 Welcome back everyone. This is the third video in the series about doing research or conducting independent research Today we're going to be talking about the scientific method or the method of actually Conducting research the general method that applies to finding new knowledge, okay, so In the last video you might have seen We were talking about asking questions or getting in the habit of asking questions about things in our everyday life Things related to our our topic that we want to research But just basically getting in the habit of asking questions about Everyday objects that way we're not taking anything for granted or at least we know Where our knowledge stops for normal objects? Once you start thinking in that mode And I recommend you go watch the video if you haven't yet Once you start thinking in that mode, then you will start to question Or you'll start to realize where your knowledge ends Also in your area of expertise or whatever area you're working on Where your knowledge ends is where you can start asking questions and questions that are hopefully relevant and interesting You're essentially finding by doing this questioning process. You are finding a problem a problem so you're finding a problem and This problem is essentially in the form of a question, right? You you want to know I think my example was about markers. You want to know how does the liquid stay in a marker? If you don't already know that well, that's a problem that you can solve So what you're actually doing is having making some sort of problem statement Right, so some sort of problem statement that we're going to be working on You have a problem in the form. Hopefully of a question. So how does Fluid Stay In a marker, right? So we have some sort of problem statement or a question that we're asking This is something it's not a yes or no question, right? It is something where we're actually understanding more about the way that fluid stays in a marker for example But If you don't know the answer to this already It's not very obvious how you would find that out unless you make your own experiment or somebody else knew how to do it Next once we have our problem statement and we have a question or something that we want to solve then we make a Hypothesis or a guess about What the answer to this question is so actually I should number this one to hypothesis So hypothesis here, you might have I hear people all the time saying I have a theory but actually what they're talking about is some sort of guess a theory The way that most people use it is just saying a guess hypothesis Is actually a guess about what the answer is based on? My limited knowledge or the knowledge that I currently have so I make some sort of hypothesis How does fluid stay in a marker? Okay? Well now I have to make a guess about how does the fluid stay in a marker? So how does the fluid stay in a marker? I think the tip is Made of I don't know Fluid repellent material I don't something like that. I don't even know what a hypothesis here would be so tip repels Fluid now I'm already pretty confident that that's not true because I can write on the board right so tip repels fluid Okay, so the tip of the marker repels fluid so all the mark fluid stays in the marker Okay, it's reasonable Except the fact that we can actually write on a white board which already kind of proves us false But we'll get there so I have some sort of problem that I want to answer and then based on that question that I'm answering I make up some guess. Okay, and that guess Doesn't necessarily have to be right. Hopefully you're right, but you still have to show that you're right even if I say the tip repels fluid well that is It's possible. It's feasible like it's something it's something that could be true right But that's actually not how a marker stare the fluid stays in a marker now the thing about a hypothesis a really good Hypothesis is a true or false statement. So here tip repels fluid I framed this in terms of true or false see the true or it's false does the tip repel fluid true or false Now you could get into some some details about some parts of the tip do repel fluid other parts Do not so it's kind of true and false at the same time But basically what we want is a hypothesis that is either true or false There's nothing really in the middle if there is something in the middle that means your hypothesis is not specific enough Okay, and then you probably have to come up with another hypothesis or a sub hypothesis to prove and then that hypothesis Should also be true or false Okay, so we want our hypotheses all hypotheses whether it's the main or the sub hypothesis To be true to be a true or false question. It should be falsifiable. Okay Now the tip repels fluid Do I really care if that's true? Well, I hope that I'm my guess is right But I don't really care if this guess is right or wrong Because once I go through this whole process, I will know the truth Okay, so even if it's not that the tip repels fluid Well, that probably means that I've found the real reason and the real reason might be actually more interesting than what I've come up with here So I'm not I don't really care if my hypothesis was true or false. I want to know more about the issue How does the fluid stay in a marker? Okay, so that's really what I want to answer and then my guess is this if my guess is wrong that I can make another Hypothesis but very often through trying to show that my hypothesis is correct by trying to report or deny it I Often find why the fluid stays in the marker as a as a different reason So I'm not too worried about the hypothesis right now. It's a really important starting point But if I prove it wrong, hey, that's also something That's also something potentially interesting. Okay, so the next or third I should say Once we have our problem statement something we want to find We have a guess about what the answer what the reason is that we're looking for in a true or false statement Then the next thing we should do is go do background research And I will have another video By background about about background research completely separate because it is such an important topic That a lot of people just kind of skip over because they don't They don't really see the need to do background research, but it is very important And if you can get background research right your your total research Will be much more amazing actually much more interesting because you will actually push Where knowledge is rather than being a little bit behind where everyone else is so background research is basically looking up What every? Everyone else has talked about in relation to Fluid staying in the marker now what most people do is say okay I want to know about fluid the ink staying in this marker So I'm going to go to Google and search for how does ink stay in a marker? And then I find every paper with a title ink stays in a marker The problem with that is that maybe you didn't search for the right thing Maybe authors aren't actually talking about ink staying in a marker But maybe they're talking about ink staying in a pin So for example I'm talking about my case is a marker But another author that we found their paper or that we missed their paper Let's say was actually talking about ink staying in a pin right so a pen And a marker let's say the mechanism for the ink staying in them was the same But because I was focused on my marker hypothesis. I completely neglected Pin I say ah pins aren't related to markers. How can they be related? No way Well, actually, maybe they're talking about exactly the same thing just a different case of the same thing so What tends to happen is that people have a couple key words in mind, which is a good place to start Key words in mind to look for background research that's related to what I want to talk about So we have key words related But then we also need to look for different concepts That are related to those keywords, but maybe not in the same area as us Okay, so in this case, I'm focused on marker So my keyword would be marker, but if I look about research if I look at research about pins Maybe I find something relevant to my marker research and very often we do In digital forensics, for example, if you're talking about forensics or you're talking about the way that data Is structured in a computer? well looking at information theory looking at compression algorithms Even though it might not be directly related to Digital forensics the way that those algorithms work or the way that those areas deal with data Is very related sometimes to what we do in forensics even though It's not exactly the same thing looking at those areas might be related Dealing with like sociology dealing with people Very often it's somehow related to digital forensics because people are the ones using the computers So if you're more technical you might focus on actually the data side if you're more investigation focused You might focus on the society side. How do people use computers? But then if you're talking about society then psychology can come into play so you can look at psychology papers and apply it back to digital forensics as well, so My point here is that whenever you're doing your background research. Don't just focus on One particular thing. This is what I'm researching and this is the only thing that's related to it. No Many things are related to your topic and you should know a lot about this particular Your particular area you should know a lot about your particular area But also try to study outside of your area as much as possible and see how that relates in okay That again goes back into asking questions about Basically everything not just your area now. You do have to focus otherwise. You'll never finish a paper, but Try to gain as much knowledge outside as you can and Background research again is one of the most important things you can do. I'll talk a lot more about it later Okay, so once we do background research, let's say that we find out that's people Have been doing some research related to this but they didn't actually find or they're not actually researching exactly what we Want to find or we couldn't find any research on that We found some things that are related to it and maybe they help us to structure some sort of experiment so Once we have our background research then we start If we can we try to use background research as a jump point if we can't then we start to structure our experiment Okay, and this experimentation is basically data gathering Okay, so background research and then we start Constructing our experiment, which is a whole process by itself We'll talk about experimentation a little bit later, but building up your experiment and gathering data from that experiment because we need to prove or we need to show Try to show that our tip repels fluid. We need to show whether that statement is actually true or false Okay, if the statement is false Then that does not answer our question if the statement is true that partially answers our question Right, if the tip repels fluid, maybe that's one reason Why the fluid stays in a pen for example, but that's only part of the answer We'll talk more about Comprehensiveness, I guess later Okay, so starting to structure our experiment or design our experiment and then trying to find data from that experimentation Then afterwards once we have our data, what do we have to do with the data? Always analysis Okay, so we capture our data Okay, and then we do some sort of data analysis. What does this data actually mean in relation to our original question? Sorry in relation to our hypothesis, and what does our hypothesis mean in relation to our original question basically? gather our data do analysis of the data and then try to either Support or deny our hypothesis notice. I didn't say prove or disprove because it's very hard to prove something Sufficiently think of it like a court a court doesn't know that somebody Murdered somebody else, but they have evidence that supports that that person did murder them. They'll probably find them guilty So we're trying to either support or deny our hypothesis And with the data we've gathered through our analysis. Okay, and then the last part of this is Reporting so doing some type of reporting or really what this means is kind of Reflecting on what all of this actually means in terms of our original question. So now our hypothesis What does this analysis mean for our hypothesis? The reporting stage is normally where If this alone was interesting enough where you would be writing papers things like that But normally what happens is after your analysis You get more questions Okay, so after your analysis If you're very lucky and you've already structured everything properly you have a good answer All right, you already have an answer to what you're trying to to show But what tends to happen is that after your analysis you will generate more questions About whatever whatever it is that you're trying to show so your reporting stage then is Reporting about what these questions are and then figuring out how it's going to fit With our problem statement and do we need to generate new hypotheses? Normally at this stage we come up with a new Hypothesis Okay, if we don't come up with a new hypothesis and we probably answered what what research question we had sufficiently enough at least But this whole process just keeps going until you're satisfied with whatever the original problem statement or problem area was So this is very basically the scientific method Okay, so very basically the scientific method come up with a problem statement What problem are we working on make a guess about the answer to the question that's being asked? Do background research to see if anyone else has already answered that question or can anyone else provide? Some support for you to start from maybe they've already planned an experiment or something like that that you could reuse to try to solve your problem and then Do experimentation or design your experiment experimentation method and gather data so do experimentation gather data from that experiment then analyze the data Through this analysis, you'll probably You're trying to figure out. Do you support or deny your hypothesis? Most likely you will have more questions after your analysis so you need to do reporting either produce the results of your experimentation or and Generate new hypotheses to try to talk more about your problem statement So this is just a repetitive process. You just keep going over and over this until you are satisfied With your Answer to your problem statement and by satisfied. I mean you have enough data to support or deny what you think is true I don't mean you found data that says that you are definitely right I mean that you found the truth related to that question, okay? At no stage do I actually Have a particular opinion about what something should be I have a guess about what it could be But if my guess turns out to be completely wrong Hey, I at least I learned something and that might be worth reporting either in a journal paper in a thesis in a blog post Something like that, right? So as you do experimentation try to report or give that knowledge to everyone else because People do use it people do gain a benefit from it even if you find out that what you thought was true is not true So let's say that your hypothesis is this the tip of this pin repels fluid And then you find out that actually I have no evidence to support that the tip of this pin repels fluid well Someone else might come along in six months and say I have an idea the tip of pins repel fluid Right, then they're going to do their background research And if you did not talk about your negative results or That this was found to not be true if you'd never posted that anywhere Then the next person that comes along is also going to do the same research, right? So reporting is extremely important for getting the information out there Even negative results even though journals don't publish negative results very often. It's still very important so I recommend running blogs or just Writing a paper and putting it on archive or just on the internet somewhere and people will find it and it will be helpful to somebody Even if it's only one person you've helped somebody and you need to be documenting all of this anyway for your own research So your documentation is already there. It'll help you to document better It'll give you practice on writing papers and you might actually help somebody in the future Even if it didn't prove what you wanted it to prove. Okay, so negative results extremely important I have no bias about my hypothesis and what it should be Yeah So this is the general process if you stick to this You should you shouldn't go too too wrong the places where you can get mixed up and Yeah, the places where you tend to get mixed up first off Problem statements are normally way too general the ones that students usually make that I see problem statements are very vague very general Not actually answering or asking a specific question Then hypothesis is Also too general it's not specific enough. There's no way to test it. There's no way to measure it It becomes very difficult to Even understand. What are we actually testing? If we make a hypothesis that's too vague So you need to be pretty specific about your hypothesis as specific as you possibly can be I should say Background research the biggest problem here is it's not comprehensive enough Most people do a Google search and then find maybe you know five papers skim them maybe look at the abstracts or something like that and then that's their background That doesn't help you at all. You should be reading papers as much as you possibly can And your background research should be as comprehensive as possible. That will give you the best Benefit for your own work. That way you don't end up doing what somebody else has already done before and There's so many benefits to background research, but anyway, most people don't do a comprehensive background background research then with experimentation and data gathering the biggest challenge here is designing The experiment designing the experiment properly. So it's actually answering or finding data Related to your hypothesis without being biased and hopefully as correct as possible. So experimentation design is a huge area that is People think is easy, but it's more difficult than it seems okay So experimentation a bit of a bit tricky there analysis A lot of people are not very good at statistics actually. So once you gather your data Doing an analysis over that data and doing statistics and Actually finding out how your statistics or how your data might be biased in some way that's not as straightforward as you think depends on the types of studies you're doing but Statistics is usually where most people get kind of caught up on the analysis side So much so that most a lot of people don't really do statistics over their data They'll just say yeah, I saw what I wanted to see there it is but in most cases that's not comprehensive enough and you do more studies to actually show that this is a Normal thing or not. Okay, and then reporting The biggest problem that people have with reporting is not Documenting enough so through this whole process a Researcher should be should have a researcher notebook and should be writing down every single thing that they're doing Every single thing that they're thinking how did they do background research? How did they design their experiment all of those things should be documented now those documentation that documentation is just your notes I'm doing this today today. I had this idea. Maybe I should do my statistics this way right Those notes and actually what you did you should be documenting them all the time The biggest problem that people have is not Documenting everything all the time. So whenever they go back to write their reports. They think How did I design that experiment exactly? Was it a five or was it a six? Right, so then what ends up happening is in their reporting They're not specific about the way that they set up their study Because they don't really remember they wrote a program a long time ago And it worked for them. So yeah, you guys can get the same results as me just Do with the same way I did it, but I'm not gonna tell you how do I did it most of the papers that I read I cannot Verify their findings they just tell me this is the truth But there's no way to verify it because they didn't tell me how they got that truth or the data is not available And you can't be reproduced They didn't tell me how they're did their analysis or exactly how they set up their experimentation all of that should be in your documentation and For your reporting if you're writing in paper you should make sure that everyone else can do the same study as you It's getting a lot better now that we can actually write scripts and then upload them to github Then people can see your scripts and possibly see your data Download it from github and then run the experiments themselves. So Reporting the biggest issue is that people didn't document properly and then their reports or their paper is very Vague on how they did something. Okay, so those are some catches To think about whenever you're you're designing these for your problem statement I hypothesis try to be as specific as possible and as clear as possible So you're not Not too vague For example on this just one example If I say How is this pin? How or how is this marker, right? What exactly am I talking about? How is this marker? Well, I could be talking about the design Right, so the aesthetic design the visual design. I could be talking about the liquid inside I could be talking about the function of the marker But how is this marker is way too vague, right? It could be talking about any aspect of this marker So I need to pick very specific language to describe exactly what I'm looking at About the marker. Okay, so problem statement hypothesis as specific as possible as clear as possible Language is hard to be very clear Really comprehensive background research good experimentation design. We'll talk about that later statistical analysis good analysis, which is normally comes back to statistics so practice doing statistics if you if you're not good at it and then Reporting reporting in a very clear way a very transparent way and a way that can be verified by anyone else who wants to do it Which means comprehensive and complete, okay? And then if you find results that you actually weren't expecting or weren't Hoping for let's say it doesn't it's not consistent with what you thought Still report that somewhere either in a blog post or just a PDF online Or a journal if they'll take it Because negative results are really important for the scientific community Whenever somebody else reads that then they won't make the the same exact study that you made next time because they know that actually this it Wasn't already Was found to already not be supported okay, so this is pretty much it for the scientific method it is kind of a Repetitive flow until you actually get enough data and enough hypotheses supported to To answer your problem statement, okay? It's not just a one one time through finished It's very rarely that way unless you design a very good Experiment or you design a very good I guess piece of research then you might be able to do it one time, but most of time it is a flow You will get more hypotheses more questions as you go along Okay, so I think that's it for today. Thank you very much