Really enjoyed the latest episode on the importance of truth in cancer research. It was not only informative but also entertaining. Your dedication to transparency shines through, making a complex topic accessible and engaging. Thanks for casting a light on the crucial facts and questioning the credibility where it's due. Keep up the great work!
These are very funny stories about how to make sure you get the results you want. I would never have thought to do these. Perhaps I have too much belief that we should be doing good science. It does remind me, however, of a situation my sister ran into in graduate school. A sociologist had developed a study which asked very ambiguous questions of non-native English speakers. The results were not what he wanted, no matter how he interpreted the answers. His solution? Change the statistical test! It has been 40 years since she described it to me. I can't recall if he went from chi-squared to a 't' test or Poisson, but the alternative test made no sense for the data being analyzed. But it somehow gave him a statistical significance to prove his hypothesis....
I just had a conversation at work today where I mentioned -- it's good to broaden one's toolbox, so you can learn which metrics/tests are going to be the most meaningful for the situation you're working on.
I come from an academic background, and I learned which tests were appropriate for what use. I took several grad-level Probability classes, where we were looking at a variety of Limit Theorems which are the underpinnings of a lot of these statistical tests. I learned the theoretical basis of the Central Limit Theorem (and where it doesn't apply), etc. So it does pain me when I see this stuff mis-used.
My point is trying to get at truth, so I'm not trying to just try anything until I get the result I want. I WANT THE TRUTH! I WANT IT, DAMMIT!
Really enjoyed the latest episode on the importance of truth in cancer research. It was not only informative but also entertaining. Your dedication to transparency shines through, making a complex topic accessible and engaging. Thanks for casting a light on the crucial facts and questioning the credibility where it's due. Keep up the great work!
Thanks for the kind words!
These are very funny stories about how to make sure you get the results you want. I would never have thought to do these. Perhaps I have too much belief that we should be doing good science. It does remind me, however, of a situation my sister ran into in graduate school. A sociologist had developed a study which asked very ambiguous questions of non-native English speakers. The results were not what he wanted, no matter how he interpreted the answers. His solution? Change the statistical test! It has been 40 years since she described it to me. I can't recall if he went from chi-squared to a 't' test or Poisson, but the alternative test made no sense for the data being analyzed. But it somehow gave him a statistical significance to prove his hypothesis....
I just had a conversation at work today where I mentioned -- it's good to broaden one's toolbox, so you can learn which metrics/tests are going to be the most meaningful for the situation you're working on.
I come from an academic background, and I learned which tests were appropriate for what use. I took several grad-level Probability classes, where we were looking at a variety of Limit Theorems which are the underpinnings of a lot of these statistical tests. I learned the theoretical basis of the Central Limit Theorem (and where it doesn't apply), etc. So it does pain me when I see this stuff mis-used.
My point is trying to get at truth, so I'm not trying to just try anything until I get the result I want. I WANT THE TRUTH! I WANT IT, DAMMIT!
Truth is admirable for its impartiality and humility: it applies universally without bias and operates without ego, promoting fairness and honesty.