There are several contexts in which the test scores may not perfectly reflect reality.
When rspec runs multiple tests globbed together, the individual results files are not
written out; only the summary. That means that the individual tests will only sum up to
the numbers in the summary if you run each test individually first.
Individual tests might contradict each other. For example, one lab might ask students
to manage a user, and the next might instruct them to remove that classification.
A student may choose to perform the tasks in the lab in a slightly different fashion.
For example, if we ask them to make a training user, they might instead
manage a testing123 user. They've still effectively completed the tasks
set before them.
Please use your own discretion when interpreting these numbers.