Calculating trends and averages on skewed data. Any help would be greatly appreciated!

[ranopano]

I'm running analysis on an Audit that has been distributed over the course of a year. There are currently about 40 or so different audits. Each audit contains 90 questions.
When the audit was first published, the results we were getting were skewed. For instance, many of the answers came back at a 100% success rate, but over the course of a couple months, that number dropped and is now averaging anywhere from 50-90% success rate, depending on the question.

I'm trying to determine the trend for these values and see how we have performing over time. However, because the data was so skewed initially, I'll get results stating we are performing 50% worse than when we started. I don't think this is an accurate representation, so I was hoping to get some feedback from you guys to see if there is a better way of calculating this.
Below is a sample table that kind of illustrates what I am referring to.[TABLE="class: grid, width: 500"]
<tbody>[TR]
[TD="align: center"]Date[/TD]
[TD="align: center"]Question[/TD]
[TD="align: center"]Score[/TD]
[/TR]
[TR]
[TD]1/1/15[/TD]
[TD]1[/TD]
[TD]100%[/TD]
[/TR]
[TR]
[TD]1/8/15[/TD]
[TD]1[/TD]
[TD]100%[/TD]
[/TR]
[TR]
[TD]1/15/15[/TD]
[TD]1[/TD]
[TD]89%[/TD]
[/TR]
[TR]
[TD]1/22/15[/TD]
[TD]1[/TD]
[TD]56%[/TD]
[/TR]
[TR]
[TD]1/29/15[/TD]
[TD]1[/TD]
[TD]57%[/TD]
[/TR]
[TR]
[TD]2/5/15[/TD]
[TD]1[/TD]
[TD]55%[/TD]
[/TR]
[TR]
[TD]2/12/15[/TD]
[TD]1[/TD]
[TD]59%[/TD]
[/TR]
[TR]
[TD]2/19/15[/TD]
[TD]1[/TD]
[TD]55%[/TD]
[/TR]
</tbody>[/TABLE]

Note, this is only an example. The actual data has 90 different questions and has 40+ different audits.

So as you can see, the initial 100% values are very skewed (this has to do with testing procedures).
How would I determine things such as trends and averages with this? The 100% values will have a noticeable impact on my results, so how would you figure out where the data normalizes?


Appreciate any feedback!