Test Analysis Documentation
back to test analysis
Enter data
Enter your test data into the textarea boxes in a tabdelimited format. It is easiest to do this by copying and pasting data from a spreadsheet. If you only want to analyze one test, then paste it into the box for Text A. If you want to analyze two tests, such as a pretest and postest or the same test given to two different student groups, then paste data from one test into the box for Test A and paste data from the other test into the box for Test B.
Data should be formatted as shown below. It should NOT include scores, but rather just the raw data. The first line contains the question numbers on the test. The second line contains the answer key. Subsequent lines are students' names and responses.
Name 1 2 3 4 5
Key b e d d c
name1 c c c c
name2 d b e c d
name3 d e e e
name4 a b d e c
name5 d a d c c
You can include as many questions as you want and as many students as you want. Blank responses should be included since the program handles them accordingly.
Select statistical functions
Available functions are:
Single test statistics

Standard Deviation

standard deviation of the mean

Mean

average score

Median

median score  50% of scores are greater than the median and 50% of scores are below the median.

Maximum

greatest score

Minimum

lowest score

Twotest statistics

Normalized Gain Classwide

The normalized gain is calculated as <g>=(postpre)/(100pre) . This is basically the gain divided by the maximum possible gain. The classwide normalized gain is calculated using the average postest score and average pretest score for the class.

Normalized Gain Individualized

The normalized gain is calculated as <g>=(postpre)/(100pre) . This is basically the gain divided by the maximum possible gain. The normalized gain individualized is calculated using the average of individuals' normalized gains.

Select view options
You can choose whether or not you want to see individual student scores or if you simply want to see the statistical results.
You may also choose whether or not to exclude outliers. Outliers are defined as students who skip more than 25% of questions. You may modify this; however, it requires having your own configuration file. Please email Aaron Titus at titus@mailaps.org if you want to have your own personal configuration file.
Customize question analysis
It can be helpful to analyze only certain groups of questions on a test. For example, if a certain group of questions test the same objective, then you may want to calculate the average normalized gain only on that particular objective.
Question groups have already been defined for certain tests like the TUGK (Test of Understanding Graphs  Kinematics) or the FCI (Force Concept Inventory). However, you may define your own question groups. Enter a question group as commadelimited question numbers with no spaces. For example you might type for group 1 1,5,12,16 and for group 2 3,7,13,14 . Check the checkbox for each question group that you want to analyze as a unit.
Customized statistical functions
It's possible to modiify the default behavior of the program; however, it requires creating your own personal config file. Please email Aaron Titus at titus@mailaps.org if you are interested.
