Douglas Adams joked about this phenomenon by describing that people will vote for lizards they hate so long it means that the wrong lizard won't win. In a zero-sum game, that result is a Nash equilibrium.

It is still a lizard that get voted into office at the end of the day.

None of this is responsive to my comment. Hillary Clinton did not win in a squeaker over Sanders; she crushed him. The voters made a clear choice, and it wasn't for "the right kind of lizard". This is the Democratic primary we're talking about, and it wasn't a "lesser of two evils" race there.

If you are asking why Clinton won over sander, while sander has a higher opinion rating, feel free to write your suggestion. Popularity opinion is shared between the left and right while same can't be said when talking about the vote for party candidates.

Some would point to the political events that the dc leaks talked about.

> If you are asking why Clinton won over sander, while sander has a higher opinion rating, feel free to write your suggestion.

In favorability or approval (whatever that means in the context of a nominee) ratings? They mean different things, and aren't always directly related[1]. I could definitely see how Sanders would be liked more, but people would believe he's less likely to do a good job.

1: http://www.gallup.com/poll/14797/presidential-approval-vs-fa...

To suggest an explanation myself, voting where majority vote win has a different result that voting where highest ranking wins. When there is extreme high and extreme low opinion about choices, there is commonly a third option which is rated higher in average.

Say I rated taco, pizza and Italian salad as 10, 8, and 2, and you rated them as 2, 8, and 10, we can easily see that pizza wins even if neither of us has it as our primary pick. pizza in this case has highest opinion rating while the other two choices are quite far behind.

I don't dispute that, but it's not what I was referring to. Some polls ask what people think of the president (favorability ratings), and some are more specifically geared to asking about the job the president is doing (approval ratings). The article I linked to from Gallup points to situations where they do not always correlate as you would expect. For example, Bill Clinton in 1993 and 2000. In 1993 he had 59% favorability and 49% approval,and that just about reverse by 2000 when he had 60% approval, but only 46% favorability. People disliked him after the scandal, but they also conceded he was doing a good job.

So, what I wonder is what questions are actually asked regarding candidates? There isn't any data on them doing the job itself yet (unless they are an incumbent, which none are this cycle), so are they asking about how much they like the person, like the policies, or how confident they are the candidate can perform well? The answer to that question may lead to different likely reasons for the question at hand. Sanders is a likable guy. I like him. I'm also not convinced his policies, as expressed on the campaign trail, are feasible. Conversely, Hillary Clinton isn't very likable. I don't really like her. I do agree with many (but not all) of her policies, but most importantly I think her policies are feasible and have a chance of being implemented if she's elected. I am, in fact, the exact sort of person that might have expressed a favorable opinion for Sanders but voted for Hillary.