#biostats

[will-the-old-med-student]

I wanted to pre-study everything from the whole year, but quickly realized that 2 weeks would be too little for that, so I tried to stick to the first semester, but even then, it's a little bit heavy.

It's all the "non-medical stuff" during semester 1: Biophysics (which includes thermodynamics and I need to spend a lot more time to understand the differences between all those formulas because it is a nightmare), Biostats (which is very okay/easy so far but it's just the begining so I'm keeping my guard), Chemistry (I am super surprised I understood mesomery and even chirality, highschool me would be freaking out) and Biocell/Histology (admittedly the most "medical" one, and the easiest to me so far). Apparently we do some embryology during the first semester, but it's in the same topic as Anatomy so I need to sort which is really anatomy and which is really embryology.

If I don't finish all the studying before entering the year, I think it's fine. Having it in my mind already for a lot of it will absolutely help for the year. I'm so happy to feel rather comfortable with a lot of things that I see already.

I had the opportunity to speak with another person who decided to go to medschool a couple of days ago too, he gave me some advices. He worked so much for it, but we didn't have the same highschool background, and he didn't do as much pre-work as I did. I'm hoping to have an easier time, but I'm targetting a good rank, so we'll see. I figured, if I want to be in the top 10~20 (which is about 16/20 average notes), and end up top 40~50 (about 15.3/20), I'll be pretty happy still.

[newassignmenthelpaus]

Biostatistics Assignment Help

Assignment Help is a valuable academic service designed to assist students and researchers in the field of biostatistics. Biostatistics is a crucial discipline that applies statistical methods to biological and health-related data, playing a vital role in research, clinical trials, and epidemiology. This specialized assistance offers students the opportunity to excel in their biostatistics coursework and research projects. Biostatistics Assignment Help not only enhances academic performance but also fosters a deeper understanding of statistical principles, thereby contributing to the advancement of research in fields such as medicine, biology, and public health. It is an indispensable resource for those seeking proficiency in biostatistics.

[kaeyas-nalgas]

I feel like the password to my poipiku posts is so easy. what if I decided to be like that one artist I saw once who put entire algebra problems to solve for the passwords

[gayofficial]

does anyone else feel like the dumbest person alive or is it just me and myself but after staring at a screen for several hours straight

[irhabiya]

math is so fun why am i in medicine again

[finmoryo]

one more month and then I'm free to actually work on writing fic for once

[0126p]

literally losing my mind studying biostats rn i hate it!!!!

[ohcitron]

hello i sent your worlds specialist fish to my marine biologist friend and his immediate reaction was "my god" "need that as a tattoo" . this isnt asking for tattoo permission i just would like you to know that your worlds specialist fish has fans

this makes me want to cry i'm so glad everyone loves worlds specialist fish. i'm giving permission regardless please let your marine biologist friend know i'm truly so pleased he'd ever consider it as a tattoo

[scrimshawed]

truly humbling to look up a math tutorial and see it's listed under pre-algebra

[bare1ythere]

I'm so so so so tired and burnt out of schoolwork and the terrible news is im an idiot so. i have to spend 8 hours a day studying when my siblings can stop at reasonable times and do fine. I'm in a hell of my own making 👍

[capybara]

i’m so far out of the loop on that entire tina fey bit i need a translator

[gottagobuycheese]

so close...freedom is Right There....

[cupcraft]

im at that stage of my semi-hiatus/studying for a standardized test where im so burnt out i dont gaf anymore like my exams in 1 week to the dot and i canttt be assed to care but i still have so much to finish before i do more practice exams gawqhghshgsh

[angstmongertina]

I do want to add that the argument isn’t just about MRIs. The paper argues for the importance of multiple testing corrections, which is relevant in far more science than just imaging analysis. That’s something that can happen in any field in which tests are being performed. The MRI example is just one such case.

This essentially comes down to what we mean when we say statistically significant. If you think back to your science classes, you may remember learning about null and alternate hypotheses when performing an experiment, or designing a test. The idea is that the null is sort of the "default," and our goal with our experiments, much of the time, is see if the null hypothesis is false. You may hear it as "reject the null." Typically, we feel comfortable rejecting the null at p<0.05. This means that there is a 5% chance that any differences we're seeing in our data compared to what we expect is entirely by chance. In other words, it's pretty unlikely that the results we're seeing are by chance, so instead we can think "well, maybe there's an actual difference."

Now, that's great for one test. But let's say we're doing 100 tests. And each one rejects at p<0.05. And let's assume that all of them truly have no difference. Well, based on our cut-off for our p value, on average, we would expect 5 of those 100 tests to reject, even though there is no true difference. This is what's called a false positive, and is a danger whenever we're doing tests like this. There's always a chance that we find an association that isn't real. But, we do our best to try and limit them.

Now back to the dead salmon. As a disclaimer, I don't work with imaging data, so anyone who actually does imaging analysis, feel free to correct me, but as I understand it, when you're looking at MRIs, you're looking at tons and tons of small regions, and looking for associations between them and the stimulus. Essentially, you're doing tons and tons of tests. And if you continue to use the same cut-off of 0.05, you're going to end up with quite a few false positives. Which is what happened in the dead salmon study. If I recall correctly, they found a region with association in the optic area of the fish brain, indicating that the very dead fish can still see things. Which is clearly impossible, since the fish is dead. Instead, what was happening was that a cluster of tests just so happened to be false positives in a way that might make it appear as though there was an association with that region of the brain. And since the fish was dead and should not have any brain activity, that was very clearly a case of false positives.

So, what can we do? Well, multiple testing corrections is essentially a method that can be used to combat this. There are a few different ways to adjust the p-value, or otherwise control how many false positives you get, but the basic idea is that when you're doing many tests, like in the case of MRIs, or other studies like Genome Wide Association Studies, it's very very important to adjust the p-value, so you don't end up with a bunch of false-positives. And any scientists performing multiple tests should be wary of these kinds of results. It's not always as easy to tell as when we're looking for brain activity in a dead fish, but the results can be equally unrealistic.

In summary, please check how many tests you're running and perform multiple testing corrections accordingly, even if you're not working with MRIs. False positives don't care about what field you're in. They come for all.

one of the best academic paper titles

[brilliantsnafu]

At this point I think I should just start lying about my masters. Never got it. Since no one wants to give me a fucking job bc of it.

[irhabiya]

our senior epidemiologist is so intelligent you can tell he knows the field inside and out but god if only that knowledge translated well into his lectures😭