Selecting only the evidence that supports your conclusion
Climate change deniers love to point to cold winters.
“Record snowfall. So much for global warming.”
They’re ignoring the global temperature trends, the melting ice caps, the shifting growing seasons, the increasing frequency of extreme weather events. They found one data point that fits their position and presented it as if the rest of the picture doesn’t exist.
This is cherry picking.
Selectively choosing evidence that supports a conclusion while ignoring evidence that contradicts it. It’s different from confirmation bias from Essay 6. Confirmation bias is the unconscious tendency to favor confirming information. Cherry picking is the active, deliberate presentation of only favorable data as if it represents the whole. Confirmation bias is something your brain does to you. Cherry picking is something somebody does on purpose.
The pharmaceutical industry has done this systematically.
Drug trials produce mixed results. Companies publish the positive ones and quietly file the negative ones away. Doctors and patients then make treatment decisions based on an incomplete picture, because the studies that showed the drug didn’t work – or caused harm – were never made public. This is why mandatory registration of all clinical trials became a major reform push. The public record was being constructed from cherries.
Politicians do it constantly.
Want to prove the economy is strong? Cite employment numbers while leaving out wage stagnation and housing costs. Want to prove it’s failing? Cite housing costs and ignore employment. Both moves can cite real data. Neither is giving you an accurate picture. And both sides will accuse the other of cherry picking while doing it themselves on the very next slide.
Crime statistics are particularly vulnerable.
Cherry-picked neighborhood data, cherry-picked time periods, cherry-picked offense categories. You can make crime look like it’s going up or down depending on which slice somebody chooses to present. Headlines rarely specify which slice they’re reporting on. People who want a specific narrative can usually find a way to build it from cherries that are all individually real.
The tell is the absence of contrary evidence.
Good analysis acknowledges the data that complicates the conclusion. If every single piece of evidence somebody presents points in exactly the same direction with no nuance or contradiction, something is being left out. The world is messy. Arguments that are suspiciously clean are usually hiding something. Real data almost always contains points that argue against the cleaner story.
There’s a version of this that’s specifically dangerous.
Cherry picking counterexamples in a way that looks like healthy skepticism. Somebody picks the one study that contradicts a strong consensus and treats it as decisive, while ignoring the hundred studies that agreed with the consensus. The one study is real. The consensus is also real. Treating the outlier as the whole story is the move. Science denial runs on this specific version constantly – find the dissent, amplify the dissent, pretend the rest of the field doesn’t exist.
When somebody presents you with data, the question isn’t just whether the data is real.
It’s whether it’s representative. What’s the contrary evidence? What does the full dataset show? Who’s ignored in the presentation? If those questions can’t be answered, the argument isn’t built from evidence. It’s built from cherries.
The cherries are real.
The orchard is the lie.