First – please say which census you refer to, as each one can use different ways of suggesting entries. Its normally based on the definitions from the original census forms.
However, you did mention 1921 the other day, so I have assumed that one. 😊
I can’t answer the question “deliberate or not?”. That sheet had been around so long (before kiwitrees I think) I don’t even recall who provided the definitions.
I can tell you that particular census is coded to provide “M”, “W”, or “-“. Same for both males and females. The calculation also requires MARR, DIV, or _NMR family tags; and they must have dates ( for checking against the census date).
As per above, I can’t tell you why “-“ is used instead of “S” for unmarried people.
Sorry. Didn’t even think to mention 1921 Canadian. Didn’t know it’s legacy. No problem. was funny having young children married . The Canadian census site (Archives Canada) doesn’t give any recap like Ancestry or FamilySearch to review.
----- [updated: 06 Feb 2022] MacOS: 10.15.7 (Catalina) Safari 14.1.2; Firefox 96.0 PHP 8.0 FastCGI; mySQL 5.x
Alter-Drukarsh Connections 3.3.10b/the Garelicks <family members only>
The code is very simple, not much to go wrong. So if it is suggesting children are married incorrectly I would recommend you first double check your data. If they married later, is that marriage date wrong?
If they are OK, give me a link and I’ll look further into it.