Você está na página 1de 27

Analyzing test items

Using Microsoft Excel/LibreOffice Calc*

*LibreOffice Calc use the same formula syntax.

Prepared by: Lim Ming Hui


Preparation: Data Frame
Some Observation
● Each variable is placed in a new column
● Give a unique ID for each student (e.g. I used
“6K_2” to indicate that the sample is from 6K
class, and 2 represent student)
● Remember write out the ID in your exam paper!
● Copy the original raw data and paste it to a new
Book, named as “Analysis”, please exclude
the ID column while copying the original raw
data
● Rationale:
– You will not “hurt” your original raw data if you make
some accidental modification to your raw data
– The analysis process must perform sorting
ascending
Before Start: Important!
● In your original raw data worksheet, please
calculate the total marks of each student obtain
(using =SUM(aX:aY) formula in each row)
● Then, calculate the average mark of each
student (using =AVERAGE(aX:aY)), and
variance of each student (using
=VAR.S(aX:aY))
AVERAGE
VARIANCE
● After that, copy and paste the result to the
“Analysis” workbook (another new column),
remember paste as
● *Excel user refer to the
help documentation.
This step is similar.
Sort Ascending
● Selecting the raw data
(don’t select all!), and
then go to Data: Sort
Ascending
● *MS Excel please
refer to the help
documentation
regarding Sort
Ascending
● You will receive a
similar* dialog box,
select Current
selection
● Repeat this step to
other question item
● *MS Excel user
please refer to help
documentation
● Look down to get some empty space, and then enter: Mean,
Variance, Count, Difficulty Coefficient, Discrimination
Coefficient, Validity of Test, Reliability of Test, Sum of
Variance and Sum of Discrimination
Step 1: Calculate Basic Statistics
● Calculate the mean and variance of each
question item:
– Mean: =AVERAGE(Cell:Cell)
– Variance: =VAR.S(Cell:Cell)
After you type =AVERAGE( ,
select the data range,
And then close it using ) and
Hit enter!.

Your formula should be like this:


=AVERAGE(aX:aY)

*X and Y are the row number


*a is the column name in alphabetical
Let’s try it!
● Do the same thing for the variance calculation.
● Remember that the variance formula is
=VAR.S(aX:aY)
Save time!
● After you have successfully calculated the
mean and variance, you can continue to work
through the other question item, but it will waste
some time for manual input.
● To quickly calculate the remaining items, select
the first two:
● Notice that you will see a black small box near
the right corner, put your mouse to the black
box and you will see a small “+” pointer, and
then just drag it!
● *(Instructor demo is required)
Step 2: Calculating totals
● Go through to the last
column, and insert a new
column named TOTAL, which
computes the total mark for
each student.
● Formula:
=SUM(aX:aY)
● Lastly, remember to calculate
the average total mark!
Step 3: ID and Sum of Variance
● Calculate the total student and write it on the
COUNT.
● Calculate the Sum of Variance, which means
that sum all of the question item variance.
(Using =SUM() )
Step 4: Difficulty Coefficient
难度系数
● This calculation requires the mean of each
question item.
● Type the formula:
= 1 – (MEAN/Total mark of the question)
● Example: (Mean of question 1 is 3.6, total mark is
5)
= 1 – (3.6/5)
● Remember to drag it after calculated the first one!
Step 5: Discrimination Coefficient
区分度分析
● This calculation requires the data have been
sort ascending
● Calculate the 27% range (High and Low) of the
sample: Example, if I have 30 samples, my 27%
range: 30 * 0.27 = 8.1. If appear decimal
point, round off to the nearest integer. So my
27% range: 8. I need the first 8 sample as the
low and the last 8 as the high.
● Use 2 different color to indicate the low and
high range in your workbook:

LOW

HIGH
● So, the formula for discrimination coefficient:
= ((Mean_High – Mean_Low)/Total mark of question)
● Example,
Mean_Low = 2, Mean_High = 6,
Total mark = 10,
= ((6 – 2)/10)
= 0.4
● Remember to drag it after calculated the first one!

Calculate the SUM OF DISCRIMINANT after you have
calculated all question item discrimination index
Step 6: Reliability of Test 信度分析
● This formula require average total mark and
= (Total_Question_Item/Total_Question_Item – 1)
* (1 – (SUM OF VARIANCE / Total_Variance))
Can write directly 19, since total
● Example, Question: 20, 20 – 1 = 19

Total_Question_Item = 20, SUM_VARIANCE =


13.111, Total_Variance = 52.988
= (20/19) * (1 – (13.111/52.988))
= 0.79218
Step 7: Validity of Test 效度分析
● This formula require Sum of Discrimination
= (Sum_of_Discrimination / Total Question Item)
● Example,
Sum_of_Discrimination = 14.786
Total Question Item = 20
= (14.786/20)
= 0.7393
Result Interpretation

Difficulty Coefficient 难度系数

Value Range Description

≤ 0.2 Too easy

0.2 < x ≤ 0.4 Suitable

0.4 < x < 0.8 Difficult

≥ 0.8 Very Difficult


Result Interpretation

Discrimination Coefficient 区分度分析

Value Range Description

> 0.4 Excellent

0.3 ≤ x ≤ 0.4 Good

0.2 ≤ x ≤ 0.3 Normal

< 0.2 Bad


Result Interpretation

Reliability Index 信度分析

Value Range Description

< 0.5 Bad

0.5 ≤ x ≤ 0.8 Good

> 0.8 Excellent


Result Interpretation

Validity Index 效度分析

Value Range Description

≥ 0.4 Good

0.2 ≤ x < 0.4 Normal

< 0.2 Bad

Você também pode gostar