Iteman requires two files for input: the control file and the data files; both are described below. The approach is inspired by statistical analysis software, which often has a data tab/variables tab.
In the future, Iteman will run the report directly on data from tests delivered inside Assess.ai, and there will be no need to format input files. However, you will still have the option to use Iteman by uploading input files.
Examples of these files are included for download - Add Link? ASC recommends that you use the files as a template rather than recreating the files from scratch.
The Control File
The control file is a list of metadata for each of the items in your tests. An example is shown below.
There are six columns, all of which are required.
- Item ID: This is the name, ID, or accession number of each item; it will be used to label items in the output.
- Key(s): Correct answer(s) as A,B,C,D if unscored multiple-choice; 1 if items are scored dichotomous, such as open-response items scored as correct or incorrect; For polytomous items: + if positively scored or – if negatively or reverse scored.
- Number of alternatives. For a 4 option multiple-choice item, this is 4. For items that are scored 0 or 1, the number of options is 2. However, we recommend re-coding pre-scored multiple choice items so that 1=correct and 2=incorrect, as starting at 0 is typically reserved for partial credit items.
- Domain or content area: unique domain label, such as Biology or Quadratic. Maximum is 50.
- Inclusion status: Y = Yes (included in the analysis); N = No (not included); P = Pretest (calculate the stats but do not count towards score).
- Item type: M = Multiple-choice items with responses that begin at 1 or A. For scored multiple-choice items, see P below; R = Rating scale items: polytomous items with responses that begin at 1 or A); P = Partial credit items with numeric responses that begin at 0 (e.g., 0, 1, 2, 3). This includes multiple-category partial credit items, and dichotomously scored multiple-choice items (scored 0 or 1).
The Data File
The Data File is the file that contains the responses by each examinee to each item. Responses can be alphabetical (A,B,C,D… or a,b,c,d…) or numerical (1,2,3,4), where A = a = 1, etc. It follows the standard approach of rows/examinees and columns/items or observations.
This file contains three initial columns that are used in the analysis and reporting.
- Examinee: This is the ID number or name of each examinee to be used in the scoring output file.
- Group: Use this to define groups, which might be useful to analyze. For example, you might want to find descriptive stats of performance for males/females. Advanced users can use this to define differential item functioning.
- Inclusion Code: As with items, you can set this to Y to analyze every examinee, but, if you want to leave out some examinees (e.g., those that did not finish the test), you can simply set this cell to N rather than deleting rows from the data set, which can be cumbersome for large datasets.