July 23, 2014 at 8:10 pm
Comments posted to this topic are about the item Data Decisions or Instinct?
July 24, 2014 at 3:02 am
In relation to this question I would recommend everyone read at least the first section of 'Thinking, Fast and Slow' by Daniel Kahneman. It goes into great detail about the way our instinctive and rational thought processes interact with a particular emphasis on judgement and decision making.
The short version is that while instinct can offer valuable insight from the qualitative, rather than quantitative realm it is not to be fully trusted. Instinct will often substitute a much simpler question for the one you are actually asking e.g. 'Is this product any good?' with 'Do I like the salesperson?'
The book also emphasizes just how poorly even statisticians are at thinking in terms of probabilities.
The overall message is take a look at where your instincts want to take you, but ensure that the data backs it up, because your emotions/instinct are too quick to jump to incorrect conclusions.
July 24, 2014 at 3:06 am
The more I have learned about cognitive biases the more I wonder how anyone has ever got anything right. However much you think you have taken that step back and thought it through these biases will always play their part in how we interpret the data and come to whatever conclusion, at least beyond a certain point of obviousness. That point of obviousness always seems to be passed quicker than you think, too.
July 24, 2014 at 3:38 am
I think that we should be right to consider our instincts. As opposed to using that to override statistical analysis we should use it to question the analysis done and to drive further queries. This way we provide quantitative results which have considered the qualitative aspects too. To ignore the facts are wrong but so is to ignore our questioning of them too.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
July 24, 2014 at 6:18 am
Steve,
Does this really apply to data professionals? I have been a programmer and data provider since 1982 and I have never been asked to interpret the data. It was always someone else's job to do that. I provide the data, the reports, and the Excel spreadsheets. I think that it is dangerous for the same person who is providing the data to be making decisions based on it. I just have to make certain that the data is correct and ahs not been corrupted, and, that it is reported correctly.
As far as how people make those decisions... I have seen both ends of the spectrum. One manager I know hardly reviews any data goes by emotion and gut feel. Another one pores over data for a very long time before coming to decisions. Amazingly, both tend to get it right. There are others who want to use data to justify their preconceptions and get angry when the data is "wrong" (different from their preconceptions). Most people are somewhere between the extremes.
Thanks,
Tom
July 24, 2014 at 6:30 am
People including leaders constantly ignore the facts. "This country has dangerous weapons, we must invade!", "So and so says the right words about an emotional issue, so I'm voting for them.", "I don't care what the numbers say, climate change doesn't effect me...!", and so on...
I had someone tell me that complex applications needed to be hosted in the "cloud" because that's cheaper. They totally didn't want to hear the details of Stack Exchange, which buys their own servers to power one of the internets busier sites, because it's actually cheaper. Or they use a combination of Microsoft and OSS technology, because they are smart and pragmatic.
Once you learn that most humans are irrational, then you learn to deal with it. Or you may exploit the situation for your benefit, if that's your style...
July 24, 2014 at 6:41 am
RP1966 (7/24/2014)
In relation to this question I would recommend everyone read at least the first section of 'Thinking, Fast and Slow' by Daniel Kahneman. It goes into great detail about the way our instinctive and rational thought processes interact with a particular emphasis on judgement and decision making.....
Good book
....
The overall message is take a look at where your instincts want to take you, but ensure that the data backs it up, because your emotions/instinct are too quick to jump to incorrect conclusions.
If a person can learn from experience, and Kahnerman examples were Grand Master Chess players, Fire controllers, then yes intuition based on valid experience is very useful and can short cut the process. But, and it is a big but, where there is not verifiable experience then people can make mistakes, and then use data and false arguments to support their intuition all with good intentions.
July 24, 2014 at 6:52 am
We should certainly look to data to guide us and perhaps even justify our decisions, but we can't forget that the human brain is still an important part of any computational exercise. We need employees that you use their judgement, in collaboration with data, to make the best decision for our organizations.
I agree.
While stories are great, I think the story should complement the data. People should be able to use their judgement and make informed decisions and many times that may involve relying on the gut feeling.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
July 24, 2014 at 7:07 am
Data should inform our decisions, but data rarely sparks truly creative thoughts. Those 'leaps' often seem to come out of nowhere.
Our instincts are not just random junk. They are subconscious evaluation processes tuned my millions of years of evolution, particularly in the area of dealing with other humans (whether others in an organization, or customers (marketing) or competitors) friends or enemies. Various experiments have found that we can be remarkably accurate in our evalutation of others, even though a popular theme of movies is the person 'everyone is wrong about'.
The best judgement comes from a synthesis of honest data and instinct.
...
-- FORTRAN manual for Xerox Computers --
July 24, 2014 at 7:22 am
Steve, the way you describe it I think you're confusing ideas with evidence.
Data is never going to tell you what to do directly. For that you'd need an idea.
Once you have the idea the next part is collecting the evidence to tell you its a good one.
If you have several 'competing' ideas the evidence can then tell you which one is the best of the bunch.
Without the story line the collected data is just that: collected data. (e.g.if you had a database of the location of every atom in the Universe it still wouldn't tell you a whole lot about the Universe).
Chris.
July 24, 2014 at 7:47 am
From a slightly humours point of view, often there is quite a resemblance to Chinese Whispers in Data Collection, Processing and Analysis processes, which in turn makes one at least question any decisions or findings based on the "data".
😎
July 24, 2014 at 8:04 am
Chris Cradock (7/24/2014)
Data is never going to tell you what to do directly. For that you'd need an idea.Once you have the idea the next part is collecting the evidence to tell you its a good one.
If you have several 'competing' ideas the evidence can then tell you which one is the best of the bunch.
The problem with ideas is that they are going to impose a narrative on the data. If you are not careful you can impose a false narrative by collecting the wrong data, after data collection you also have the same problem in that a false narrative can be imposed by selective interpretation of the data.
It can be the difference between:
'Can we get data to prove X?' and 'What is this data telling us?'
You said: 'Once you have the idea the next part is collecting the evidence to tell you its a good one.'
Often you are better to look for data that will prove your idea is a bad one - try to challenge your hypothesis rather than support it.
July 24, 2014 at 9:38 am
OCTom (7/24/2014)
Steve,Does this really apply to data professionals? I have been a programmer and data provider since 1982 and I have never been asked to interpret the data. It was always someone else's job to do that. I provide the data, the reports, and the Excel spreadsheets. I think that it is dangerous for the same person who is providing the data to be making decisions based on it. I just have to make certain that the data is correct and has not been corrupted, and, that it is reported correctly.
Maybe some, maybe not others. I've often worked with business people to identify patterns and understand data. Sometimes it's a back and forth as I find business people I've worked with often haven't understood how to interpret large sets of data, or how averages, counts, distribution, and other statistical measures impact things. On the other hand, I don't always understand the importance or impact of a particular piece of data, and I can work with them to better understand how to structure things.
As an example, years ago (far too many, like 18), we were trying to determine how to plan for inventory of a few products. Demand can vary, but the salesman suspected some broad patterns. However it wasn't quite apparent from the data. I would look at patterns, go talk with them, get feedback on what they saw based on my patterns, and eventually we could make a decision on how to predict a portion of our demand (70% or so). Then the ordering people would need to apply their knowledge for the rest.
July 24, 2014 at 9:46 am
SQLRNNR (7/24/2014)
While stories are great, I think the story should complement the data. People should be able to use their judgement and make informed decisions and many times that may involve relying on the gut feeling.
Yes, though the gut needs some data support. Not 100%, but no 10% either. I'd hope for a preponderance of support, at least 60-70% supported by data.
July 24, 2014 at 9:55 am
OCTom (7/24/2014)
Steve,Does this really apply to data professionals? I have been a programmer and data provider since 1982 and I have never been asked to interpret the data. It was always someone else's job to do that. I provide the data, the reports, and the Excel spreadsheets. I think that it is dangerous for the same person who is providing the data to be making decisions based on it.
Tom, I agree that no decision should be made by the one who gathers and supplies the data without the work of others. However, in the decision process the one gathering the data is often crucial to the evaluation of that data. Any filtering, aggregation, normalization, and other valid or required manipulation of the data to become information needs to be explained to those pouring over it.
I have also been in the business for some time, and I have gathered both data and decision packages. I have often been involved in the decisions or at least in the discussions that led to the decisions. That should not be rare, but common.
But I have never been nor will I ever be the only one reviewing the data and making the decision. After decades of being in this and like business I have, like you, considerable knowledge of the processes, practices, validity, and quality of the data. When we were a rookies making a report, we followed certain instructions and did as we were told. Now after the years of experience I would hope that we are exercising the knowledge we gained concerning the business that our contribution is far greater, and our opinion considered. That is what we do.
M.
Not all gray hairs are Dinosaurs!
Viewing 15 posts - 1 through 15 (of 20 total)
You must be logged in to reply to this topic. Login to reply