We have had the same experience with KMx and we have experimented with substitute values and they have fixed the problem you described. We too, have tried a mix of different test question types. Basically, we added a manual advance button on each page that modifies each test question variable to a placeholder like * or x. This will eliminate the navigation problems you describe and the test should score fine. The weird part is the SCORM data that results. Here's what we know about the impact on SCORM data output from Lectora when modifying the question variable manually with * as a default for no answer:
True/False: Generates cmi.interaction.student_response data line and *
Multiple Choice/single answer: Does not generate cmi.interactions.student_response data line or value
Multiple Choice/multiple answer: Does not generate cmi.interactions.student_response data line or value
Matching: Generates cmi.interaction.student_response data line and ,
Drag and Drop: Generates cmi.interaction.student_response data line and ,
Hot Spot: Does not generate cmi.interactions.student_response data line or value
Hope this helps.
