It could simultaneously be true that including a good salary range increases the number of applicants who are good fits for the role, and also true that including a good salary range decreases the average quality of candidates.

One way you might see that effect is if both candidates who are good fits for the role and candidates who are bad fits for the role apply more often when a high salary is posted, but the number of bad fits increases faster than the number of good fits (e.g. because there is some subset of people who will send their application to every role that pays over a certain threshold whether or not they are qualified).

If it's sufficiently costly to distinguish qualified from unqualified candidates, the company might be better off not showing a salary range, even accounting for how it causes good people not to apply. That approach does feel like an inelegant hack to get around their inability to easily tell whether someone would actually perform well in the role though, so addressing that root cause would be better in that scenario if they could figure out how to do it.

Another, simpler option:

When you put a range of $100-120k, the very high quality candidate who won't accept an offer less than $150k doesn't apply.

Or, put another way:

> It could simultaneously be true that including a good salary range increases the number of applicants who are good fits for the role, and also true that including a good salary range decreases the average quality of candidates.

Part of "good fit" could be "willingness to accept compensation in range". If the job range says $10/hr, the average quality of applicants will go down because the MIT Ph.D.'s won't apply-- but you weren't going to hire them at $10/hr anyways.

I think parent understands your point, but, it's implied that the salaries were high, so your example doesn't fit.

> Our salary offerings are very aggressive to the developer's benefit

Your example covers the case of "mid-range salary", overqualified applicant doesn't want the job, but wasn't going to get hired anyways because overqualified. Parent's example covers "very high salary", underqualified applicant wants the job just because the number is high. Given the "salary offerings are very aggressive", then why would we be talking about a "mid-range salary" case.

Not who you're asking, but I take "very aggressive to the developer's benefit" with a mountain of salt, because everytime I've come across someone saying that the eventual comp was extremely poor and they were either deliberately lying or completely out of touch with the labor market.

Every time I have a recruiter explicitly tell me that they are "top of market value comp" and keep reiterating it, I already know there's going to be no stock or bonus and the total comp is probably 1/5 of what I get now and "can't" do a signing bonus. I have not had one that doesn't meet this benchmark yet.

The ones that actually pay top of market rate I reach out to, not the other way around.

Yes, that is definitely a possibility. My point was "the average quality of candidates goes down when you post a salary range" can be true even if the salary range is very good.

It's possible, but I don't know that we should rush to believe that's the case based on an anecdote of a single flawed experiment when there's simpler explanations.

(Especially since it's only a problem if the higher pay motivates more bad applicants that are hard to distinguish from good applicants. If it encourages 100 people who have no relevant experience to apply, that brings the average down, but it doesn't increase the probability of making a bad hire.)

It does however increase the level of effort to find the wheat.

You make a good point, except it's moot in this particular case. There is no mention of averages. The best candidates simply responded to the one with no salary range posted.

> The ads performed equally well in regard to total responses with the better candidates responding to the ones without salary ranges