« first day (2 days earlier)      last day (19 days later) » 

Nick Ulle
Nick Ulle
01:20
@TheDemonLord to claim that ideas about implicit bias are "not derived from a statistical analysis" is misleading. Implicit bias has been studied extensively by cognitive scientists and psychologists over the past 30 years, typically through statistical analyses of human subjects. One of the most well-known examples is the implicit association test (IAT). It would be more accurate to say that there is still much debate about how to interpret these results
 
1 hour later…
JJrodny
JJrodny
02:49
Agreed, @NickUlle. @TheDemonLord: if you don't believe that implicit bias is real, and you're open to learning, I have good news! You don't have to run any experiments yourself to prove it's non-existence. Hundreds of talented PhDs in Cognitive Science and Psychology have all already put in the leg work to create complex experiments trying to take into account every other possible factor to rigorously study the effect of implicit biases.
If you really are open to learning more about your world, I strongly encourage you take a look at a few of them, they're really interesting! Simply Google Search Implicit Bias Review: scholar.google.com/scholar?q=implicit%20bias%20review
This is a good one, cited 1200 times about the effect of implicit biases in the medical field: bmcmedethics.biomedcentral.com/counter/pdf/10.1186/…
@FrankHopkins I agree that corporate trainings may not be able to cover everything in the past 30 years' research into implicit biases, but I'm sure there are some smart PhDs out there who can explain it better than a corporate HR department.
I'm hopeful as time goes on we'll be able to reference a good set of literature on implicit biases that are more easily digestible than having to read published papers from Google scholar, but for now, take a look there! There's lots of hard experiments lots of smart people did to prove whether or not it's real.
Frank Hopkins
Frank Hopkins
03:19
@JJrodny thx for the links but it seems no one claims there is no "bias" or associations at all. my point is also not that a training needs to include all the background research, though having references for people who would like to look deep and a compelling summary can probably be helpful. My issue would more be with
a) rather than telling people they have biases a training would be helpful if it helps them identify biases with negative effects and giving them methods to avoid that - and most of those methods do not to be tailored towards certain groups or particular group trainings, take for instance blind interviewing methods or CV methods. Offering height adjustable desks, voice to text tooling etc. pp.
and b) that in the current case it's unlikely subconscious bias will play a big role
and/or can be countered more efficiently by OP by addressing potential issues upfront rather than asking whether there was some training, that may or may not have addressed stereotypes people may have about autistic colleagues.
and c) that trainings that go for stereotypes might risk reinforcing those stereotypical thinking. I.e. my point in that regard is that a training can be positive or negative, and isn't a good metric to look for in itself. Now OP might address that by digging deeper but they might also just start looking for more generic metrics rather than get hung up on particular …
TheDemonLord
TheDemonLord
@NickUlle - Not at all. The theory was derived from an initial belief that people harbored unconscious prejudices and that these manifested in discriminatory behavior. Then as a post-hoc 'proof' they used reaction times when presented with different stimuli to prove that this is true. There is no statistical evidence that the different reaction time results in discriminatory behavior - Add to the fact the 'Training' seems to have the opposite effect reinforces that this is bunk.
Frank Hopkins
Frank Hopkins
03:44
that meta study also seems to mix subconscious biases as measured by association tests and stereotypical treatment; the latter seems reasonable to lead to worse treatment since it represents professional misbehaviour (using stereotypes rather than proper statistics combined with individual medical history/context, though both are somewhat related^^).
The few IAT examples mentioned I've brushed over so far from that paper don't convince me that bias is the cause and not just a symptom (e.g. of different socialisation clashing).
and at least that stereotype part would also be a good argument that trainings that provide proper statistics and tooling to base decisions on to reinforce proper diagnosis methods would probably be more efficient than trying to tell doctors that they are biased. Yes, well done that can also help if part of it is providing data that challenges their stereotypical thinking. But imho medical treatment is also a muddy field anyway, because you inherently need to use statistical generalized data.
I..e you need to take gender, age etc. into account to know how likely a treatment will have what effect and thus decisions can easily be skewed by statistically unsound data etc. That's not something you would do for regular hiring processes etc.
TheDemonLord
TheDemonLord
@JJrodny - Most of the studies referenced in that study I would categorize as 'Junk' - small sample sizes, the assumption method (which presents an extreme scenario) - other of the larger samples (several thousand) appear to be self-selected and the conclusions are dubious. For example - claiming that a preference against psychiatric patients is due to unconscious bias as opposed to a conscious bias/experience would seem to be a ridiculous claim, would it not?
 
2 hours later…
JJrodny
JJrodny
06:02
It's important to be able to learn, and change our preconceived notions based on facts presented to us. By learning and admitting we're wrong, we grow. I know how hard it can be when presented with facts that run counter to our belief structure.
There's a whole body of literature on how facts don't change minds, especially politically related ones, which are formed based on emotions. Take a look at this article, which is very interesting. Maybe it can help us all grow and be open to learning something new snopes.com/news/2022/08/11/cognitive-bias-research
TheDemonLord
TheDemonLord
06:42
@JJrodny - Everything you've said, could equally be applied to you - let me ask you one simple question: Assume for the moment that all the literature on Unconscious Bias is correct - if this is the case, then it should stand to reason that Unconscious Bias training would improve peoples behavior. Unfortunately the literature on this is that at best, it does not improve it and at worse actually increases discriminatory behavior. How do you explain that?
Furthermore - The link between a difference in reaction time measured in Milliseconds between and In-Group and an Out-group is missing several causal links to get from mere reaction times to manifest behavior. Lastly - As someone who does genuinely enjoy Science and the Scientific method - seeing what passes currently is lamentable - the Grievance Studies Affair, in particular, is the best exemplar of what I mean.
 
9 hours later…
Frank Hopkins
Frank Hopkins
15:23
@JJrodny are you saying that bias trainings cannot work because people won't change their minds based on facts? ;) Or that they need to be manipulative rather than factual due to that? ;) Indeed some opinions are hard to change, but that argument seems to be more than doublesided in this discussion^^ (note this comment is made with an amused semi-serious undertone as I'm aware it was not the point you wanted to make, but the argument can easily be applied that way...^^)
@TheDemonLord do you have any example or an aggregation of literature about the effectiveness of bias training, given you argue that there is literature that shows it does not improve behavior.
 
4 hours later…
TheDemonLord
TheDemonLord
19:13
@FrankHopkins - The main one being: A Meta-Analysis of Procedures to Change Implicit Measures - the Conclusion at the end remarks "we also found little evidence that changes in implicit bias translated into changes in explicit bias and behavior, and we observed limitations in the evidence base for implicit malleability and change."
There were small changes in the short term with some methods, but didn't change peoples overall behavior and didn't stick in the long run.

« first day (2 days earlier)      last day (19 days later) »