Hacker News Reader

Comment by educationdata | Hacker News Reader

It is strange that the study does not report raw numbers of instance. How many happened after post card release vs. not after post card release? I can't find the raw numbers of such. Only raw number of breaking down by race.

Also, "Of the 167906 calls to the child abuse hotline for children aged 5 to 11 years, 2017 (6.7%) were verified as physical abuse cases". Only 6.7% verified???

View on HN

>Only 6.7% verified???

I have complaints about the quality of the study, but this is not one of them. Keep in mind that anyone in Florida who suspects abuse is required to report it to the hotline. Probably more significant is the more traditional mandated reporters, who are trained to report any suspicion of abuse, out of both an abundance of caution and to protect themselves from liability. By design, such hotlines have a high false positive rate.

I am not blaming the high false positive rate (although itself is problem). The problem is that this kind of high false positive rate has a huge impact on the quality of the study. Be aware the followed analysis does not account for the error in reporting and verification. There are simple too many noise.

I wonder if it could just be that the child abuse workers can more easily verify cases over the weekends. Maybe they need to talk to parents (who are likely working other days) before they can mark it as verified?

This is such an obvious explanation that it makes me doubt the whole study.

> Only 6.7% verified???

Verified "as physical abuse cases", meaning that the rest were deemed to not be.

Could be. Then the study should report how many were verified as abuse, how many were physical abuse.

Reply on news.ycombinator.com