(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
 . . . . . . . . . .



Maybe we lost because of racist algorithms [1]

['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.']

Date: 2024-12-18  

My wife lives in the same HH, in Pennsylvania, as I do, is as “ educated” and is the same race as I am so the only defining characteristic that would flag her for gotv is her gender. Yet, I never received ANY gotv texts or calls. I did at least get one Trump oriented Text btw so this tells us something interesting about who they thought they could convert(and I am a registered Dem which in PA is publicly available). She received DOZENS of texts and calls and emails. We had a door knock too and her name was first.

Certainly, it is logical that she has more propensity to vote Harris so should be a priority over me. Priority though given diminishing returns would suggest maybe 2 or 3x as much contact (and I’m being generous) as me but I got NONE. So this got me thinking, who else as a voter demographic has historically lower turn outs or less propensity to convert for a vote for Dems? Latinos and African Americans come to mind. We had lower turn out of non college AAs and Latinos. College education is positively associated with turn out. It occurred to me, knowing a bit about how these algorithms can work from my digital marketing and analytics experience that if someone didn’t think too hard and just wrote the code to look for “high propensity” voters to cycle contacts and didn’t put some cap on frequency or some progressive correction on consecutive hits etc. that what could easily happen is the higher propensity (more elastic) voters would get hit a billion times and the lower propensity or “convertible” voters—according to the algorithm—much less maybe even none. Certainly my wife was contacted way beyond any meaningful marginal benefit. So I know that much is true and have heard from others as well. So if I am ignored who else might have been?

Ok so this pure speculation, however it is speculation based on real issue that could have played a role that I have seen again and again in my professional life. Algorithms are not objective, they reflect the biases of their designers or the data they train on (or more often, both). So if they were not trained carefully they may have wasted contacts and therefore dollars saturating some individuals and leaving out others who should have been contacted. I’m not sure I’d ever be given access to see these to verify :) but I’m pretty confident this is at least part of the story given other clear fails from the campaign in adjacent areas. Everything I have seen—granted the best stuff they have is probably kept under wraps, but I doubt this would change my mind—suggests really naive and simplistic contemporary tactical executions. There were election fundamentals playing a role here as you can see in my other articles as well as others on the site but we need to fix everything, big strategy and tiny tactical details.

[END]
---
[1] Url: https://www.dailykos.com/stories/2024/12/18/2289274/-Maybe-we-lost-because-of-racist-algorithms?pm_campaign=front_page&pm_source=more_community&pm_medium=web

Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/