Nostr Web Client

This is an academic thread. The first principle in any academic search is to differrentiate between association and causation.

An assocation means to things look related. It may be causing each other, it may be not.

For example, look at the first image. At one glance, you may think that if you eat more margarine, you are more likely to get a divorce. however, this association sounds ridicolous (obviously) and the causation is doubtful.

unfortunately, such association occur in our everyday lives. we see event A, we see event B, we conclude that they are related. but they are not. to determine if an association is causation, we look at the 11 rule of bradford-hill criteria. see image 2.

Disturbia 2y ago

Now, a lot of people tell me, " i find this study from a research". but the question is, "which one?" so it is a good idea to learn about the hierarchy of study design. The highest level of evidence are clinicak practice guidelines, which combines the best of expert opinion with research findings, then meta analysis of trials, followed by meta analysis of other study types.

randomized controlled trials are the next best, but it is very hard to design. so most human setting study settled on quasi experimental design.

now where do you draw the line for association vs causation?

at cohort. anything below cohort is association (although you can use regression analysis to make a bit of causation) and from cohort onwards, the causation gets stronger

Reply to this note

Please Login to reply.

Discussion

Disturbia 2y ago

While scientific studies are important, it usually fails to understand the exogenous factors that cause a problem, such as the question of why people defaulted chemotherapy or why people choose to have unprotected coitus.

such question are best answered by qualitative studies as in image 1.

question is, how do you decide if the qualitative study is trustworthy? you use a specific analytic criteria

Disturbia 2y ago

now, there are medicine that were known to work in scientific studies, but fail to replicate the same result in real life. why? because these studies fail to consider external factors like the environment or emotion. thus, it is always good to observe community trials. there are studies done on real human in a real life settings. is it hard to do community trials? definitely. you have to make a whole community to agree to a community studies / trial

Disturbia 2y ago

let us refresh some basic school science.

an independent variable is a thing that we think to cause something,

while a dependent variable is that something.

for example, if we worry that eating chicken can cause lung cancer:

independent variable : eating chicken

dependent variable : lung cancer.

if we are exploring multiple factors, we can simply replace independent with the word "risk factors" or "determinants".

Disturbia 2y ago

Now that you have learnt about All these, pet us practice together.

let say your momma told you that her grandma used to consume Hibiscus Rosa Sinensis when she suffered pains, like dysmenorrhea.

so what are the independent and dependent variables?

yes, they are:

independent variable : consume Hibiscus Rosa Sinensis

dependent variable : reduces pain

now head to consensus.app website and type :

Does Hibiscus Rosa Sinensis reduces pain?

Disturbia 2y ago

you will get relevant answers based on different researches along with their explanation. you may need to use either a 10 years or 5 years filter so you dont get really outdated studies

Disturbia 2y ago

if you click on the researches you will get more information such as the study design. this is where you can use your knowledge about hierarchy of study design evidence, especially if you find different studies with conflicting results.

Disturbia 2y ago

i hope my post is beneficial to you. thank you!