Speaker 1: Listen to me, you are so close with your research or science fair journey as a student. It is essential that you make a great research plan. In this video, I'm going to show you exactly how I planned my research process and wrote a proposal in order to win international awards at top-level science fairs, publish science research papers in actual journals, make a genuine contribution to my field of interest, and of course, learn a ton of valuable skills along the way that have taken me a long way from a high school student to now a college student doing research. There are nine steps in this process. First, let's talk to Sahil Sood, a sophomore at Harvard College who has published research papers, research conference papers, and written over 20 abstracts. Yo, Sahil, how does one come up with a rationale for their research project or science fair project?

Speaker 2: In terms of a rationale, I think it's best to look at this in context of an example. So let's take a very common project, if you're using artificial intelligence in some capacity to try to solve the issue of diagnosing or treating neurodegenerative diseases. From there, I think there's kind of two ways that you can take the rationale. You either have some kind of personal connection to the issue, why you're doing this project, because you're personally invested in some way, whether you yourself have faced this disease or you have a family member that faces this disease in some capacity. You're then able to link your personal story and your experiences anecdotally to what exists in terms of the research project itself. The second way to look at the rationale is more from a global standpoint. A lot of the issues that are being solved in science fair research specifically are ones that impact thousands, if not millions of people. And so you can discuss that you're pursuing this research project in order to better the quality of lives for those people that face this disease on a daily basis. And that kind of enables it so that you're able to see the purpose of your research and what the downstream impact of that research should be.

Speaker 1: Thanks, Sahil. That was really helpful. Over here, I'm now screen sharing one of my past winning research project proposals. This is from 2022 at the International Science and Engineering Fair, where I won first prize. I then turned this research project into a published research paper in BMC Bioinformatics in high school as a first author. And just so you guys know, I see a lot of discussion about high school journals and some predatory online ones where you just pay a lot of money, pay to play. And I would highly like to condemn all of those. You guys should genuinely go after legitimate journals that are peer reviewed, that will actually make an impact on science. You know, in high school, I was able to publish this research paper. And now it's sitting at something like 11 or 12 citations within just a year or two. And that's really exciting for me because it shows that actual researchers, actual scientists are using my project, my technology. And I want you guys to be able to make legitimate contributions as well, and not just do things for college apps. So that being said, let's jump into the rationale here. I think it's really important to highlight that I always like to have kind of a two or three pronged attack, right? If we're trying to prove to someone, why is this problem so important? Why was I motivated to solve it? And we're going to want to have multiple ways we can attack it. If we just go all in on one strategy, it's possible that the reader or a science fair judge may not appreciate that they don't resonate with it, or they don't agree with it. And so that's why having multiple layers, kind of like you peel back the onion is really important. In my rationale, the first paragraph, I give this really broad overview of the entire field. I'm solving one kind of niche problem in the whole field. But I like to zoom out at the very beginning just to motivate why this is important. If someone is just hopping in a science fair conference and listening to my speech, or someone is reading my research paper and they have no idea why this is important, they need to understand that. And so I like to give these statistics. These are the types of things you would see from like a McKinsey report or a financial analyst, right? $125.8 billion, and it's growing, and this is so important. But in the second paragraph, we do cater to that scientific audience by zooming in a little bit more on the specific technology and science that we're examining. And this here is about the problem of codon optimization, and I'm motivating why that problem is so important. As you can see here, I'm actually citing studies. And this is where I like to talk about pathos, ethos, and logos in my research projects and in my science fair coaching that I give, link is in the description. And this is a great example of how I use it because I'm appealing to credibility by citing past researchers and showing that this is what they said. And based on this, this is the conclusion that I'm making of why this is important. And so overall, I hope you guys can kind of just understand that when you're approaching a rationale and motivating why you're looking to solve this problem, you should really approach it like peeling back the layers of an onion and give multiple reasons and lines of defense of why this problem is important, why you wanted to go after it and solve it. So the next piece of the puzzle is coming up with a research question for your plan or proposal. And so question is not really a problem statement, but it's a question that you as an inquirer, as a scientific person is trying to answer. And so I'd really encourage you guys to focus on novel questions. I talked about how to come up with research questions in my past videos. I literally have a 25 minute guide that I'll link in the description down below where I show you from start to finish how to come up with your research project topic. But coming up with the question is just slightly different where we actually want to answer something that hasn't been answered before or that we're interested in. And I think novelty is really important here because otherwise you just have students doing the same sort of projects again and again, which are really good for skill development, but don't really offer any contribution to science, which is fine, but why not? Why shouldn't we have students and young people making legitimate contributions? Welcome back to Google Docs. As you can see here, this is our research question. Does a codon optimization tool guided by sequential and contextual information present in a cell factory's genome help improve the efficiency of recombinant expression? And so I think it's really important to note here that in your question, it should be something that you're genuinely trying to answer that hasn't necessarily been answered before. So as you guys can see over here, you might have a hunch, which is leading into your hypothesis that yes, utilizing the sequential and contextual information will improve the efficiency, but that is the question we are trying to answer. So I want you guys to carefully consider your research question and the resulting hypothesis that you come up with based on that. But how you can actually come up with something that's really novel is using this tool called sciencefair.io. I'll link it in the description down below as well, but it was made by this really handsome guy in the bottom left. I have no idea who this guy is, but he is super handsome and he's definitely not me. But basically you guys want to type in your topic or any sort of interest you have. Let's say you're interested in adeno-associated viruses or AAV, hit enter. It's going to search through thousands of past winning research projects and then based on that, help you come up with a new winning one that is a spinoff based on it. So you basically get to stand on the shoulder of giants who have done work previously and now come up with a novel idea that you can go after. Another site I want to show you is called typeset.io and so basically you guys can type in your topic. Once again, let's just say it's very broad, AAV. We have no idea what it is and then we want to come up with a research question based on it. Well, we can scan through these insights that it gives us based on past studies in order to come up with a new question that is novel and I think I really want to focus on that novelty piece. I know I'm just belaboring that point over and over again. The reason why is because so many students will just do projects that have already been done before and as a result, their work just doesn't get recognized. So I really want you guys to focus on things where you can make a legitimate contribution because it will not only be rewarding for you, but it will also help society and help fields. As you guys probably saw, the next part of the research process is writing your hypothesis. And for this, I'm inviting a researcher from the Broad Institute, Shreya Bhatt, who happens to be a sophomore at Harvard and one of my closest friends, who has won ISEF multiple times and also published research papers.

Speaker 3: So a hypothesis is basically an informed guess or prediction about what you think is going to happen based on your research question. And I think the best way to go about developing that hypothesis is just having a very good sense of what's currently out there, maybe current treatments, and then based on your own knowledge of the biology or chemistry, what you think is going to happen next. So I would say my number one tip for students trying to come up with a research proposal or question is that your contribution doesn't actually have to be that, you know, revolutionary or transformative, even if it's a small step in the right direction. You can then, from that information, come up with more interesting questions and that's often how it works out in science anyway.

Speaker 1: So the next piece of the puzzle are the objectives. And this varies depending on the type of project you're doing. If you're doing science inquiry, this is something where you're trying to answer a question and usually you'll conduct an experiment to do so. Then you're going to want to come up with these two different pieces. One is variables. So usually you have an independent and dependent variable that are influencing your experiment. In addition, you have constraints. This will be present for all sorts of projects. It could be things like resources, time, or money that enable you to do or don't do your experiment. Then for engineering, there's a different set of things. There's the engineering claim, which is the claim you make that by the end of your project, by the end of your research, you will have a device or an innovation or some end product that can do the following things or can solve the following problems and that's your claim. Your criteria are basically how you evaluate your solution, how you evaluate if your claim has been successful. And so typically it could be something like if you're developing in treatment, you have over a 90% effectiveness, or if you're developing a diagnosis tool, you have over a 90% accuracy or something like that, a way to quantify or measure how successful you were and based on that decide yes or no, was this a success? And finally, you also have constraints similarly like resources, time, materials that go into the project. As you guys can see over here, I'm going to go to my engineering goals that I had for my project. I have an engineering claim, which is that in this project, a tool will be created and it will improve these different factors. And then I have these primary and secondary endpoints, which are kind of like the criteria that I use to evaluate the success of my claim. The next piece is to come up with your methods and design your experiments. This can be really tough, but I want to give you some pointers so that I can lead you in the right direction. I think the best way to frame this is by taking a look at your research problem or the research question that you're looking to answer and identify the type of data that you're going to need to answer it. So let's say you have a research problem. What you're going to want to do is try to understand what category your problem comes under. It could be one of these three things or something else that I don't have listed, but I'll give you some examples. One is establishing a relationship. So this could be something like a cause and effect relationship, like sharks and the amount of cheese that people eat. This is just like a scientific inquiry kind of study that you might be interested in, and you're looking to establish a cause and effect relationship. If so, now we need to come up with what type of data we're going to look at. This should either be quantitative or qualitative, and I'll get into that in just a second. Let's say our problem is more dealing with describing the characteristics of something, like we're looking at a new type of bacteria, and we're looking to categorize the bacteria amongst usual traits. So we're going to do a series of assays on that bacteria to understand how it functions. And so in order to do that, we're going to need to, again, get some quantitative data or perhaps qualitative data, like if we run some assay and look at the color of the bacteria and what color it turns into. Then finally, there's a gain a more in-depth understanding. And so this is something that can be like, let's say you have a research problem where you're dealing with something that's completely new. In this case, you're trying to gain a more in-depth understanding about that topic. And so you're going to want to study it in kind of like an observational setting or something like that. And so once again, think of the type of data that you might need to answer your question. Now engineering type of problems are different because in those cases, you have a clear solution and you're working towards that claim that I was describing earlier. And for that, you're going to want to design a series of methods that best tackle that. And I'll get into that in just a second. So now in types of data, there's two main types. There's quantitative and qualitative, and I've kind of discussed this already, but one other consideration you have to make is whether you're going to collect the data yourself or if you're dealing with others as data. So as a student researcher, especially, I think this is something that's really important because when I first started out doing science research, especially in middle school, I had no data whatsoever. And I would email some professionals, I would do all this stuff, and that got me nowhere. But what I ended up figuring out was I didn't have to necessarily collect the data on my own. I didn't need to get permission from the hospital to let me run XYZ tests when the data from those tests that were used in other scenarios were already published online. And I could now use that data and apply it to solve some other type of problem while citing the original creators and curators of that data. So that's an amazing way to use other people's data. But you're going to need to decide on whether or not you need to collect that data yourself. So if you're collecting that data yourself, then you're going to need to design those experiments and trials to collect that data. So in terms of the methods, that is exactly tying into this first bullet point here, sampling methods or a criteria for your sources. So are you collecting that data from yourself, like you're running some technique in a lab, or are you collecting that data from others, whether that's online or literally sampling other people? And so I think that's a key distinction that needs to be made when you're coming up with your methods. The next step is coming up with the tools, sources, and materials that you're going to need, whether that's something like a programming language to a laptop computer, all the way to the reagents that you need to run your experiment. This should be listed in kind of like a materials list, especially in your research proposal or plan. And that's going to save you a lot of time later on because you have all of those materials listed and you know exactly how much your solution is supposed to cost and what tools and materials you need before you can actually get started. And that'll help you think about everything that you need. And then finally, understanding how the variables are going to be measured or were measured. Typically in methods, you write everything in the past tense because they were conducted in the past tense. But if you're writing a proposal or a plan, obviously you're going to conduct it. So how are those variables going to be measured, whether that's quantitatively measuring something or qualitatively looking at how something changes. So here's some example headings for what you could include in your research plan. This is either something mentally you can have running or physically, like write it down on a Google document or on a piece of paper, just like how I had. Some example headings would be data collection, where you're collecting all your data, a curation process, wrangling it all together, then prepping and processing. So let's say you want to take all that data and use it in some statistics software like R or Stata. That's where you would list information about that, as well as any software that you might need, programming languages, things like that, as well as languages, images, and interpretation. Because oftentimes if you're dealing with qualitative data, you need a good way to describe it. And so that's where you can come up with those methods in your head beforehand to plan for the future. And then also I want you to, once again, consider why your method is new. Oftentimes science is driven forward because of technology. You come up with new technologies and those technologies in turn allow you to conduct better science. And so to come up with those new technologies, you need to forge new methods. And so I want you to also consider if any of the methods you're using aren't new. And don't be afraid, like as a high schooler, as a middle schooler, or whatever age student you are, you can come up with new methods too. I was able to do it in my research and some of the speakers that I've included in this video like Sahil and Shreya also did it in their research at a very early age in high school. And so I think it's completely reasonable for students to be able to come up with new methods as well. In fact, students have some of the highest rates of fluid intelligence out there. So definitely don't doubt yourself. You can come up with something new. That being said, I do want to briefly hit on the engineering type of points. Once again, I personally had an engineering type of project. And so oftentimes I would look at guides like this and I would be like, huh, none of this stuff really like properly fits what I'm trying to do. And so that's why I have that procedure section here listed out for you guys. This is a literally a bullet point list of what I had roughly planned to do for my research project. And it was about right. At the beginning, I wanted to curate, you know, a list of genes, train the model, I'm going to use this technique for processing, I'm going to use this to quantitatively describe the data that I'm collecting, right, measure the free energy, measure the codon adaptation index of all of the sequences in my data set, then I want to do some processing, once again, create some scripts is actually getting into the software methods that I was just telling you about, create some more software methods, once again, create, you know, all software methods and specifics of that. And then finally, conducting those statistics in whatever software. And then, you know, once again, going through those processes of calculating the statistics, and then finally evaluating if the solution we made matches our initial engineering claim. And so that is the the biggest part where the engineering method varies from the scientific inquiry is on that final step, where instead of saying, Hey, did we succeed? Like, did we answer our question, you instead need to go back and look at your claim and see if you need to iterate. Engineering is all about iteration. And that's why it requires a lot of perseverance. Because in science, often you just fail. And you know, that's that with engineering, you make these iterative improvements over and over on your design until you reach the claim or meet the criteria that you had specified earlier. And that's exactly what this is about. We're testing it against these other benchmark genes testing it against other optimization methods. And that is exactly what I listed in this procedure. So now that you've come up with your methods, there is a very, very high chance that as a student researcher, you've realized that there are some methods that you just don't have access to, right, it could be some fanciful neurotechnology that costs millions of dollars. And as a student, there's just no way you could get your hands on that. And that's completely normal. Now, one way to solve this is just by adapting. And I'm going to give you 10 solutions for that adaptation process. And don't get worried, myself and some of the other speakers I had on this video have used the exact same techniques when we were students. So the technique number one I'd like to give you is leveraging online databases, there are tons of free online databases, some for biomedical spaces that I know in particular are TCIA, TCGA, Kaggle, NCBI, these are really great resources. And I've also done some research and I found for environmental studies, check out NASA's Earth data, I found it to be very, very extensive. In addition to using online databases, the second technique I have for you is open source software. So let's say you have like a very specific application for something that you're not sure of. There are often free libraries made in Python and in R that can do those things that you're looking to do via software. So oftentimes learning coding can really come in handy, spending 15 or 20 hours watching some online tutorials on YouTube on how to use those software packages can save you a lot of time and money later on when you're trying to search for some crazy thing that you don't have access to. Number three is do-it-yourself science. This is for all the builders and engineers out there. I've met tons of people at the International Science Fair who win the top awards by doing kind of frugal science, where they'll take some super expensive method that costs tens of thousands of dollars. In fact, some of them that I've met, it's like they were trying to do some original project and then they realized like, oh, I don't have this technique. And then they built that whole technique and that's their engineering project. And so I'd highly, highly recommend that. Like if you find some technique that costs tens of thousands, hundreds of thousands of dollars and you're like, there's no way I could do that. But you have some way you think, hmm, what if I apply my machine learning skills plus this frugal science to build this? Go for it. I think that's like an excellent way to do research at a young age. The fourth solution I have for you is to actually just reach out. If you have local universities nearby, you can often just cold email them and you'd be surprised by how many are willing to respond. So just email specific professors, read their literature, read their papers so you understand what they're working on and send targeted emails to them. If you send 10 to 15, there's a chance that you might get one or two to respond. The other technique I have for you is using online forums like mine. I have a school community, which I'll link in the description down below, and I give free resources for students to do research. The other technique I have is kind of like this library resources. People often don't realize, but their local libraries have a ton of sources, especially for the social sciences and those types of research. So consider just asking your local librarian. The next is like virtual labs and simulation software. I've noticed a lot of students do simulation projects in place of, you know, you can't get access to a real human's tumor, but you could simulate it online. The other thing I'd suggest doing is asking your school or your teachers for their resources. Then you can also check out networking opportunities like attending local STEM symposiums and things of that nature to ask experts. And finally, you can apply for grants. All right. So I do want to speak briefly about this, but when I first started science research, I had no idea what ethics were, what risk and safety assessments are. And so later in college, I started doing neuro ethics research and I found it to be very interesting doing normative writing and studies of that nature. I have only recently started it, and so I'm still a learner in that regard. But so far, my thoughts are that integration between researchers and scientists with ethicists is very important. And so as you grow in your science and research journeys, I think incorporating ethics into your research is very important and specifically looking at the techniques you're doing and making sure that they are ethical and that you're following proper safety and risk assessment guidelines. So in a practical sense, this is kind of like what it would look like. For me, it was very simple at the time because I didn't know too much about ethics, but, you know, must use precaution around computational resources, keep any and all patient information confidential if obtained. You know, my project was dealing with bacterial data. And so, you know, this was something that I just kind of added because I wasn't even sure about what risk and safety would look like. But now, you know, way later, I kind of realize how risk and safety can truly make an impact on our society. And so I highly encourage you to think critically and carefully about it. Some things to get started with are really just looking at this word cloud that I pulled up right here. I think it's important to consider different people's perspectives, especially in like medical research, where there is inherently like different procedures being used by different physicians and different clinicians. And so that's really important. And so oftentimes, those can have societal impacts that affect different people differently, based on their socioeconomic status and other principles. So I think it's really important that whenever you're dealing with anything, be ethical about it and consider truly risk and safety. And finally, one note I'll leave you with is especially because you all are student researchers, if you're dealing with data, make sure that it is accurately represented. Throughout your entire study, you do want to stay clear of plagiarism and things of that nature. You know, it's just not worth it. And it could not only harm your career, but harms like literal people if your research goes on to get published and so forth. So I highly encourage you to stay ethical in your research and factor in risk and safety assessments always. If you're still watching this video, it's clear that you still really, really care about research or science fair. And I want to help you guys as much as possible. So thank you so much for watching so far. And let me help you by teaching you a very cool technique I learned towards my junior and senior year of high school. And that's dealing with this idea of peeling back the onion and just having infinite layers of depth to whatever you do. And that's in this data analysis section. So oftentimes, when people are looking to show that the results are significant, that they're quote unquote, good, they'll report something called a p-value. This is a statistic that allows us to calculate the chance of something occurring. And if it's really, really low, that means wow, our results are significant, right? That's just a rough understanding of what it is. But this technique is inherently flawed. And oftentimes, even if you report a really low p-value of 0.0001, what does it actually mean? What does it actually tell us something about our work? And one of my mentors around three or four years ago told me something that I will never ever forget. And that is carefully looking at what these metrics actually mean. Just because you have a 99.999 number or a 0.00001 number, it doesn't necessarily mean that your results are good. And real scientists will uncover this about your research. So when you submit your paper for peer review, or if you are competing in a science fair, your judges and your peer reviewers will point this out unless you do it properly. And so what I'd recommend doing is peeling back our onion as many layers as possible and just having as many tests as possible. Be as rigorous as possible with your scientific method here. As you guys can see here, I tested on six different metrics. That is really strong. In my previous year's work, I tested it on only one metric. And I got third place at the Regeneron International Science and Engineering Fair. I added five more metrics, improved my method, and curated a benchmark set that I then compared my method on compared to other optimization methods. And it performed even better. I got first place. So as you guys can see, adding more layers to that testing process is really essential and it'll help you out a ton. All right, we're finally here to the last piece, and this is compiling your literature. So if you search up Rishabh Jain Literature Review on Google, you will literally come up with my video where I show you step by step on how to do a literature review. I'll show you the best software, how you should find research papers and past works related to your research question in the most effective way possible for students. This is a 23-minute video guide that you should 100% watch. But in case you've already watched this, I'm going to give you something that I learned more recently. As you guys can see here, this is from, again, my winning ISF project back in high school, where I had 30 whole resources crammed on this tiny page. Well, what I learned about a year later in my senior year of high school is that having more sources is not necessarily better. I think in middle school and early high school, people are like, let me just cram as many citations or as many sources as possible because it looks better. But oftentimes, it distracts from the point you're trying to make. And so what I'd recommend doing is you carefully consider each of the sources that you're including. Rather than just pasting in 30 and saying that, hey, this looks really good because I've done more research. What you could instead do is focus in on 10 and actually memorize the names of the authors who are on those research papers, as well as the years that those papers are published in. This will allow you to effectively prove your credibility to the judges, whether it's in a science fair symposium or a science fair judging session, or even when you're writing your research paper, you will remember those initial papers that you cited to come up with your actual project. So I'd highly recommend that you instead focus this down to a list of around 10. And you would call these select references or salient references so that the reviewer knows that you've actually looked through many papers, but these are the 10 most important ones. As you go on to submit your research proposal to grants and things of that nature, oftentimes they will have a limit on the number of sources that you can even include. So if you're still watching this video until the very, very end, you are quite literally part of the 1% of people who are genuinely driven to do research. You're not just clicking on these videos and then clicking off and saying, hey, I'm not going to follow through with any of this. You're probably actually trying to act upon it. And that's why you spend so much time on this video. And to you, I really, really appreciate that you decided to spend this much time because not only are you supporting me and my channel, but you're supporting yourself. I genuinely think that this advice will take you pretty far in your journey. And I want to further provide you as much advice as possible. And so for that, I'd recommend two things. Number one is that you join the new school community that we created. This is linked in the description down below. Basically, I'll be posting these kind of full length guides in that school community along with just genuinely useful advice. There's a lot of forums online these days that just waste a lot of time, but we want to create a really actionable space for high achievers to get things done. And that's what that school community is all about. The second thing is I do kind of help students one-on-one as well through coaching and mentorship kind of things. And so I recommend you check out the link also in the description down below through my new company called sciencefair.io, where we're offering a masterclass that I teach on winning science fair, as well as coaching calls for personalized feedback on winning your science competitions. So with that, I hope this was really helpful. And if you want to see another video, I'd love to see your comment down in the description below on what that is all about. And thank you for all of your support that you've given me by spending a whole 30 minutes on this video. With that, I will see you guys in the next one.