I have found a link to the recent survey and I have one question: what are these users' languages? Do they speak English? The result is rather meaningless if all 5 users are English speakers, as they obviously will look for English Wikipedia only and do not need any secondary links, but it would be much more interesting to get a point of view of someone who is not an English speaker and who can be looking for a secondary link.
Topic on Talk:Wikipedia.org updated page layout
The participants in this survey research were ones that offered us their names and email addresses as part of a survey that was conducted on the Wikipedia.org portal site. They were all willing participants and we did not make any determination on who would participate in the research, only that they have freely given us their contact information and were ok with signing the release form, so that we can publish their comments and concerns.
We made no effort to select participants based on language, nor did we ask them what languages they spoke, read or write.
I can guess from the fact that all comments are in English, screenshots are from English Wikipedia and only wiki mentioned is enwiki that all 5 users where English speakers. If this is the case, checking links to non-English Wikipedias on users who don't use them is not really meaningful.
What I am referring to is how one of the participants commented on this change: "I wouldn't use this very often. Before I'd click on random languages... how is it now?". This means that this person is less likely to use secondary links even for fun. At the same time other people use them for finding wikis in languages they speak, and I don't think we want them to be less likely to find these wikis.
@NickK, you might be interested in reading this analysis of the A/B test for the discussed change (Edit: oops, I see in a topic below you may have already read this!). While the survey contained English-speaking participants (I'm not sure we collected information on first/second language or nationality) the analysis used a random sampling of visitors. The analysis echos the survey - that the change resulted in positive interaction with the portal.
@CKoerner (WMF): I already commented on this survey below (or above, or in parallel... can't properly refer to another thread in Flow, sorry) in #Use of the drop-down menu. Basically the result seems to be that users who don't use these links (in particular those most interested in English Wikipedia) indeed get a more positive experience, while users interested in these links are significantly (4.7 times) less likely to use them with a new layout. As an editor in a language which is likely to be in a dropdown menu for many readers I cannot qualify this result as a positive one.
All language links will still be accessible, for fun or otherwise; some will be in the top ten links around the globe in the browser's preferred languages and some will be in the language dropdown.
I apologize for not having more participants that used languages other than English, but that was not primary focus of the research we were doing at that time, with those participants.