I am working on a project for scoring and reporting student tests. The data will be imported from a text file (generated from a scantron form). Once imported, it will be scored (checked against a key). Each test question must be identified to address a particular state mandated skill (objective). I need to be able to report the data in format like (+ is a correct response):

Question #____1___2___3___4___5___6___7___8___9

John Doe_____+___A___B___+___+___+___A___B___+

Jane Doe_____C___+___+___+___+___A___+___B___+

I have looked at other ("professional") solutions that use a flat file. For each question there is a correct response, student response, state objective, and a summary field for reporting. The average test will have 60 questions. In this format each record will be comprised of approximately 300 fields. It would take days to create all of the fields, not to mention placing them on the layouts for reports. If I were getting paid by the hour, this might be OK but I'm not.

I decided to create a couple of relational files to simplify the design of the project. I have a test_key.fp5 file and a student_responses.fp5 file that are related (one(test_key.fp5) to many(student_responses.fp5)) based on a question_ID field. test_key.fp5 has the following fields; question_ID, correct_answer, and state_objective. student_responses.fp5 has student_ID, student_response, and score (calculation). In this format student_responses.fp5 will become very large (8000 students take 4 test with 60 questions each = approxomately 1,920,000 records) but the database is only made up of a few fields. Additionally, with this type of format, I am having trouble creating the type of report I want (see "Columnar report with headers and summaries" in "The Right Brain" forum).

How should I design the solution? Ideally, I want a solution that will not take me weeks or months to design, creates the reports I need, and runs somewhat efficiently.

I appreciate any help you can provide.

I think the related file approach that you have taken is good. The fact that student_responses.fp5 will end up with a lot of records isn't a problem. Consider the difference between a long skinny file and a short wide file. Either one has to hold the same amount of information.