b. Analysis Technic of Items
1. Validity
Validity is the accuracy of measuring item valid. Test need to determined in order to know the quality of test. Divide Anas
Sudijono, 2011: 164 validity in two kinds, namely: a
Validity of test Validity of test is used to measure question competensively.
The following types of the validity of tests are: 1
Validity of Rational Validity of rational is validity earned that by logical.
Devide validity of rational have two kinds, namely: a
Validity of Content The content of the test represent all the materials.
b Validity of Construction
Validity of contruction is validity that build the test, regulate thinking aspect that already mentioned in
special instruct goals.Suharsimi, 2009: 66. 2
Validity of Empiric Validity of empirical is accurancy in measuring
something based on empirical analysis Anas Sudjiono, 2011: 167. This is kinds of empiric validity:
a Validity of devination
The validity of devination is a condition that shows how far a test can show its ability to predict accurately
what will happen in the future Anas Sudjiono, 2011: 168.
b Validity of Comparison
Validity of Comparison is the tests can show the relationship between a first test with the next test Anas
Sudjiono, 2011: 176-177. b
Validity of Item According to Anas Sudjiono 2011: 182, the validity of an
item is the accuracy of measuring the items. To calculate valid of item can use biserial point, that formula is:
pbi =
� − � �
Note:
�
pbi
= coefficient biserial correlation M
p
= average score from subject which are true for item Mt = total average score
St = standar deviation and total score P = students proportion which is answering true
Q = students proportion which is answering false q = 1-p
Suharsimi, 2009: 79 Correlation index biserial point
�
pbi
can be result from calculate with r
table
on the significant level 5 appropriate with amount students.
2. Reliability
Reliability is the extent or degree of consistency of an instrument. Reliability test with regard to the question whether a
scrupulous and test can be trusted in accordance with the criteria have been set. A test is reliable if always give the same result if
working on the same group on a different time or opportunity Zainal Arifin, 2011: 258. The total reliability can calculate with
formula KR-21:
11
= �
2
− − �
2
− 1 Note:
r
11
= total reliability of test K
= amount of item test ax
2
= score variation = average score
Nana Sudjana, 2011: 19 According to Anas Sudijono 2009: 209, to giving
interpretation coefficient of reliability in general using foundation it’s:
a. Reliability of the test obviously have been the high reliable if
r
11
same or more than large with 0,70. b.
Reliability of the test obviously have been including the high unreliable if r
11
less smaller than 0,70. 3.
The Level of Difficulty Difficulty is a real number between 0 and 1 which expresses a
measure of the difficulty of the item, intended as the proportion of