I like how people with no medical expertise or ability to understand statistics tell others on the internet that they don't think people who actually have these skills are right, and that they need more skills or training. How about you? If they need more skill, then you're a complete rookie so I'd suggest gaining some before criticising others. As a statistician, I tend to believe DeeJay is right. The shit you see on the internet is mostly not strongly supported or inferred correctly from the data. Plus, there is no such thing as correct conclusion in Statistics, because Stats is about taking samples to ESTIMATE the true parameters of a population, parameters that you will never ever be able to calculate precisely. There is a thing called confidence interval, and the bigger the interval, the greater the confidence that you will have a good estimate. If you see a number on a website and a number on another website, this is because samples are different and tests are different and also the estimators might be biased. It's always better to have an interval estimation rather than a point estimation, because one number like 6 doesn't tell you anything (or almost anything).
Do girls really want 5 inches
↑ View this comment's parent
← View full post
I like how people with no medical expertise or ability to understand statistics tell others on the internet that they don't think people who actually have these skills are right, and that they need more skills or training. How about you? If they need more skill, then you're a complete rookie so I'd suggest gaining some before criticising others. As a statistician, I tend to believe DeeJay is right. The shit you see on the internet is mostly not strongly supported or inferred correctly from the data. Plus, there is no such thing as correct conclusion in Statistics, because Stats is about taking samples to ESTIMATE the true parameters of a population, parameters that you will never ever be able to calculate precisely. There is a thing called confidence interval, and the bigger the interval, the greater the confidence that you will have a good estimate. If you see a number on a website and a number on another website, this is because samples are different and tests are different and also the estimators might be biased. It's always better to have an interval estimation rather than a point estimation, because one number like 6 doesn't tell you anything (or almost anything).