In today’s tech-driven economy, data is essential for gaining new insights, making decisions, and building products.
In fact, there is so much data out there, that the quantity of it is doubling every two years—and by 2025, there will be 175,000 exabytes of data in existence.
This is an unprecedented figure, and it’s hard to put into perspective. To give you some sense, a single exabyte is equal to 1,000,000,000 GB of data, and five exabytes has been said to be roughly equal to “all of the words ever spoken by mankind”.
Common Fallacies With Data
As you can imagine, digging through all of this data can be quite the challenge.
Data comes in many different forms and not all of them are easy to analyze. As a result, it is tempting to take shortcuts with data, or to try and fit the incoming data we receive into our pre-conceived notions of how things ought to be.
15 Common Data Fallacies
How do we avoid painting a bullseye around the arrow, so that we can interpret the meaning of data in a logical, consistent, and methodological way?
The key is to understand common mistakes that people make with data, and why these errors skew our interpretations.
Examples of Fallacies
Here are four examples of fallacies, and why each is considered a faux-pas by data scientists.
1. Survivorship Bias
When people analyze the qualities it takes to be a successful entrepreneur, we typically look at the existing population of established entrepreneurs for clues. However, by limiting our sample just to this “surviving” group of entrepreneurs, we run the risk of survivorship bias.
There are lessons we can learn from all of the entrepreneurs who have failed—they are just much harder to find. Integrating that data into the story can help complete a much fuller picture.
2. False Causality
Did you know that there is a 95% correlation between the marriage rate in Kentucky and the amount of people who drown each year from falling out of fishing boats? (See it, and other bizarre correlations here)
Does this mean that there is some sort of relationship between the two variables?
Finding a high level of correlation can happen simply by chance—but awarding false causality is one of the most amateur statistical mistakes in the book.
3. The Gambler’s Fallacy
If the roulette wheel turns up black for 26 times in a row, does that mean that it will revert back to red on the next spin?
It’s easy to say that the odds don’t change, but imagine being in the moment. The Gambler’s Fallacy happens with data analysis as well: just because something happens unusually frequently over a period of time doesn’t mean that nature will “even it out”.
4. The Cobra Effect
Data can be used to measure progress in achieving business goals, but what if there is incentive to game these goals?
Wells Fargo, in an effort to upsell existing clients, introduced an incentive called “eight is great”. In short, their employees were encouraged to sell eight accounts per customer, which could take the form of credit cards, savings accounts, and other financial services.
In an example of good intentions gone awry, Wells Fargo employees began breaking the rules to meet their targets. Millions of unauthorized credit card and deposit accounts were opened based on this perverse incentive, and the bank was eventually ordered to pay a $142 million settlement.
How Much Do Americans Trust the Media?
Media trust among Americans has reached its lowest point since Trump won the 2016 presidential election.
How Much Do Americans Trust the Media?
Media trust among Americans has reached its lowest point in six years.
Gallup began its survey on media trust in 1972, repeating it in 1974 and 1976. After a long period, the public opinion firm restarted the polls in 1997 and has asked Americans about their confidence level in the mass media—newspapers, TV, and radio—almost every year since then.
The above graphic illustrates Gallup’s latest poll results, conducted in September 2023.
Americans’ Trust in Mass Media, 1972-2023
Americans’ confidence in the mass media has sharply declined over the last few decades.
|Trust in the mass media||% Great deal/Fair amount||% Not very much||% None at all|
In 2016, the number of respondents trusting media outlets fell below the tally of those who didn’t trust the media at all. This is the first time that has happened in the poll’s history.
That year was marked by sharp criticism of the media from then-presidential candidate Donald Trump.
In 2017, the use of the term ‘fake news’ rose by 365% on social media, and the term was named the word of the year by dictionary publisher Collins.
The Lack of Faith in Institutions and Social Media
Although there’s no single reason to explain the decline of trust in the traditional media, some studies point to potential drivers.
According to Michael Schudson, a sociologist and historian of the news media and a professor at the Columbia Journalism School, in the 1970s, faith in institutions like the White House or Congress began to decline, consequently impacting confidence in the media.
“That may have been a necessary corrective to a sense of complacency that had been creeping in—among the public and the news media—that allowed perhaps too much trust: we accepted President Eisenhower’s lies about the U-2 spy plane, President Kennedy’s lies about the ‘missile gap,’ President Johnson’s lies about the war in Vietnam, President Nixon’s lies about Watergate,”
Michael Schudson – Columbia Journalism School
More recently, the internet and social media have significantly changed how people consume media. The rise of platforms such as X/Twitter and Facebook have also disrupted the traditional media status quo.
Partisans’ Trust in Mass Media
Historically, Democrats have expressed more confidence in the media than Republicans.
Democrats’ trust, however, has fallen 12 points over the past year to 58%, compared with 11% among Republicans and 29% among independents.
According to Gallup, Republicans’ low confidence in the media has little room to worsen, but Democrat confidence could still deteriorate and bring the overall national reading down further.
The poll also shows that young Democrats have less confidence in the media than older Democrats, while Republicans are less varied in their views by age group.
GDP7 days ago
Visualizing U.S. GDP by Industry in 2023
Brands2 weeks ago
Ranked: Fast Food Brands with the Most U.S. Locations
Economy2 weeks ago
Visualizing 30 Years of Imports from U.S. Trading Partners
Markets2 weeks ago
Ranked: The Biggest Retailers in the U.S. by Revenue
Globalization2 weeks ago
The Top 50 Largest Importers in the World
Maps1 week ago
Mapped: Which Countries Recognize Israel or Palestine, or Both?
United States1 week ago
Ranked: America’s Best Universities
War1 week ago
Ranked: Share of Global Arms Imports in 2022