On the other hand, I often see studies where I'm quite curious to reproduce it and I just can't. Just today: saw a 2007 study from a university finding a correl

On the other hand, I often see studies where I'm quite curious to reproduce it and I just can't. Just today: saw a 2007 study from a university finding a correlation between sleep quality (deep sleep phase duration) and exposure to a GSM-frequency signal for 3 hours (a long, but not unusual/absurd, phone call) before sleep, even though the participants could not tell better than chance whether they had been exposed... would love to see if that reproduces even though the p value was not bad, but I just don't have the people nor the radiation calibration equipment nor the sleep lab nor anything. Even so much as a survey is hard to properly reproduce, definitely more than a day's worth of work if you need to go out and find a random pool of surveyees. Opening up a spreadsheet in your favorite software sounds fairly approachable by comparison, assuming it has no fancy university-level mathematics in it. Whether it is fun to talk to people for a survey vs. sit through pages of arithmetic is a whole other question, though.