#OldDataFallacy - Proposed New Concept

[Oxpiu]

Hi, My name is Philip. I am new to this forum. My background is mathematics and statistics and I did some philosophy studying in university around 20 years ago. I am proposing a new fallacy name today: #OldDataFallacy.

#OldDataFallacy - #ConfoundingBias argument where pioneers of a new perspective of knowledge get challenged with a #ConfirmationBias argument from advocates of a traditional data approach that portrays the novel approach as dangerous based on old data sources. The pioneers advocate a need to transition to a new path with different statistical indicators and new databases validating the new path as safer. What I am hinting at is that what is dangerous for one generation is safe for another as technology matures.

This can be very frustrating as the pioneers in one generation often become the luddites holding the next generation away from growth. Often there is a lot at stake including health outcomes and financial resources. My passion is to help make #ConfoundingBias issues easier for people to understand as many researchers don't engage with the mathematics behind confounding bias. In it's most simple form confounding bias is essentially based on differentiation skills and I am arguing that these are often underrated skills in data analysis.

[FlashDangerpants]

Can you find a layman's example for this? I hope it doesn't involve excluded middles, I rather like my middle, it's possibly my best bit.

[Impenitent]

local surgery to remove the bullet as opposed to amputation...

-Imp

[Oxpiu]

Layman's Example

Business Model for Winning Baseball Games in the movie Moneyball known as Sabermetrics

In 2002, Billy Beane challenged the Baseball industry using the Sabermetrics business model to win baseball games and was considered crazy by people in the baseball community in 2002. When the Oakland Athletics baseball team started winning a lot of games, the validity of the business model was proved to work in practice and not just in mathematical logic. The impact was so huge that it was put in a movie. Today, sabermetrics is a lot more respected in the baseball community. Sabermetrics refocused the baseball industry to different statistical indicators than traditional statistical indicators that baseball scouts focused on.

[FlashDangerpants]

Layman's Example

Business Model for Winning Baseball Games in the movie Moneyball known as Sabermetrics

In 2002, Billy Beane challenged the Baseball industry using the Sabermetrics business model to win baseball games and was considered crazy by people in the baseball community in 2002. When the Oakland Athletics baseball team started winning a lot of games, the validity of the business model was proved to work in practice and not just in mathematical logic. The impact was so huge that it was put in a movie. Today, sabermetrics is a lot more respected in the baseball community. Sabermetrics refocused the baseball industry to different statistical indicators than traditional statistical indicators that baseball scouts focused on.

OK, I think we have that over here now, football fans in England have started talking about some very confusing statistics such as XG for expected goals that are a result of applied datascience in some manner.

So this fallacy of old data, how does that work? Presumably it does something other than looking cooler if you hashtag it? Is it a formal or informal type of fallacy?

[phyllo]

Hi, My name is Philip. I am new to this forum. My background is mathematics and statistics and I did some philosophy studying in university around 20 years ago. I am proposing a new fallacy name today: #OldDataFallacy.

#OldDataFallacy - #ConfoundingBias argument where pioneers of a new perspective of knowledge get challenged with a #ConfirmationBias argument from advocates of a traditional data approach that portrays the novel approach as dangerous based on old data sources. The pioneers advocate a need to transition to a new path with different statistical indicators and new databases validating the new path as safer. What I am hinting at is that what is dangerous for one generation is safe for another as technology matures.

This can be very frustrating as the pioneers in one generation often become the luddites holding the next generation away from growth. Often there is a lot at stake including health outcomes and financial resources. My passion is to help make #ConfoundingBias issues easier for people to understand as many researchers don't engage with the mathematics behind confounding bias. In it's most simple form confounding bias is essentially based on differentiation skills and I am arguing that these are often underrated skills in data analysis.

Doesn't sound like a fallacy. It sounds like people use existing data and analytics until the new stuff is shown to be better.

That seems to have also happened in your baseball example ... "validity of the business model was proved to work in practice and not just in mathematical logic".