Prove It
The Replication Crisis. More and more research is coming out to show just how difficult it is to accurately replicate experiments. How many things we accept as fact are just wrong.
A climate crisis, a virus crisis, a food scarcity crisis…we’re going through a bit right now. Underpinning all of these threats is the thorn in science’s side right now…the replication crisis. In short, more and more research is coming out to show just how difficult it is to accurately replicate experiments. Sure, a study may prove some theory with statistical significance but it’s incredibly difficult to repeat the experiment and get the same results.
Research found 40 of 100 of the top economic research papers could not be replicated. Another study found that most conclusions in economic research are exaggerated by at least a factor of two, and a third are exaggerated by 4x. That being said, economics is fairly accurate. Other fields like psychology see immense difficulty in replicated results.
The obvious question arises, what are we accepting as fact that simply isn’t true? Let’s run through some commonly held beliefs that may not be so statistically significant.
The first is the signature “marshmallow test”. You’ve seen it before. Kids are left in a room with a marshmallow and told not to eat it. The original conclusion was that those kids who abstained from eating the marshmallow had greater willpower and were more likely to succeed. 2 decades later, the experiment was more rigorously performed and no correlation was found between high school success and willpower at 5-years-old. In fact, it was found that family wealth was the biggest explanatory variable; poorer children were more likely to eat the marshmallow given the instability and less certainty in their day to day lives.
More directly related to the markets, an incredibly influential paper in 2010 found high government debt was correlated with low growth. The findings affected government decision making until it too could not be replicated. In 2013, researchers found coding errors, data discrepancies and an unsound method rendered the results worthless. The revised results found high levels of government debt encouraged growth.
Let’s be very clear, this isn’t anti-research or science. The purpose of this piece is to encourage critical thinking. Especially when investing, questioning what everyone else accepts as fact creates opportunity. On top of that, it highlights just how much more powerful research which has been replicated is compared to single-study findings.