I can barely get my head around why this (retracted dome) is or could be perceived as an issue or something likely to give rise to errors. I have to agree with Leigh, above. Have errors actually been observed through using the flashmeter this way? If so, how were the results qualified as directly related to the collar around the retracted dome? Or are we boxing at shadows?

I am particularly intrigued by this:

In such cases, the Sekonic and Gossen would seem to inject a lot of error into the measuring process.
Speculation is one thing, but results from actively experimenting should answer the question.