Thanks for that interview. The relevant part is around 45 minutes in, give or take.
Unfortunately, I don't find Richard Arnold's recollection to be credible in this interview, at least without further corroboration. There are numerous reasons why, including:
1. In this interview, Arnold implies that Roddenberry did rewrites on "Family" in season four, something that's directly contradicted by Moore in
this interview with Ronald D. Moore and Ira Steven Behr.
2. Arnold doesn't even say what Roddenberry supposedly did to the script of "Watchers." So, there's nothing really to address there, besides the vague claim that Roddenberry did something to it.
3. However, Arnold does imply that Roddenberry is specifically responsible for removing the sexual encounter between Data and Ard'rian in "The Ensigns of Command." However, according to Memory Alpha, this is contracted by a recollection by the episode's director Cliff Bole given in the
Star Trek: The Next Generation Companion 2nd ed on page 101, where Bole says that the reason that the sexual encounter was removed from the episode was because the budget for the episode was cut at the last minute [
http://memory-alpha.wikia.com/wiki/The_Ensigns_of_Command_(episode)].
4. The idea that Roddenberry would have a problem per se with Data being sexual in "Ensigns," which without any nuance is how Arnold recollects the situation, is hard to believe, given not only Data's behavior with Yar in "The Naked Now" but also the fact that Data's line about being fully functional was lifted out of Roddenberry's own
The Questor Tapes.
5. Roddenberry already had a history of mis-characterizing how other writers portrayed principal characters in the context of creative conflicts. The most famous is his version of his objection to an early draft of "The City on the Edge of Forever" when Roddenberry stated incorrectly that Harlan Ellison had Scotty dealing drugs. This is a problem if Arnold is basing what he understands to have happened on what Roddenberry told him, and that issue would potentially apply to every episode discussed in the interview here.
6. There are other, differing accounts circulating the web about the development of "Ensigns," including why Snodgrass used a pseudonym on a draft. These accounts have nothing to with contributions from Roddenberry. People can find them just as easily as I did, but I won't repeat them here, since I cannot corroborate them.
This doesn't mean that Arnold is wrong about "Watchers," but without corroboration his claim that Roddenberry made any contribution to the script is not something that can be credibly accepted just because Arnold said so in this interview.
It's also worth mentioning that what Moore said in the interview I mentioned in point #1 is consistent with the idea that Roddenberry made changes to episodes before "Family," but that remark isn't specific enough to glean anything about "Ensigns" or "Watchers."