University of Memphis
About ITD
Projects
Services
Facilities
Support
Learning
Resources
Security
A to Z List
Glossary
Spectrum
Deals and Discounts
Get Started.
Get Services.
Get Help.
Information Technology Division
Text Only Print    

Test Scoring / Scanning

Tests Scoring / Scanning Information

Table of Contents

Test Request Form

  1. Identification Section
  2. Statistical Analysis (total test or by section)
  3. Weight Factor Sheet
  4. Course Number
  5. Section Number
  6. Last Question Marked
  7. Number of Unmarked Questions
  8. Number of Secondary Answer Keys
  9. Test Result Sort Keys
  10. Histogram Distribution
  11. Inter-Item Correlation Matrix
  12. Number of Copies
  13. Base Score
  14. Form Numbers
  15. Login Number

Coding Control Sheets (Answer Sheets)

Introduction to Test Scoring Services

This manual contains information and instructions needed to use Information Technology's Test Scoring / Scanning service:

  • Information for filling out the Test Scoring Request Form.
  • Instructions for filling out the Control Sheets needed to process your sheets.
  • Common problems in filling out your Test Scoring Forms.
  • Explanation of Test Scoring Results and Statistical Analysis produced.

A Test Scoring Request Form must be completed and submitted with each request for test scoring. The information entered on the request form is used for three purposes:

  • Identifying your test.
  • Specifying the way you want your test scored.
  • Specifying any of the available options you wish to use.

Test Forms:

Test Scoring Requests Forms can be obtained in the I/O Room, AD 112 during normal.

Test Request Form

The following explains the use of each of the numbered items on the Test Scoring Request Form:

Sections that use the "square" fill-in areas should be filled completely with a your #2 pencil.

Section 1
  1. Complete this section with the identification information requested:

    * Include your VAX USERNAME in this section and we'll e-mail you the results when the test is done!
    Section 2

  2. An analysis is performed on the total test automatically. If the test is given to more than one section, you may request a separate analysis for each section by marking the box labeled TEST & SECTION.

    Section 3

  3. Mark the box labeled YES if there is a Weight Factor Sheet. A Weight Factor Sheet is used if some questions are to count more than others. A question with a weight factor of "3" would count three times as much as a question with a weight factor of "1". A weight factor of "0" means that a question is graded but is not to be used in determining a students grade. An instructor may wish to include such questions merely as an aid in teaching.

    CAREFUL ! - Entries for Sections 4 - 8 must be right-justified. Otherwise, the entry will be interpreted incorrectly.

    Section 4

  4. Fill in the Course Number. The Course Number is used for identification purposes and must match the course number on the Primary Answer Key.

    Section 5

  5. Fill in the Section Number of the course entered in Section 4. If more than one course section is being scored, then the lowest section number should be entered. It MUST ! match the entry made on the Primary Answer Key.

    Section 6

  6. Fill in the Number of the LAST QUESTION marked on the Primary Answer Key.
    NOTE: This number does not, necessarily, indicate the actual number of questions being scored.

    Section 7

  7. Fill in the total number of unmarked questions preceding the last question marked on the Primary Answer Key. Questions left blank on the Primary Answer Key are ignored by the statistical analysis.

    Section 8

  8. Fill in the number of Secondary Answer Keys. This item is used if a question has more than one correct answer. Specify only the number of Secondary Answer Keys. Do NOT count the Primary Answer Key in this total.

    Section 9

  9. You can have the test score listings sorted in any or all of the following orders:
    1. Alphabetical order by name.
    2. Social Security Number, (lowest to highest).
    3. Test score, (highest to lowest).
    A listing will be generated for each choice. Listing the test results by Social Security Number can be useful when scores are posted or when names are omitted. If this section is left blank the listing will be by Alphabetical Order.

    Section 10

  10. If you want a Histogram (Bar Graph) of the test score distribution, fill in the box labeled YES. Otherwise, fill in NO or leave blank. If a Histogram is requested, you will receive one for the total test. If the Statistical Analysis was requested by TEST & SECTION, a Histogram will be produced for each section.

    Section 11

  11. If you want an Inter-Item Correlation Matrix, fill in the box labeled YES. Otherwise, fill in NO or leave blank. The Correlation Matrix can ONLY be produced for the total test and NOT for each individual section. The Inter-Item Correlation Matrix measures how strongly one question is related to another question.

    Section 12

  12. Mark the number of copies of test results desired. If this section is left blank, you will receive one copy of the results.

    Section 13

  13. The default base score is 100. To specify a base score other than 100, fill in section number 13. The scores are computed according to this formula:

    SCORE = (NUMBER RIGHT / NUMBER OF QUESTIONS) * BASE SCORE

    If a base of 30 is specified on a 20 question text, a student with 13 correct will have a score of 20: ((13 / 20) * 30). If you request a Histogram, a base score that is a multiple of 10 should be used, since the Histogram divides the base into ten intervals.

    Section 14

  14. Fill in the Form Number. The Form Number is required when you give more than one test within a section. This form number must be the same as the form number on the answer key under GRADE OR EDUC. The following numbers should be used:
    A1 = 1 B1 = 4 C1 = 7
    A2 = 2 B2 = 5 C2 = 8
    A3 = 3 B3 = 6 C3 = 9


    Do NOT use the 0 in the second column. For each individual student, the form number is added to the right of the section. For example: students in section 12 taking a test with the form number A3 would have 12A3 listed under the section heading on the test result listing.

  15. Login number: leave blank. Test Scoring Services will complete this section.
name section

Control Sheets (Answer Sheets)

How to Code a General Purpose NCS Answer Sheet: (Form No. 4521)

  • Under NAME (top left), code AAAAA (left justified). This indicates that this is the Primary Answer Key.
  • Your department should be coded in the last four positions of the NAME section (right justified).
  • For multiple forms ONLY: The form number must be the same as on the Test Request Form. Enter under the EDUC section according to the following table:

    A1 = 1 B1 = 4 C1 = 7
    A2 = 2 B2 = 5 C2 = 8
    A3 = 3 B3 = 6 C3 = 9


  • The identification number must be coded 000099991 in columns A-I.
  • The course number must be coded in columns K-N. This must be the same as the course number entered on the Test Request Form, (Section 4).

    ID section

  • Code the section number in columns O-P. This must be the same section number entered on the Test Request Form, (Section 5).
    • Answers: Mark the correct answer to each question. If a question has more than one correct answer, you MUST use a Secondary Answer Key(s). No question may require a student to mark more than one choice for a given answer.
    • Questions to be omitted should be left blank. These will be ignored by the grading and statistical analysis process. If you want to include a question but not have it count toward the student scores, see the section on Weight Factor Sheet.

Coding a Secondary Answer Key

  • If any question has more than one correct answer, a Secondary Answer Key must be coded. No question may require the student to make more than one choice (i.e. "Mark the three most correct answers" is not an allowable question).
    Identification coding. All identification coding is the same as on the Primary Answer Key EXCEPT:
    1. NAME should be coded AAAAB for the second answer key, AAAAC for the third answer key, etc...
    2. IDENTIFICATION NUMBER (columns A-I) should be coded 000099992 for the second answer key, 000099993 for the third answer key, etc...
  • Use the Secondary Answer Keys to code alternate correct answers if a question has more than one correct answer. Mark only one alternate answer per question for each Secondary Answer Key. Do not mark questions which have only one correct answer. Use only as many of these Secondary Keys as necessary. The question having the greatest number of correct answers will determine the number of Secondary Answer Keys required.

Coding a Weight Factor Sheet

  • If all questions are not to be weighed the same, a Weight Factor Sheet is required.
  • Identification coding is the same as on the Primary Answer Key EXCEPT:
    1. NAME must be coded AAAAF, (left justified).
    2. IDENTIFICATION NUMBER, (column A-I), must be coded 000088888.
  • Answers: The Weight Factor Sheet should be used only if one or more questions has a weight value other than 1. The codes A through E will assign weight values of 0 through 4 respectively, to the questions. All questions left blank will default to a weight value of 1. Use the following table to determine the weight value you wish to assign each question:

    A = 0 Question not used in scoring exam.
    B = 1 Default or base weight value.
    C = 2 Weight value TWO times that of B.
    D = 3 Weight value THREE times that of B.
    E = 4 Weight value FOUR times that of B.


Coding a Section Header Sheet

  • If a TEST AND SECTION analysis was requested in Section 2 of the Test Scoring Request Form, then each section of a course, (after the first), must have a Section Header Sheet. The first section uses the Primary Answer Key for identification.
  • Identification coding. All identification coding is the same as on the Primary Answer Key EXCEPT:
    1. IDENTIFICATION NUMBER (columns A-I) must be coded 000077777.
    2. The Section Number (columns O-P) must be coded with the Section number of the Course Section that is to follow.

    NOTE: DO NOT fill in the Answer Key. The Section Header Sheet is only used for identification -- it separates sections within the same course.

Student Answer Sheet

  • Identification coding. All identification coding is the same as on the Primary Answer Key EXCEPT:
    1. NAME. Last name should be coded first, then a blank space and then the first name.
    2. IDENTIFICATION NUMBER should be coded with the student's Social Security Number, unless some other identification number has been assigned.
  • Answers: Only one choice per question is allowed. Students should use a #2 lead pencil and make reasonably clean erasures. NOTE: ABSOLUTELY NO INK PENS ALLOWED.

    close section

    Closing Sheet

    One Student Answer Sheet MUST be filled out with five ZZZZZ's in the NAME area of the sheet, (left-justified). The program running the tests will see this "closing page" and know that this is the end of this test.

Hints for using Test Scoring Forms

Test Scoring Services has recognized a number of common problems associated with test scoring forms. In an effort to minimize errors, please check all forms turned in for the following problem areas:

  • The number one problem occurs when filling in the sheets; please, always check twice.
  • Students must use a #2 lead pencil; ink is NOT to be used.
  • Students should properly mark the section number and if more than one test form is given within a section, the form number.
  • Students should properly mark their answer choices. The improper marking of answers causes an excessive amount of omissions to show on the student analysis. The instructor may wish to check the computer analysis against the Student Answer Sheet where excessive omissions are indicated.
  • The instructor should properly separate sheets of the different sections and / or test forms to be run. Any sheets that are submitted to Test Scoring Services as a group with an Answer Key will be scored with that Answer Key. If a student from an incorrect section or test form is submitted and run, the instructor may spot the discrepancy by noting an incorrect section number or test form number by the name of the particular student.
  • The instructor should properly mark the Answer Key. It is particularly important that the instructor mark the Answer Key accurately and complete the identification area in accordance with instructions provided. Failure to do this may cause a delay in getting the test results.
  • The instructor should be sure to submit Section Headers when several sections are to be scored together, and physically separate the sheets of one section from those of another section.
  • The instructor should properly mark the Test Scoring Request form when submitting tests for scoring.
  • The instructor should submit the forms in the following order:
    1. Test Scoring Request Form *
    2. Primary Answer Key *
    3. Secondary Answer Key(s) (if any are needed)
    4. Weight Factor Sheet (if needed)
    5. Student Answer Sheets (first section) *
    6. Section Header Sheets (if there are more than one section)
    7. Student Answer Sheets (for next section, etc...)
    8. Closing Sheet *
      * MANDATORY

Test Scoring Statistical Analysis:

  • Test Score Summary. The test score summary consists of the following:

    1. All scores that have occured for a particular test are listed in the first column.

    2. The second column simply lists the number of correct responses marked by the students to obtain the corresponding scores.

    3. T-scores are given for each score in the third column. The T-score is an indicator of how well a student does compared to the other students in his/her class. The T-scores always have a mean of 50 and a standard deviation of 10, regardless of the mean and the standard deviation of the raw test scores. Because of this, the T-score is independent of any particular group or test, so that it may be used for comparing scores of two different tests or sections of a course. The first section has a mean of 75 and a standard deviation of 11.6, while the second section has a mean of 87 and a standard deviation of 7.6. To determine what score a student must make in section 2 in order to rank the same as a person making an 80 in section 1, first it is necessary to find the corresponding T-score for 80 in section 1, which is 54. The instructor then finds the raw score corresponding to the T-score 54 in section 2, 90. Therefore, someone who scores an 80 in section 1 ranks the same as a person who scores a 90 in section 2.

    4. The frequency (column 4) of a score is simply the number of students receiving that score; it is given for each score listed in the score summary.

    5. The cummulative frequency (column 5) of a score is the number of students who have received a grade of less than or equal to that particular score.

    6. For each score, the cumulative percentile (column 6) is listed for a student who has received that score. It is computed by the formula:

      CUM. PERCENTILE = (CUM.FREQUENCY/NUMBER OF STUDENTS) * 100

    7. The number of students, mean, and standard deviation are listed at the bottom of each summary. The mean is merely the arithmetic average of the scores received. The standard deviation is a measure of variability of the scores about their mean. A large standard deviation indicates that the scores are spread out; whereas, a small deviation indicates that the scores are close together.

  • Histogram. (optional)

    The student score Histogram is a graph of the informataion in the test score summary that gives the instructor an overall picture of how students performed on a given test. On the score axis, the scores are divided into ten even groups, and the frequency for each group is computed by finding the total number of students receiving a score that falls in that group. On the frequency axis, the length of each frequency interval and the number of students per interval are completely flexible. The computer will use the combination of interval length and number of students per interval that will allow for the maximum size graph. This procedure may cause Histograms to be scaled differently from section to section. The following is an example of a test score summary (Figure 1) and its corresponding Histogram (Figure 2).

    Figure 1

    Test Score Summary

    Course - 3711 / Department - Math

    12 - 15 - 95

    SCORE # CORRECT RESPONSES T-SCORE FREQUENCY CUM. FREQ. CUM. PERCENTILE
    46 46 36 2 2 17
    53 53 44 4 6 50
    54 54 46 1 7 58
    66 66 59 2 9 75
    68 68 61 2 11 92
    71 71 65 1 12 100

    NUMBER OF STUDENTS = 12 MEAN = 58.1 STANDARD DEVIATION = 8.7

    FIGURE 2

    STUDENT SCORE HISTOGRAM

    COURSE - 3711 / DEPARTMENT - MATH

    12 - 15 - 95

    Student Score Histogram

  • Item Analysis.
    1. The item analysis consists of a question by question summary and analysis. By each question number there are six columns indicating the number of people who marked A, B, C, D, E, or omitted that question. An asterisk indicates the right answer.
    2. Next to these columns are two columns indicating each questions difficulty -- percent correct and percent correct with omissions not considered. (The latter is useful in gauging the difficulty when a large number of omissions are due to a lack of time rather than question difficulty.)
    3. Next to these columns is the net discriminant index. This index may be used to determine how well a question discriminates between upper and lower level students. The index is scaled from -1.0 to +1.0 with 0.0 as the poorest discriminator. A large positive index means that a good student is more likely to answer the question correctly than a poor student. On the other hand, a large negative index means that the reverse is true. A question with an index near 0.0 does not separate the upper and lower level students very well. In general, a question's power to discriminate is independent of its difficulty. The exception to this is that questions of medium difficulty will tend to have higher indexes than questions which are extremely hard or extremely easy.
    4. If a Weight Factor Sheet is used, the relative weights of each question are given in another column.
    5. At the bottom of the item analysis, the Kuder-Richardson reliability coefficient is listed. This is a measure of a test's internal consistency or homogeneousness for the group to which it was administered. The coefficient generally ranges from a low of 0.000 to a high of 1.000.
  • Inter-Item Correlation matrix. (optional)

    The correlation matrix measures how well each question on a test correlates with every other question, that is, how strongly two questions are related. For instance, if students who answer question 1 correctly generally answer question 2 correctly and students who answer question 1 incorrectly usually answer question 2 incorrectly, then one can say that there is a strong positive correlation between question 1 and 2. If, however, on that same test, students who answer question 1 correctly generally answer question 3 incorrectly and students who answer question 1 incorrectly generally answer question 3 correctly, then one can say that there is a strong negative correlation between questions 1 and 3. The correlation index ranges from the highest negative correlation of -1.0 to the highest positive correlation of +1.0. Coefficients near 0.0 indicate that two questions are independent of each other. An example follows:


    		 1 2 3
    1 1.0 0.8 -.9
    2 0.8 1.0 -.8
    3 -.9 -.8 1.0

    Notice that question 1 correlated with question 2 will always be the same as question 2 correlated with question 1. Because of this, only the lower triangle of values are listed in the test scoring matrix.


    		 1 2 3
    1 1.0

    2 0.8 1.0
    3 -.9 -.8 1.0

    The Inter-Item Correlation Matrix can be used in several ways. For example, it can be used to test whether knowledge of one item is related to knowledge of another item. It might also be used to test whether two basicically equivalent questions with different wordings are seen as equivalent by the students.

- E N D -
Last updated: 07/10/2008 16:15:05
Copyright © 2008 University of Memphis | Important Notice Contact Us | Got a Question? Ask TOM