Stanine 1 is the smallest, and at the
middle is Stanine 5.
To calculate standard norm, the test is administered on a large group and standard deviation
and mean of scores are found out. Now the score to be analyzed is changed into some
standard score (Z or t etc.), and the thus obtained standard score is analyzed. The value of
standard score tells how less or more a student is from the mean. A test-maker makes a
table for changing raw scores into standard scores, with the help of which the raw scores of
the test can be changed into standard scores. For example, Usha Pande has scored 60 marks
in a mathematics achievement test, and on that test, the value of Z-score is 2 for 60 score; it
would mean that Usha Pande obtains +2 more marks than an average girl in mathematics
achievement. 2Z = 0Z + 2Z). In the same way, if the value of t-score of Nishi’s English scores
is 75, it would mean that she obtains 25 t-score than average girls. (75t = 50 t + 25 t or 75 =
50 + 25). In the same way, Sten and Stanine scores can also be analyzed.
Generally, standard scores and norms appear to be similar, but this is only an illusion. In
fact, there is basic difference between the two. According to Ross and Stanley, the term
‘standard’ has objective inherent in it, that is, it tells about our goal. While ‘norm’ indicates
present achievement. We want to proceed towards the standard which remains stable, while
the degree of stability is deficient in norms.

Z-score is generally plus/minus 3 and t-score between 20-80, while Sten-score and
Stanine-score are between 0-9. Relative Z, t, Sten and Stanine value to a score tell
that how more or less that score is from the mean (0, 50, 5.5 or 5 respectively.)

11.4 Characteristics of Norms

It is essential for norms of a test to have the following essential qualities :
1. Novelty.
2. Representation.
3. Meaningfulness.
4. Comparability.
1. Novelty : By being novel or up-to-date is meant that the norms should not be outdated, that
is, norms should not be constructed on the basis of test scores which were administered a
long way back because time interval can effect change in students’ abilities. For this reason,
if an intelligence test constructed in 1990 will not be suitable to analyze scores of students
on a test administered in 2004, as it would not be proper. Therefore, the norms should be
changed from time to time so that they remain novel and up-to-date.
2. Representation : By representation of norms is meant that norms should be developed from
the scores obtained from the representative group, whose scores have to be analyzed.
Therefore, if some skill of class 9 students has to be analyzed, then the norms too should be
developed on the basis of scores obtained from class 9 students. Besides, these students
should be equal to other students in other traits. Test norms should be constructed on the
basis of scores obtained from a large group. A small group cannot represent the whole

154 LOVELY PROFESSIONAL UNIVERSITY