Our lecturer gave a good example of distinguishing type 1 and type 2 error: Type 1: A doctor is looking at an old man telling him he is pregnant. Type 2: A doctor is looking at an obviously pregnant woman telling her she is not pregnant.
Thank you, Micheal. Could you please check this. I think I'm right In the first statement, the doctor said the man is pregnant. This makes it positive. We know a man can't be pregnant, making it false. As evident, the doctor's judgement is wrong. The blame is on the doctor. In the second statement, the doctor says a woman with all looks of a 🤰is NOT pregnant. The statement was negative in tone. It was factually wrong too. That is, it is a false negative. Perhaps, the doctor can justify later, the symptoms ( just imagine, don't overthink) were not sufficient. This is beta. Blame is on the data. I'm appearing in a test tom. Not a student of stati. too. Will be happy if you help to oversee this comment. Thank you
@@AbdulRahman-ir5zn In these scenarios the null hypothesis would be that there is nothing in effect taking place. That is, no one is pregnant or that there is nothing significant taking place in the population. In the first situation (type 1 error), the doctor (or researcher) is rejecting the null hypothesis when he or she should not have. So yes, you are right, the doctor is at fault here for rejecting the null hypothesis even though the results clearly suggest there is nothing going on (no pregnancy). Or, this could be from the small chance of obtaining significant results from a sample population even though there is nothing significant taking place in the population. In the second situation (type 2 error), the doctor is accepting the null hypothesis (that there is nothing going on), even though clearly the results/observations indicate that the null hypothesis should be rejected. In a proper research setting, you are right, this would happen if on the slight chance the sample population is showing no relationship between variables when in fact there is a relationship in the population. I hope this helps clarify the difference. Disclaimer: I am no expert on this, just a first-year Psych student.
Here's a better way to remember type 1 and type 2 errors: Type 1 error: think of a girl who apparently likes you and is perfect for you but you reject her (thus rejecting a true null, committing a type 1 error) Type 2 error: think of a girl you like but she isn't right for you say she's a gold digger but you settle for her anyways (thus accepting null when it's actually false, committing a type 2 error) Above statements are statistically wrong if you think about it but are merely for understanding type 1 and 2 errors conceptually. Story credit: Utkarsh Jain, Fintree
Nah this doesn't work. Type one is something boring is happening but you think that something special is happening. In what way is a girl liking you a "false alarm"? This is why your example makes no sense.
For me, best example is like justice system. Null is EVERYONE IS INNOCENT UNTIL PROVEN OTHERWISE. Type 1 is you wrongly reject true null-> send innocent one to prison (false positive) very serious problem. Type 2 is you wrongly accept false null -> believe innocent although person is guilty (false negative) , more like no enough evidence scenario.
The story at the end is misleading. The null hypothesis should ALWAYS be a statement referring to status quo. It should be "She does not like you back."
Agreed. The null hypothesis should be that there is no relationship, or in the case of his example, the girl doesn't care about him. A Type I error is a false positive, in this case thinking the girl likes him back when she actually does not (ie status quo).
You got it backwards. Type II errors are WAY worse. For example, a Type I is when the fire alarm sounds even though there is no fire (false alarm is false positive) A type II error is when the fire alarm DOESN'T sound when there actually is a fire (false negative)
I think your example is looking at it from the consequences of a fire, but you have to look at it too from the null hypothesis. What is the null hypothesis in your example? The null hypothesis is regarded as "sacred" since it is the status quo, and so erroneously rejecting it is serious therefore a type 1 error. Not rejecting the null hypothesis which is a type 2 error is less serious because it could be that there was not much evidence to sway you from that position, in your case no alarm. Take for example a judge who acquits a defendant who is actually guilty of a crime because of insufficient evidence despite popular opinion that the defendant is guilty. You cannot fault the judge because he decided based on scanty or no evidence, therefore it is a less serious error. Start by asking yourself what is the null hypothesis in your example? So from an epistemological or theoretical point of view type 1 error is more serious, but from a practical specific situation the consequences of a type 2 may be more serious. It all depends on where you look at it from, from either epistemological or practical.
I think the example is another way around: if she likes you back and you didn't invite, it is type 2 error - actual class is positive but predicted is negative; if she doesn't like you but you invite her out, it is type 1 error - actual class is negative but predicted class is positive.
Thanks for your video. Unfortunately, I believe your last example is wrong . A null hypothesis by definition assumes that nothing has changed and the status is still the same; so it should be : "The girl don't like you ". Also, Type 1 error is defined as : " taking an action when no action is needed", meaning the student did type 2 error ( Not taking action when it is needed) not the other way around. Nice video, but you got it twisted at the end. I hope you accept my feedback. And correct me if I'm wrong please. Best,
Great job! You have explained a confusing topic in a very simple, short , intuitive and probably taking the best example ❤🤣😂. I'll never forget the difference b/w Type1 and Type2 errors.
Shoudn't H0 be she doesn't like you back (ie, you didn't have an effect on her), since H1 is your desired result that she likes you back (ie, you had an effect on her)?
No it just depends on what you are testing. Are you testing whether she likes or whether she doesn't like you? Either way the null hypothesis is the opposite. So I were testing whether she likes me, my null is She doesn't like me and vice versa
Which Error Is Better By thinking in terms of false positive and false negative results, we are better equipped to consider which of these errors are better-Type II seems to have a negative connotation, for good reason. Suppose you are designing a medical screening for a disease. A false positive of a Type I error may give a patient some anxiety, but this will lead to other testing procedures which will ultimately reveal the initial test was incorrect. In contrast, a false negative from a Type II error would give a patient the incorrect assurance that he or she does not have a disease when he or she in fact does. As a result of this incorrect information, the disease would not be treated. If doctors could choose between these two options, a false positive is more desirable than a false negative. Now suppose that someone had been put on trial for murder. The null hypothesis here is that the person is not guilty. A Type I error would occur if the person were found guilty of a murder that he or she did not commit, which would be a very serious outcome for the defendant. On the other hand, a Type II error would occur if the jury finds the person not guilty even though he or she committed the murder, which is a great outcome for the defendant but not for society as a whole. Here we see the value in a judicial system that seeks to minimize Type I errors.
Re: type II errors, shouldn't you say that you "fail to reject a false null hypothesis", rather than "accept a false null hypothesis"? Since you never accept a null hypothesis, you just fail to reject it.
oh as a dr with little knowledge of statistics this explaination helps a lot for me to understand the basic type 1 vs type 2 error. Thanks and I love that short story!
Thank you for explaining this topic in a way that can be understood without confusion. Now I am going to your channel to see what you have to say about ANOVA. Thanks, again!
Absolutely a convincing example about type1 and type2 errors. But however, if I take h0: Medicine works perfectly truth: It works perfectly. h0 was rejected - type1 error. Now if the truth: it won't work perfectly, and h0 was accepted - type2 error. Here the type2 error may cause too serious health issues if we use the medicine, isn't it? can't we consider type2 error was more serious than type1 error. Any clarification from anyone would be greatly thankful.
Significant propability of making type II error makes experiment results useless aswell. You should estimate necesary minimum number of samples before experiment. And statistical power depends on margin of error too. You choose it which isn't negligible Also accepting a false hypotesis is in fact Type II Error but it is called false positive as You get positive result wrong
I didn't get how rejecting a true null hypothesis is considered false positive? should it be false negative meaning we falsely say something is negative (rejecting it) while it is positive (true) ?
Hey, Himan! The words: "positive" and "negative" do not refer to the situation with the girl. They are just terms (words) used in statistics. Let us consider your statement: "If we consider she indeed likes you back, but we reject H0 (and therefore we think she doesn't like you back), why is this a false positive ? Wouldn't it make more sense to call this a false negative ? In the sense that it looks like a negative, but in fact it is positive." Let's see the etymology of the term: It is called false positive, because rejecting H0 (so rejecting the status quo), means you believe some trait (attribute) is present ( your H1). The status quo is: "she likes you back". The attribute is: "she doesn't like you back". If you reject the null hypothesis of a test, your test is positive!!! So, if you wrongly reject the null, your test is FALSE positive. So we are referring to the test, and not if the question is referring to the " the positive thing". *** In the second case, the null is still H0: She likes you back. This time, we accept the null. So our test is negative!!! However, we wrongly accepted it, so it becomes a FALSE negative. Maybe the example was not the best, but it surely shows that false positives and false negatives do not depend on the question we are asking, but on the type of error. Best, The 365 Team
Yep! The problem is, in the previous video, according to every source I can find online, they flipped H0 and HA. H0 isn't your hypothesis, it's the counter to it. It's the "nullification" / invalidation to your hypothesis. So if his hypothesis is she likes him: H0: she doesn't like him HA: she does like him They show H0 here but not HA, which complicates matters. Because H0 is she likes him, then his hypothesis is 'she doesn't like me'. Let's break this down: true && false might as well be called correct/incorrect. If the prediction is correct -> true. If the prediction is incorrect -> false. positive && negative is what is predicted, based on the hypothesis. HA: she does not like him If yes (she does not like him) is predicted -> positive. If no (she likes him) is predicted -> negative. He predicted she doesn't like him (a "positive" prediction), but she liked him, so his prediction was incorrect (false). ∴ FP (Therefore, false positive.) This double negative is quite a pain to wrap ones head around, but I see why 365 DS did it this way.
You can't specify type 1 is more serious than type 2, it depends on the situation. Foe example: Lab testing for Cancer detection, here there could be 2 types of errors: A Cancer patient not getting treatment A healthy person getting cancer treatment Which one is type 1 error and which is type 2? Here a type 2 or false negative is: Cancer patient not getting treatment which might result in his/her death. Which is more severe, so it depends on the situation.
I think you have made a mistake in considering Type I error as more serious than Type II error . As I have been taught and experimented Type II Error is more serious than Type I Error as we are falsely accepting the null hypothesis.
The examples are the exact opposite of how I think of it. Type I is a false alarm. Conspiracy theorist believe something amazing is happening when really there isn't. Type II is when is when the magic is happening but you don't think anything is happening. You've accepted the boring result. Type II is the one where you miss out. I think of it this way because I think of the null hypothesis as the typical boring situation. Nothing really happened. Thinking something did happen is the error.
I liked this girl during school, and I think I made type 1 error during school by not confessing her. Somehow I was able to confess her after 3.5 years of college, and turns out that I made a type 2 error.
