The paper, titled “The link between low-stress bicycle facilities and bicycle commuting,” reports that protected bike lanes see about 1.8 times the number of bike commuters than standard bike lanes do and 4.3 times as many as blocks without bike lanes.

So doing the math, even an unprotected bike lane is about a 2.4 times improvement over no bike lane at all.

I find that to be an interesting contrast compared to a survey done in my city as part of a trails planning process, where planners asked respondents whether they "would recommend" use of bike infrastructure based on type. In that, only 8% of respondents would recommend using an unprotected bike lane and 14% were unsure, while 56% would recommend a bike lane with flex posts and 20% were unsure. In other words, respondents were between 3.5 and 7 times more likely to "recommend" use of a protected bike lane than an unprotected one.

I had suspected that that disparity in perception was greatly inflated compared to the actual difference in amount of use, and I'm glad to now have actual science to point to to back up that hunch. Thanks!