There’s an important distinction between non-partisan elections and primary-election reforms, such as Alaska’s adoption of the Top-4 primary (followed by RCV in the general election). Because this distinction is sometimes lost in discussions of reforms like that in Alaska, I wanted to explain the distinction.
In non-partisan elections, the candidates appear on the ballot without any party identification or association. Although Progressive Era reformers thought non-partisan elections were a “good government” reform, which would encourage voters to vote based on the characteristics of individual candidates, political scientists are highly critical (rightly so, in my view) of non-partisan voters. Contrary to romantic Progressive Era views, many voters need the party cue as an important filter to help inform their vote. That’s especially true in the United States, where the individual voter is asked to vote in far more elections than in other democracies. For this reason, most political scientists who study elections consider non-partisan elections to be bad policy.
In contrast, in primaries like CA’s Top-2 or Alaska’s Top-4, candidates can list themselves on the ballot as “prefers Republican Party” or “prefers Democratic Party” or “independent” or “non-partisan.” Voters therefore have the cue of the party a candidate prefers to be associated with. It is true that these are not the candidates of the Republican or Democratic party, because they have not been chosen by the party or its voters. But the way the candidates identify on the ballot gives voters important partisan information that voters lack in non-partisan primaries (indeed, dissenting Justices in the Supreme Court case upholding Top-2 primaries thought this structure was unconstitutional precisely because voters would associate the candidates with specific political parties).
Thus, Top-2 or Top-4 primaries are not non-partisan primaries. They are ones in which the candidates identify the parties with which they want voters to associate them. To be sure, there are empirical questions it would be interesting to explore about whether voters identify these candidates as strongly with the relevant political party as they do candidates who emerge from traditional party primaries (though as just noted, dissenting Justices feared voters would identify these candidates too strongly with the political parties with which the candidates identified themselves). But these primaries are not the non-partisan elections that the Progressive Era extolled and that are still with us for some local government elections.