Home  Computer Based Assessment Register for Testing Our Team Clients Contact us

EPEC specialises in scoring of tests, data analysis and other psychometric services. These are detailed below.

  Answer Sheets and Marking
  none Although there is an increasing interest in online testing, paper-and-pencil tests are still widely used. EPEC can design your own answer sheets which include your logo and specifications or provide generic answer sheets. After administration of tests the high quality optical scanners can scan thousands of answer sheets in a very short time after which EPEC staff manually verify the electronic data to ensure, for example, that missing responses are indeed missing and not due to lightly shaded answers. Likewise double responses are verified to ensure that two answers were indicated rather than one answer not rubbed out properly.
  Scoring and Analysis
  Scoring can (unfortunately) have different meanings. At the lowest level scoring means the number of questions answered correctly or the sum of all the marks given to answers. Unless a standardised test with known psychometric properties (such as norms) was used, such (raw) scores can be very deceiving.
Although a summary of scores can be given as a mean, mode or median and the spread of scores can be expressed as a standard deviation, a test taker can challenge the result in terms of the test’s validity and/or reliability. One general argument is that if different questions were asked, the outcome could have been different, especially by test takers who marginally failed.
Especially in high stakes testing it is important that results are based on a sound foundation. An item (question) analysis reveals each item’s difficulty and quality. Poor items can be identified and removed from the test for scoring after which an index of the test’s reliability can be calculated and evaluated. Tests with high reliability will yield precise measures of performance. By taking out poor questions, the scoring of test takers’ performance is not only more defensible, but also more precise.

Different measurement models can be used for analysis:
            Classical Test Theory (CTT): has been used for over a century in traditional or conventional assessment delivery and is used for for item- and test analysis, scoring and reporting
            The Rasch Framework and Item Response Theory (IRT): used for detailed psychometric analysis, construction of tests and scales, linking, equating and reporting
            Option Probability Theory (OPT): used for unique performance and diagnostic assessment and reporting

EPEC specialises in all of these measurement theories and have applied and used them extensively.
  Psychometric Services
  EPEC has already assisted many organisations in a wide range of psychometric services, including:

        Item banking: EPEC staff has more than 30 years’ experience of item banking and management. EPEC’s sophisticated systems allows for many types of questions, such as; multiple-choice, multiple-response, short answer, essay, Likert and fill-in-blanks. Through the powerful editor, any character, symbol, table, image or video can be added. Items are stored with full history, version, author, state (new, reviewed or approved), key words, various statistics and other information.

        Data analysis: EPEC has specialised knowledge and extensive experience in data analysis using Classical Test Theory, Rasch Measurement Theory and models of Item Response Theory for multiple-choice, extended matching, short answer, clinical, surveys and other formats.

        Test development: In addition to EPECs wide range of tests, EPEC has the capacity, knowledge and experience to develop new tests to suit your requirements.

        Test construction: Whether a fixed length traditional test or an adaptive test is required, EPEC can assist with constructing an optimal test for your purposes meeting blueprint requirements. EPEC uses information functions to select items that yield maximum information where required, for example, around the cut-score. Parallel tests can be constructed to match the same blueprint if security, time zones or other issues are of concern.

        Test delivery: We can deliver tests in paper-and-pencil format or online as traditional, adaptive or diagnostic based on EPEC’s own measurement theory called Option Probability Theory; see Computer Based Assessment page.

        Standard setting: EPEC has already assisted many clients with standard setting exercises using Angoff, Beuk, Hofstee, Objective and other methods. For longitudinal programs, EPEC can ensure through linking and equating that the same standard is maintained over time; irrespective of differences in cohorts and/or difficulties of different examination papers.

        Workshops: EPEC can be contacted to do workshops at any required level.
  Please contact us for more information.