Food for thought about data, interpreting research, and misinformation

Joined
Apr 22, 2018
Messages
3,430
Location
MA
My economics professor studies income inequality and a couple of years ago there was serious concern about gender inequality and he was asked by the university to study income inequality at the university and report his findings to a committee. First, he set out to determine the major factors that influence faculty salaries. The major factors were: 1. Time - the longer you were there the more you made; 2. Highest degree obtained - PhDs made more than masters-level faculty; 3. Specialty - some academic disciplines paid better than others; 4. Additional administrative tasks resulted in additional pay and; 5. Gender

He was asked to report preliminary findings to the committee and immediately there were faculty who wanted to run with the idea that there were inherent biases in pay between men and women. Dr. Bob is an excellent researcher, he knew exactly why he was asked to look at this issue, and he wasn't satisfied that this was in fact reality. Driving home one day he was thinking about the problem and he realized that when he started at the university in the late 1980's fewer than 10% of the faculty in the business school were female. He also realized currently >50% of the faculty in the business school were now female which is a change that has really just happened in the last 15-20 years. When he controlled for time at the university, gender was no longer statistically significant, yet every other variable remained significant. This also made intuitive sense to him since there were still a significant number of male professors who were hired before the push to hire more women and the average tenure of male faculty was significantly longer than the average tenure of female faculty. Gender bias WAS an issue but it was addressed 20 years ago and any remaining differences in income was just leftover from that period of time. Matter-of-fact, when you looked at pay between male and female faculty hired in the past 10 years, there seems to be a gender bias towards women making more than men.

I have a research background in immunology which can be very complicated and very nuanced. It is impossible for someone who does not do that day after day to really understand this nuance. Since I've been out of that field for 20 years I no longer understand the nuance. It's the reason we have "experts" and it's the reason no one can be an "expert" in everything. Every day I see lay people pretending they understand the nuance of complicated situations, having never seen the raw data, and really being ill-equipped to understand it even if they did. Despite my best effort, I do this myself at times. When we do this we are the committee members above who don't really understand the situation but want to run with the first result that fits our preconceived ideas. We as a society need to do better and stop pretending we understand things we really do not because it is pure chaos when we do this on a mass scale. Admittedly, while much of this is self-inflicted, we as a society need to make changes that result in increased public confidence in our "experts".
 
You cannot do a social study that is double blind like the medical industry. Income inequality is a tricky subject as you know, things take time to manifest. I am surprised it took your professor so long to realize that.

Break down the pool to finer pools and laser focus on those, and you might (not will, just might), find that today's new grad in the same field (don't compare teachers with engineers) and you may find that women have similar or more pay than men.

What may set them apart in the future is the family commitment compromise. If you look at women and men with no kids into the 40s you may find them with similar income, but if you look at men with housewives they may be higher paid than men with working wives and kids, and women in similar situations.
 
You cannot do a social study that is double blind like the medical industry. Income inequality is a tricky subject as you know, things take time to manifest. I am surprised it took your professor so long to realize that.

Break down the pool to finer pools and laser focus on those, and you might (not will, just might), find that today's new grad in the same field (don't compare teachers with engineers) and you may find that women have similar or more pay than men.

What may set them apart in the future is the family commitment compromise. If you look at women and men with no kids into the 40s you may find them with similar income, but if you look at men with housewives they may be higher paid than men with working wives and kids, and women in similar situations.
I'm not sure how you came to some opinion about the temporality of what I posted? I gave you the quick and simple version for clarity. This could've been an hour of work or week or a month...no idea.
 
Did your professor happen to write an official paper or study about it? I'd be interested in reading it.
 
As my usual short comment, watch the people who say "follow the science" very closely, because the usually are the people who aren't.
This is exactly my point as the alternative is just as untennable - don't follow the science? This crisis of confidence will snowball to the point where opinions mean more than actual reality and the new reality will be whatever someone wants it to be x 8 billion.
 
I'm not sure how you came to some opinion about the temporality of what I posted? I gave you the quick and simple version for clarity. This could've been an hour of work or week or a month...no idea.
"Driving home one day he was thinking about the problem and he realized that when he started at the university in the late 1980's fewer than 10% of the faculty in the business school were female. "

Driving home one day, this should have been realized before the study started.
 
This is exactly my point as the alternative is just as untennable - don't follow the science? This crisis of confidence will snowball to the point where opinions mean more than actual reality and the new reality will be whatever someone wants it to be x 8 billion.
I think it can be distilled down to:

Real scientists shouldn't need to say that.

Real science, real data doesn't need anyone to force it on others.
 
I can only speak for me. I do thank you for posting. I'm 50 years old and a few decades ago scientists seemed to have greater transparency and maybe I was naive but I trusted findings and thought/felt that despite who or where the funding came from; results were construed in an honest manner.

Current-era scientists are increasingly political and the future generations of scientists are increasingly ideological. IMO ideology and pure science do not mix. Leads to fudging of the data to prove a predetermined outcome. That predetermined outcome is often what the funder expects and at times is so bold to announce it.

Most recent example I know of is the pending assault on natural gas for indoor cooking. Study was funded by a group that outright claims to want to do away will the source altogether, globally. I could have missed but how would the study be complete without mention of a mitigating factor such as proper ventilation? A proper, unbiased study should at least mention this caveat and seek funding to determine the effect if proper ventilation is used. But that is against the grain of what the funding entity wishes to demonstrate, no??

So many modern studies show clear correlations when read but don't quite prove causation. And from my armchair, it would seem to me that many bought and paid for studies intentionally demonstrate correlation and purposely dismiss causation so that the study may be weaponized. No political side here as I'm claiming it's done by all sides in some circumstances.
 
the major factors that influence faculty salaries.
By law, a state university will hold to a strict pay scale based on years of seniority, degree level held, whether or not tenure has been granted, and various other non-gender attributes. Private universities also likely institute a similar system to avoid claims of discrimination. So someone of any gender in the same position will be on the same pay.

The question is are there institutional blocks to reaching a high position that may be gender unequal.
 
By law, a state university will hold to a strict pay scale based on years of seniority, degree level held, whether or not tenure has been granted, and various other non-gender attributes. Private universities also likely institute a similar system to avoid claims of discrimination. So someone of any gender in the same position will be on the same pay.

The question is are there institutional blocks to reaching a high position that may be gender unequal.
Yup
 
I think science should somehow adopt a pro sports model for the major global issues, teams of scientists could work on;

climate change,
AI,
genetic modifications of plants, animals, and humans,
loss of genetic diversity and habitats,
and solutions to the unsustainability of our current growth rate in resource consumption.

The public would get full transparency on research topics and who is funding it.

And just like in pro sports, for the 99.99% of us who don't really know what is actually happening in the field, we get a week by week analysis and explanation by a group of recognized experts in each of the fields.

Right now in science reporting we have the equivalent of guys who barely made the starting line up in high school basketball, talking like they are Shaq on Inside the NBA, and no one is calling them out on their lack of knowledge, extreme bias, or BS line they are passing off as fact.
 
For over a decade, I did was a "Senior Engineer" doing statistical process quality control at a multi-national company. It was my job to study whatever needed studied in our facility and within my scope of operations, and also write/run DOEs (design of experiements). I was not the sole person doing this; I was one of many.

I can attest that "science" is misunderstood by just about everyone except those who actually did the study.
I can also attest that often "science" is biased (on purpose or not; does not matter).

But, there's always a story to tell with data. It's just a matter of how well the data was collected, processed, and presented. You cannot eliminate bias, but you can account for it and try to reduce it's effect on the study. What we cannot control is the bias of the observer who reads the study.
 
Back
Top