The document discusses test validity and reliability. It provides formulas for calculating validity, including point biserial correlation, and shows sample calculations for item validity. Reliability is assessed using split-half reliability and Pearson product-moment correlation. Sample reliability calculations are shown for several test items. The results indicate some items have high reliability while others do not. Guidelines are provided for interpreting reliability coefficients.

1. INTERPRETING TEST SCORE
A. VALIDITY
In terms of test validity, we can show the tests to the colleagues for face validity,
compare the course objective and the test items for validity, check whether the students
respond in the way they are expected in doing the test for response validity, and calculate the
point bi-serial correlation for item validity using the following formula:
rpbi = Point bi-serial Correlation
Coefficient, i.e. item validity coefficient.
Mp = Mean score of testees correctly answering the analyzed item.
Mt = Mean score of the total score.
SD = Standard deviation of the total score.
p = Proportion of testees correctly answering the analyzed item.
q = Proportion of testees incorrectly answering the analyzed item.

2. No
Nama
Peserta
Butir Soal
x x2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
1
Heny
Almaida
               12 144
2 Fifi                12 144
3 Lisa                10 100
4
Muh.
Tasrif
               10 100
5 Sofia                10 100
6 Linda                9 81
7 Herma                9 81
8 Rasmi                9 81
9 Hasrianti                8 64
10 Nurasia                7 49
11 Ramlah                6 36
12 Hardiana                6 36
13 Kasmiati                6 36
14 Zufiana                6 36
15 Nursyam                5 25
∑x 2 1 10 10 7 10 8 10 11 10 8 14 11 5 8 125 1113
Mp 10.5 6 8.8 9.2 8.5 8.4 9.6 9.2 8 8.8 9.5 8.5 9.1 10 9
Mt 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3
SD 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1 2,1
p 0.13 0.07 0.67 0.67 0.47 0.67 0.53 0.67 0.73 0.67 0.53 0.93 0.73 0.33 0 53
q 0.87 0.93 0.33 0.33 0.53 0.33 0.47 0.33 0.27 0.33 0.47 0.07 0.27 0.67 0.47
rpbi 0.37 - 0.28 0.32 0.59 0.08 0.05 0.64 0.59 - 0.22 0.32 0.05 0.32 0.62 0.56 0.34
The following steps are recommended for calculation, taking item number 1 as the
sample of calculation.
1. Determining the proportion of testees correctly answering the analyzed items:

3. 2. Determining the proportion of testees incorrectly answer the analyzed items:
q1 = 1 – p1 = 1 – 0.13 = 0.87 q6 = 1 – p1 = 1 – 0.67 = 0.33 q11 = 1 – p1 = 1 – 0.53 = 0.47
q2 = 1 – p2 = 1 – 0.07 = 0.93 q7 = 1 – p1 = 1 – 0.53 = 0.47 q12 = 1 – p1 = 1 – 0.93 = 0.07
q3 = 1 – p3 = 1 – 0.73 = 0.27 q8 = 1 – p1 = 1 – 0.67 = 0.33 q13 = 1 – p1 = 1 – 0.73 = 0.27
q4 = 1 – p4 = 1 – = 0.33 q9 = 1 – p1 = 1 – 0.73 = 0.27 q14 = 1 – p1 = 1 – 0.33 = 0.67
q5 = 1 – p5 = 1 – 0.47 = 0.53 q10 = 1 – p1 = 1 – 0.67 = 0.33 q15 = 1 – p1 = 1 – 0.53 = 0.47
3. Calculating the mean score of the total scores:
4. Calculating the mean score of testees correctly answering the analyzed items.
6 .6
5. Calculating the standard deviation of the total score:

4. 6. Calculating the item validity coefficient:
 Test item 1
 Test item 2
 Test item 3
 Test item 4
 Test item 5
 Test item 6
 Test item 7
 Test item 8
 Test item 9

5.  Test item 10
 Test item 11
 Test item 12
 Test item 13
 Test item 14
 Test item 15
B. RELIABILITY
In terms of test reliability, we can use single-test single trial method with split-half
reliability, applying Pearson product moment correlation and Spearman-Brown odd even
modal correlation this calculation may be processed through SPSS program, based on the
level of significance of 5. The formula of Pearson product moment correlation is as
follows:

6. rxy = Pearson product moment correlation between variable x and y
N = Number of students taking the test
∑x = sum of variable x
∑y = sum of variable y
∑xy = sum of multiplication of variable x and variable y
∑x2
= sum of square x
∑y2
= sum of square y
- Test Item 1
No x y x2
y2
Xy
1  12  144 12
2  12  144 0
3  10  100 0
4  10  100 0
5  10  100 0
6  9  81 0
7  9  81 9
8  9  81 0
9  8  64 0
10  7  49 0
11  6  36 0
12  6  36 0
13  6  36 0
14  6  36 0
15  5  25 0
2 125 2 1113 21
The result of this calculation is then analyzed using Spearman-Brown odd ven model
correlation to see the realibility of the test.

7. rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = 0.55)
- Test Item 2
No x y x2
y2
Xy
1  12  144 0
2  12  144 0
3  10  100 0
4  10  100 0
5  10  100 0
6  9  81 0
7  9  81 0
8  9  81 0
9  8  64 0
10  7  49 0
11  6  36 6
12  6  36 0
13  6  36 0
14  6  36 0

8. 15  5  25 0
1 125 1 1113 6
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = -0.78)
- Test item 3
No x y x2
y2
Xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 0
5  10  100 0
6  9  81 9
7  9  81 9
8  9  81 9
9  8  64 8

9. 10  7  49 7
11  6  36 0
12  6  36 0
13  6  36 6
14  6  36 6
15  5  25 0
10 125 10 1113 88
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = 0.46)
- Test item 4
No x y x2
y2
Xy
1  12  144 12
2  12  144 12
3  10  100 10

10. 4  10  100 10
5  10  100 10
6  9  81 9
7  9  81 9
8  9  81 0
9  8  64 8
10  7  49 0
11  6  36 6
12  6  36 0
13  6  36 6
14  6  36 0
15  5  25 0
10 125 10 1113 92
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.71)

11. - Test item 5
No x y x2
y2
Xy
1  12  144 0
2  12  144 12
3  10  100 0
4  10  100 10
5  10  100 0
6  9  81 9
7  9  81 9
8  9  81 9
9  8  64 0
10  7  49 0
11  6  36 0
12  6  36 6
13  6  36 0
14  6  36 0
15  5  25 5
7 125 7 1113 60
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers

12. To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = 0.18)
- Test item 6
No x y x2
y2
Xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 0
5  10  100 10
6  9  81 9
7  9  81 0
8  9  81 0
9  8  64 8
10  7  49 0
11  6  36 6
12  6  36 0
13  6  36 6
14  6  36 6
15  5  25 5
10 125 10 1113 84
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.

13. rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.76)
- Test item 7
No x y x2
y2
Xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 10
5  10  100 10
6  9  81 9
7  9  81 9
8  9  81 0
9  8  64 0
10  7  49 0
11  6  36 0
12  6  36 0
13  6  36 0
14  6  36 0
15  5  25 5

14. 8 125 8 1113 77
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.77)
- Test item 8
No x y x2
y2 xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 10
5  10  100 10
6  9  81 9
7  9  81 9
8  9  81 0
9  8  64 8
10  7  49 0

15. 11  6  36 6
12  6  36 6
13  6  36 0
14  6  36 0
15  5  25 0
10 125 10 1113 92
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the realibility of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.71)
- Test item 9
No x y x2
y2
Xy
1  12  144 12
2  12  144 0
3  10  100 0
4  10  100 10
5  10  100 10

16. 6  9  81 0
7  9  81 9
8  9  81 9
9  8  64 8
10  7  49 7
11  6  36 0
12  6  36 6
13  6  36 6
14  6  36 6
15  5  25 5
11 125 11 1113 88
The result of this calculation is then analyzed using Spearman-Brown odd ven model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = -0.67)

17. - Test item 10
No x y x2
y2
Xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 0
5  10  100 10
6  9  81 9
7  9  81 0
8  9  81 9
9  8  64 8
10  7  49 0
11  6  36 6
12  6  36 6
13  6  36 6
14  6  36 0
15  5  25 0
10 125 10 1113 88
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers

18. To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = -0.59)
- Test item 11
No x y x2
y2
xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 10
5  10  100 10
6  9  81 0
7  9  81 0
8  9  81 9
9  8  64 0
10  7  49 7
11  6  36 0
12  6  36 0
13  6  36 0
14  6  36 6
15  5  25 0
8 125 8 1113 76
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.

19. rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.71)
- Test item 12
No x y x2
y2
xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 10
5  10  100 10
6  9  81 9
7  9  81 9
8  9  81 9
9  8  64 8
10  7  49 7
11  6  36 6
12  6  36 6
13  6  36 0
14  6  36 6
15  5  25 5

20. 14 125 14 1113 119
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = 0.43)
- Test item 13
No x y x2
y2
xy
1  12  144 12
2  12  144 12
3  10  100 10
4  10  100 10
5  10  100 10
6  9  81 9
7  9  81 9
8  9  81 9
9  8  64 0
10  7  49 7

21. 11  6  36 0
12  6  36 6
13  6  36 6
14  6  36 0
15  5  25 0
11 125 11 1113 100
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.72)
- Test item 14
No x y x2
y2
xy
1  12  144 12
2  12  144 12
3  10  100 0
4  10  100 10
5  10  100 0

22. 6  9  81 0
7  9  81 0
8  9  81 9
9  8  64 0
10  7  49 7
11  6  36 0
12  6  36 0
13  6  36 0
14  6  36 0
15  5  25 0
5 125 5 1113 50
The result of this calculation is then analyzed using Spearman-Brown odd ven model
correlation to see the realibility of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers
To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is reliable (r11 = 0.70)

23. - Test item 15
No x y x2
y2
xy
1  12  144 0
2  12  144 12
3  10  100 10
4  10  100 10
5  10  100 10
6  9  81 0
7  9  81 0
8  9  81 9
9  8  64 8
10  7  49 7
11  6  36 0
12  6  36 0
13  6  36 0
14  6  36 6
15  5  25 0
8 125 8 1113 72
The result of this calculation is then analyzed using Spearman-Brown odd even model
correlation to see the reliability of the test.
rtt = Total test coefficient reliability (tt = total test)
rhh = Product moment Correlation Coefficient between the first half and the
second
half of the test (hh = half – half)
1 & 2 = constant numbers

24. To interpret the test reliability (r11) Sudjono (2003.209) provides criteria. If the resulted
calculation ((r11) is the same or greater than 0.70, the evaluated test is highly reliable.
Conversely, if the resulted calculation (r11) is smaller than 0.70, the evaluated test is not
highly reliable. Therefore, the result of calculation is not reliable (r11 = 0.48)
No
Nama
Peserta
Butir Soal
x
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
1
Heny
Almaida
               12
2 Fifi                12
3 Lisa                10
4
Muh.
Tasrif
               10
5 Sofia                10
6 Linda                9
7 Herma                9
8 Rasmi                9
9 Hasrianti                8
10 Nurasia                7
11 Ramlah                6
12 Hardiana                6
13 Kasmiati                6
14 Zufiana                6
15 Nursyam                5
UG 2 0 5 7 4 5 7 7 4 5 5 7 7 3 4
LG 0 1 4 3 2 5 1 3 6 4 2 6 3 1 3
UPPER
GROUP
LOWER
GROUP

25. C. ITEM DIFFICULTY
IF1 = Index of facility
UG= the number of correct answers by the upper group
LG = the number of correct answer by the lower group
N = the number students taking the test
(Difficult Question)
(Difficult Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)
(Medium Question)

26. (Medium Question)
(Difficult Question)
(Medium Question)
The conclusion of items covering a wide range of difficulty levels may promote
motivation. The inclusion of very easy items will encourage and motivate the poor student.
On the other hand, the more difficult items may be necessary in order to motivate the good
students.
D. ITEM DISCRIMINATION
ID = index discrimination
N = number of students in one group (1/2N)
UG = frequency of score by upper group (upper half)
LG = frequency of score by lower group (lower half
(Enough)
(Low)
(Low)
(Excellent)
(Enough)
(Low)

27. (Excellent)
(Excellent)
(Low)
(Low)
(Excellent)
(Low)
(Excellent)
(Enough)
(Low)