EA Concepts: Share Impressions Before Credences

Hello, read­ers! I’m tri­al­ing for the Con­tent po­si­tion at CEA; as such, I’ve been asked to draft a cou­ple of posts for the con­cept map. Th­ese are meant to be close to the cur­rent style (no links in body text, fairly con­cise).

I’d love to hear your feed­back on this post. Spe­cific ques­tions:

1. What are your fa­vorite words for “be­liefs be­fore up­dat­ing on out­side in­for­ma­tion” and “be­liefs af­ter up­dat­ing on out­side in­for­ma­tion”? We’re try­ing to draw that dis­tinc­tion with “im­pres­sion” and “cre­dence”, but those may not be the best op­tions.

2. When you imag­ine this from the view of a reader who is newish to EA, and clicked on a link to read about the im­por­tance of “shar­ing your im­pres­sions”, does it make sense? Is it clear why this con­cept is useful

3. Are there any other links we should add to “fur­ther read­ing”? (In par­tic­u­lar, I think that a link to the Soviet ex­am­ple of “ev­ery­one hates the gov­ern­ment but is afraid to say so” might be rele­vant, but I couldn’t find a good ar­ti­cle sum­ma­riz­ing the ex­am­ple.)

Thanks for your help! The other con­cept draft is here.


Share Im­pres­sions Be­fore Credences

When we think through a ques­tion by our­selves, we form an “im­pres­sion” of the an­swer, based on the way we in­ter­pret our ex­pe­riences. (Even if you ex­pe­rience some­thing that oth­ers have also ex­pe­rienced, what you take away from that is unique to you.)

When we dis­cuss a ques­tion with other peo­ple, we may up­date our “im­pres­sion” into a “cre­dence” af­ter up­dat­ing on their views. But this can in­tro­duce bias into a dis­cus­sion. If we up­date be­fore speak­ing, then share our up­dated cre­dences rather than our im­pres­sions, our con­ver­sa­tion part­ners partly hear their own views re­flected back to them, mak­ing them up­date less than they should.

Con­sider two friends, Aaron and Max, who are equally good weather fore­cast­ers. Aaron has the im­pres­sion that there is a 60% chance of rain to­mor­row. He tells Max about this. Max had formerly had the im­pres­sion that there was an 80% chance of rain to­mor­row, but he up­dates on Aaron’s words to reach a cre­dence of 70%.

Aaron then asks Max for his view. Max tells him he thinks there’s a 70% chance of rain, so Aaron up­dates to reach a cre­dence of 65%. Both friends used the same de­ci­sion al­gorithm (av­er­age both prob­a­bil­ities), but be­cause Aaron shared his im­pres­sion first, and Max shared a view that “re­flected” that im­pres­sion, Aaron failed to up­date in the same way as Max.

This dy­namic ex­plains why it can be im­por­tant to share your ini­tial im­pres­sions in group dis­cus­sions, even if they no longer re­flect your up-to-date cre­dences. Do­ing so helps all par­ti­ci­pants ob­tain as much in­for­ma­tion as pos­si­ble from each par­ti­ci­pant’s pri­vate ex­pe­rience.


Fur­ther Read­ing:

Kawa­mura, Ko­hei, and Vasileios Vlaseros. 31 July 2014. “Ex­pert In­for­ma­tion and Ma­jor­ity De­ci­sions”.