EXAM SCORING REPORTS

1. Item responses of each student.  This list gives the keyed answers to the test and compares the answers of each student to these keyed responses.  This list is often useful for instructors who review test responses with students.  It helps instructors identify students who made specific errors on a test.

2. Test scores by student name.  This list is alphabetized by student last name, and it is often useful to teachers when they are entering scores in a grade book.  In addition to student last names, this list contains student ID numbers of up to 10 digits; student section-numbers and test-form numbers; raw scores on the total test and subtests; and each student's percentile and standard score for the total test.  The percentile score gives the percentage of the total class having a raw score lower than the student in question.  The standard score is a relative score that is unaffected by such factors as the size of the group tested or the difficulty level of items on the test.  It is calculated in the following way:  (a) the average score is subtracted from an individual's raw score and the difference is divided by the standard deviation.  (b)  This index is then multiplied by 100 and 500 is added to get rid of decimal and negative values.  The average of the set of standard scores is 500 and the standard deviation is 100.

3. Test scores in order of score.  Student names appear on this list not in alphabetical order but rather in order of test score--from highest to lowest.  This list is often useful to teachers when they are determining grade distributions.

4. Frequency distribution.  The frequency distribution summarizes the performance of the entire class.  This list may also be posted for students to see.

5. Item analysis.  The last section of the report contains information on the statistical quality of the examination.  This information includes, first, the number and percentage of students marking each of the possible response alternatives (preferably and usually five) for each item.  Good items are those answered correctly by a majority of students (but not an overwhelming majority); good items also have between 4 and 20 percent picking each of the four possible wrong answers
   
   Also included on the item analysis table is the percentage of correct responses on each item.  Questions answered correctly by about 90 percent or more are too easy and of relatively little value in distinguishing different levels of student achievement. Similarly, items that are too difficult, say those answered correctly by fewer than 30 percent (when there are four or five possible choices), also tend to add error to the assessment process.  This is because many if not most students getting such difficult questions right get them right by chance (i.e., marking randomly).  For accurate evaluation of student achievement, it is best to write items that between 50 and 80 percent of the class get correct.  An average score between 60 and 70 percent of the possible score usually indicates that questions of appropriate difficulty have been asked. 
   
   Another statistic reported for each item is the discrimination index.  It gives the correlation between item scores and scores on the rest of the test.  A positive discrimination index of .2 or above for an item indicates that the item was answered correctly by students who tended to do well on the test in general.  Items with negative indices should be examined carefully; students who answer such items correctly tend to do poorly overall on the test.