Test scaling features allow you to create test scores for examinees and save the results in your data table. You can choose between a sum score, average score, Kelley regressed score, percentile rank, or normalized score. A sum score is the sum of item scores for the items selected in the dialog. Missing item responses are counted as zero points when computing a sum score. An average score is the sum score divided by the number of items completed by the examinee. As such, an average score ignores missing data and computes the score as an average score for the completed items. A Kelley regressed score is an estimate of the true score that adjusts for measurement error. If reliability equals 1, the the sum score and Kelley’s regressed score are the same. As reliability decreases, the average score for the entire group of examinees is counted toward an examinee’s Kelley score. Percentile ranks are the percentile rank of an examinee’s sum score. This method of computing percentile ranks treats data as ordered integers. Finally, a normalized score is the value from a standard normal distribution that has the same percentile rank as the sum score. Normalized scores have a mean of 0 and a standard deviatoin of 1 by default. All scores, except percentile ranks, can be linearly transformed to a new scale using the Linear Transformation panel of the dialog.
- Start the Linear Transformation dialog by clicking Transform > Test Scaling.
- In the upper part of the dialog, select all of the items that you would like to count in the score.
- In the section labeled “name,” type a name for the variable that will store the scores in the data table.
- Use the Type Combo Box to select the type of score you would like to compute.
- If no further options are needed, click the Run button to execute the analysis. A new variable will be added to the data table using the name you provided. This new variable will contain the scores for your examinees. Summary statistics and a score conversion table will be displayed in the interface.
- To apply a linear transformation such that the scores have a particular mean and standrd deviation, type numeric values in the appropriate text field in the Linear Transformation panel.
- You can specify a minimum and maximum score value by providing information in the Constraints panel.
- The Precision text field allows you to indicate the number of decimal places that you would like to impose on the scores. For example, typing 0 in the Precision text field will round scores to the nearest integer.
- Click the Run button to execute the analysis. A new variable will be added to the data table using the name you provided. This new variable will contain the scores for your examinees. Summary statistics and a score conversion table will be displayed in the interface.