I'm evaluating uTest for our company and so far am very pleased with what I'm seeing.
We are single-sourcing between print and HTML and would like to be able to have all our question types supported in both media. In our print documents, we support the following test types:
2) Long Answer
3) Short answer (fill-in)
4) Multiple choice
5) Multiple answer
The first five are handled nicely by the applets, but the other two appear to have no counterparts.
For clarity, Matching would be:
Match the following terms with their definitions.
A. term1 __ Definition 1
B. term2 __ Definition 2
C. term3 __ Definition 3
Prioritization would be:
Number the steps of the following process in their order of completion:
__ Step 1 description
__ Step 2 description
__ Step 3 description
__ Step 4 description
It is possible to implement each part of a matching or prioritization as an individual question, but there doesn't appear to be a way to have a single question consist of the multiple parts.
You have done something similar with uTestAnswer10.class where multiple images are part of a single question, so I was wondering if the same is possible for these two additional types.
Each could be accomlished by using the same pick-list of numbers (prioritization) or letters (matching) for each component. The tricky part would be to get the individual parts to be scored as a single question.
Yes, both "matching" and "prioritization" question types can be handled. See pages 3 and 4 of the following demo as an example:
Those do require each option to be treated as a separate question... but you can change the "points" of each question such that they add up to the desired question weight.
OK, I see now how prioritization and matching questions can be built.
Now for today's related issue:
In order to have prioritization questions and matching questions not mess up the test numbering, I am thinking of creating an algorythm to move the answers to higher numbers outside of the range of question numbers.
For example, if Question 4 is a matching question with 5 parts, instead of assigning the answers to positions 5-10 of the answer sheet, the algorythm will assign them to questions 141-145 and enter them on the answer sheet as such.
My questions are:
1) Can the answer sheet have gaps, or must the numbers be contiguous (and present)? e.g.
2) Do the numbers have to be in order? i.e. is this ok:
Will either of these impact the scoring or any other aspect of usage.
As a followup, can the answer key have non-numeric entries? for example:
Again, I'm trying to accomodate those priority and matching questions.
In regard to having non-numeric question numbers.... Unfortunately, No. At this time, all question numbers must be integers.
The algorithm that is used to dynamically create the HTML for the answer sheet will automatically display the answers in increasing numeric order. Such as:
If any numbers in the test are skipped (ie. by the user not answering a question or by simply not having an applet for a question), then those answers will simply be left blank on the answer sheet. For instance, if a 10 question test doesn't have an applet for question #8, then the answer sheet will still be displayed as:
Note that the order in which the applets are placed on the test HTML pages... and the order in which the user answers the questions... does not affect the display of the answers on the answer sheet. They'll always be displayed in increasing numeric order.
The only exception to the above rule is that Long Answer and Essay questions will always be displayed at the bottom of the answer sheet. That is, suppose a test has this sort of question types:
Questions 1 - 5 are multiple choice
Question 6 is a long answer type
Question 7 and 8 are true/false
Question 9 is an essay type
Question 10 is a short answer type
The answers would then be displayed on the answer sheet with the "Long Answer" and "Essay" questions at the bottom: