writing the discussion section can be one of the most challenging part of a research paper and usually it takes the longest why is that it's because it's complex nobody know what disgusts actually means I recently gave a master class on how to write different sections well and make it a compelling story I took out a snippet specifically on the discussion side to talk about what discussion means how do you get started and what makes a week versus what makes a strong discussion stick around this is a very good video it's going to change how you think about your discussion section I'll see you there what's the discussion section the discussion section is to interpret your findings in the context of a research field so think about the application and implication of your research importantly readers don't necessarily see what you're seeing your results you actually have to tell them so this is a structure you'll see in the discussion section first is paragraph one usually it is a summary of all the results and the main answer to your research question the key results then you have a bunch of middle paragraphs here you're discussing findings in sequence or in importance and then in the final paragraphs we talk about limitations strengths application and also future directions what does it mean to discuss a paper when I first tried writing paper I find discussion the most difficult it is still the most difficult the main issue is I don't really know how to discuss what does it mean so let me give you a few Clues first is this an unexpected finding if it is why do you think that is very important don't sweep it a little rub because the reviewers will ask you to explain it I can share a story there was one study result that were really interesting for us and unexpected the strategy we took was you know what it's a small thing I don't think people will notice let's not talk about it if we don't talk about it maybe people won't notice it of course when we got the review back everybody noticed it all three reviewers and also the editor said can you please explain why this is happening and so we have to go back so my main point is if you have an unexpected finding that is troubling don't sweep it in a rug I just really think about it and address it the next one is concordant findings with other studies that means similar as you said why is it similar why is my study finding similar to other papers in the literature but don't just end there think about how does your study finding add more to it then the third one is this finding with other studies or maybe it's mixed what you want to include in your discussion is talk about why is it different is it in different population is it the sample size too small new theories or maybe mine is better because you can stop talking about that and then finally in the last few paragraphs typically you can also talk about implication the impact of your study finding application how can you use the study results in different settings and address potential objections and also talk about future research so this is a discussion Matrix that I create before I start writing the whole paper up the reason I like to do this is because I like to plan write the bullet points first and the outline share it with the co-author before you spend time writing those beautifully crafted paragraphs and from personal experience whenever I don't do that I get bitten at the end because um the re the co-authors like oh I I don't agree with this discussion um I don't agree with this part let's reframe everything so it is better to plan ahead think about my findings what are the studies with similar findings in the second column maybe studies with discordant findings and also generally think about the application implication and write about your thoughts and now you can create an outline once everybody's happy with that then you can start crafting the paper it's really much more efficient that way now I want to talk about weak versus a strong discussion so my first few papers I have been using a lot of week discussion and after reading more papers I realized they have differences in higher impact papers in higher impact journals via discussion sections are much much stronger and the difference is the weak one only States what studies are similar or different from our study finding and the strong ones not only say what but they State why and how other studies are similar or different from our study finding so that's a key difference let me show you a personal example from one of the drafts before we made major changes with the paper so here you can pause and take your time to read through the paragraph but really what I want to highlight is first we summarize study one basically reporting the rates of graph loss then in the second part I summarize study two and we were reporting the rates of graph loss and then the final sentence we talk about our own finding and about rejection rates so if you can see here we are only talking about the what and we just basically summarize and didn't really have much thought to it then this is the after version and how is it different first we talk about our own finding this is a hero then we talk about why there are different rates of graph rejection and here then we support okay studies showing that one type of therapy increases graft rejection and then one group of therapy is used to treat rejection and then another group of therapies shown to reduce rejection in animal studies and so here we are talking about both studies that are contradicting and also supporting but we explain why the differences so that that made it much stronger so let me show you another example here this is a paper published in New England Journal of Medicine here the first sentence they immediately say one way that our study differs from that of guilt and colleagues was the size of the dose they are saying how they are different from other studies and then in the middle paragraphs here they explain why they deliberately chose their method then finally they added a little bit of balance opinion but the smaller dose could also be okay so here again importantly it's not just about reporting studies that are different from them but explaining how they are different and why they chose this particular method now the final part I want to talk about limitations limitation can be a struggle for some people because they may become one or two and then they stop there so let me give you some few examples get started with those that are inherent to study design for example if you see cross-sectional studies you can say no temporal relationship if you see a cohort study you can say it's a cohort study therefore we can only assess Association and organization if you're writing a qualitative study you can say oh because this is qualitative study we cannot make statistical inferences only hypothesis generating if it's a randomized controlled trial you can say that results May not translate into real life setting because this is done in a highly controlled environment other limitations include limitations that are related to study instruments or method of data collection or access of data and also sample and selection of sample to small or limited to a certain population number four it could also be limitations of the researchers and able to carry out that research as planned for example dropout rate unable to recruit participants running out of time or money outside circumstances and number five generalizability very important slide do not merely State the limitations also explain how you're contributing to science despite the limitation or how the limitation is not as bad as people might think or maybe how you have tried to fix the limitation or reduce bias and how future study is needed to solve this limitation and make sure you you're specific about this so I'm going to give you some examples here so here ethnicity was missing for approximately 26 of patients so they are saying that as a limitation and then the last sentence but was broadly representative so basically you give the limitation but it's not as bad as you think the next sentence here they were also missing data on Obesity and smoking and then the next immediately they say sensitivity analyzes found that our estimates were robust to our assumptions around missing data so they basically tried to say that fix this limitation it checked the limitation and it was actually fine all right so I've shared some strategies and tips on how to write a discussion section if you like this video give me a thumbs up and I will see in the next video