highest quality statistics videos you have here sir. Thank you. I have come across with a lot of shitty information, even wrong informations for a couple of hours then I found your videos.This is of great importance to me, since I want to learn how to conduct,interpret research.
In the real world a test like this would typically be done as a t test (even if the sample size was greater than 30), in this particular video I've assumed the population standard deviation is known. If the population standard deviation is known, then we'd use Z regardless of the sample size (if we're assuming normality). The population standard deviation is almost never known, but this video comes at a point where I'm introducing the topic, and we have not yet encountered the t distribution.
Oh I see- because p value=0.211>0.05; and because H0= 780; Therefore we should reject H0, which means reject the mean the weight is no significant than 780. Was trying to understand the logic of the very last part of the video...
Sir I don't understand why do we take Ha =/= 780 as the alternative hypothesis rather than Ha < 780 as the question is about how many cereal packs are underfilled? Thanks
In the example, the stated weight on the boxes is 750 g. There is some variability in the weights, and the producer doesn't want many boxes to be underfilled, so they set the mean weight of the process at 780 g (30 grams higher than the stated weight). They feel this strikes a reasonable balance between having only a small proportion underfilled (compared to the stated weight of 750), and unnecessarily using extra product. They then check to see whether the mean of the process has drifted (in either direction) from their desired 780 g.
You said: Note: A p-value of 0.211 does not give strong evidence that the mean = 780. But it is safe to assume that the mean is close to 780, right? Because if we construct a 95% confidence interval around the sample mean, 776: We get the interval (769,782) approximately and 780 is one of the values in that interval.
Hello, firstly I would like to say how much I appreciate your videos. But I am slightly confused about the QQ plot. Anywhere I read about it, it said that it is a comparison of distributions of 2 datasets. However, you're comparing only one of them. What are the "theoretical quantiles"?
I have a video on normal QQ plots here: ruclips.net/video/X9_ISJ0YpGw/видео.html QQ plots can be used to compare two samples, or to compare one sample to a theoretical distribution. In a normal QQ plot, the sample values are plotted against appropriate quantiles of the standard normal distribution. My video on it goes into more detail.
The significance level is the significance level, and I gave it as 0.05 in this case. Since it was a two-sided test, I doubled the tail area to get the (two-sided) p-value, which I then compared to the 0.05 significance level.
That depends a great deal on the specifics of the situation. If we're using the normal QQ plot as a check on normality for a t test, say, and we happen to see some right skewness, then we might consider transforming the data (e.g. using the log of each data value, or square root) and seeing if the transformed data values are approximately normal. Or we might consider using a distribution-free procedure (a procedure that doesn't rely on the data coming from any specific distribution). But as I stated above, it depends greatly on the specifics of the situation.
We only carry out a test if we have a specific hypothesis in mind. We should have this hypothesis in mind before gathering the data used to test it. And we can carry out a test of the null hypothesis that mu = anything, but there is typically a very limited number of hypotheses that have any motivation behind them whatsoever. For the example used in the video, we could just as easily test H_0: mu = 1000 kg as H_0: mu = 776 g but testing the former makes no sense whatsoever in the context of the problem. If a reader has a different hypothesis in mind, they can make the appropriate adjustments to the calculations on their own. If we also report a confidence interval, then if a reader has a different hypothesis in mind they can quickly get a quick initial assessment of whether that hypothesis is consistent with the observed data (based on an appropriate interpretation of the confidence interval, and the relationship between tests and confidence intervals).
Why did we multiply by 2 here ruclips.net/video/Xi33dGcZCA0/видео.htmlsi=i0pT0C9uJvh1oJle&t=288? For that case you should have considered only one side and not both side as 776 was less than 780.
I suggest you limit the "should haves" until you have a deeper understanding of the situation. If your logic here was valid, then there would be no such thing as a two-sided alternative. The value of the statistic will always be on one side or the other, so if we let that determine things we'd always have a one-sided alternative. So your logic can't be valid, or nobody would ever speak of a two-sided alternative. The choice of alternative ******never never never never never ever ever ever ever ever ever**** depends on the value that we see in the sample that we're using to carry out the test. If we do that we're distorting the math and then making false statements in the end. The choice of alternative depends on the nature of the situation at hand. One should be able to write out the appropriate hypotheses without ever looking at the sample data. If the sample data has influenced your choice of hypothesis, then you've violated a fundamental requirement of hypothesis testing.
highest quality statistics videos you have here sir. Thank you. I have come across with a lot of shitty information, even wrong informations for a couple of hours then I found your videos.This is of great importance to me, since I want to learn how to conduct,interpret research.
this is way better that any paid cours or tutorial out there, short, concise and very clear , thanks sir for these high quality video!!
Bro I was able to solve this question on my own bcos your previous videos were great!!!! You saved my life. Thanks alot
In the real world a test like this would typically be done as a t test (even if the sample size was greater than 30), in this particular video I've assumed the population standard deviation is known. If the population standard deviation is known, then we'd use Z regardless of the sample size (if we're assuming normality).
The population standard deviation is almost never known, but this video comes at a point where I'm introducing the topic, and we have not yet encountered the t distribution.
Oh I see- because p value=0.211>0.05; and because H0= 780; Therefore we should reject H0, which means reject the mean the weight is no significant than 780. Was trying to understand the logic of the very last part of the video...
Sir I don't understand why do we take Ha =/= 780 as the alternative hypothesis rather than Ha < 780 as the question is about how many cereal packs are underfilled? Thanks
In the example, the stated weight on the boxes is 750 g. There is some variability in the weights, and the producer doesn't want many boxes to be underfilled, so they set the mean weight of the process at 780 g (30 grams higher than the stated weight). They feel this strikes a reasonable balance between having only a small proportion underfilled (compared to the stated weight of 750), and unnecessarily using extra product. They then check to see whether the mean of the process has drifted (in either direction) from their desired 780 g.
You said: Note: A p-value of 0.211 does not give strong evidence that the mean = 780.
But it is safe to assume that the mean is close to 780, right? Because if we construct a 95% confidence interval around the sample mean, 776: We get the interval (769,782) approximately and 780 is one of the values in that interval.
You are a god amongst men. Thank you
You are very welcome!
I came here in search of bronze. I found gold.
Hello, firstly I would like to say how much I appreciate your videos.
But I am slightly confused about the QQ plot.
Anywhere I read about it, it said that it is a comparison of distributions of 2 datasets.
However, you're comparing only one of them.
What are the "theoretical quantiles"?
I have a video on normal QQ plots here: ruclips.net/video/X9_ISJ0YpGw/видео.html
QQ plots can be used to compare two samples, or to compare one sample to a theoretical distribution. In a normal QQ plot, the sample values are plotted against appropriate quantiles of the standard normal distribution. My video on it goes into more detail.
If we would've choosen rejection region approach then it could've also given us a correct answer(as 1.25
As this is a two-sided test wouldn't our significance level have been 0.025 for which we test our p-value against?
The significance level is the significance level, and I gave it as 0.05 in this case. Since it was a two-sided test, I doubled the tail area to get the (two-sided) p-value, which I then compared to the 0.05 significance level.
how are you concluding that if p=.2 then no evidence for mean differs,. to differ how much p value required?
which software are you using to write on the black board?Please reply
What if the normal QQ plots doesnt show normal distribution. What do we do then ? Anyone pls answer.
That depends a great deal on the specifics of the situation. If we're using the normal QQ plot as a check on normality for a t test, say, and we happen to see some right skewness, then we might consider transforming the data (e.g. using the log of each data value, or square root) and seeing if the transformed data values are approximately normal. Or we might consider using a distribution-free procedure (a procedure that doesn't rely on the data coming from any specific distribution). But as I stated above, it depends greatly on the specifics of the situation.
isnt it suppose to be a t test since the population sample is smaller than 30?
why can't we test all kinds of alternatives and give corresponding p-values and let the readers decide.
We only carry out a test if we have a specific hypothesis in mind. We should have this hypothesis in mind before gathering the data used to test it. And we can carry out a test of the null hypothesis that mu = anything, but there is typically a very limited number of hypotheses that have any motivation behind them whatsoever. For the example used in the video, we could just as easily test H_0: mu = 1000 kg as H_0: mu = 776 g but testing the former makes no sense whatsoever in the context of the problem. If a reader has a different hypothesis in mind, they can make the appropriate adjustments to the calculations on their own. If we also report a confidence interval, then if a reader has a different hypothesis in mind they can quickly get a quick initial assessment of whether that hypothesis is consistent with the observed data (based on an appropriate interpretation of the confidence interval, and the relationship between tests and confidence intervals).
You are very welcome!
Why did we multiply by 2 here ruclips.net/video/Xi33dGcZCA0/видео.htmlsi=i0pT0C9uJvh1oJle&t=288?
For that case you should have considered only one side and not both side as 776 was less than 780.
I suggest you limit the "should haves" until you have a deeper understanding of the situation.
If your logic here was valid, then there would be no such thing as a two-sided alternative. The value of the statistic will always be on one side or the other, so if we let that determine things we'd always have a one-sided alternative. So your logic can't be valid, or nobody would ever speak of a two-sided alternative.
The choice of alternative ******never never never never never ever ever ever ever ever ever**** depends on the value that we see in the sample that we're using to carry out the test. If we do that we're distorting the math and then making false statements in the end. The choice of alternative depends on the nature of the situation at hand. One should be able to write out the appropriate hypotheses without ever looking at the sample data. If the sample data has influenced your choice of hypothesis, then you've violated a fundamental requirement of hypothesis testing.
how does sample mean=776?
The weight of the cereal in the 25 boxes was measured. When the average of these 25 measurements was taken, it was found to be 776 grams.
So the sample mean was given by the question?? But I can't seemed to find it OmO help me sir
It's given on the slide at 1:30.
LOL IM BLIND!! THANKS A BUNCH SIR!! >
@@jbstatistics But initially you told us that sample mean = 750
Thanks for the video
How did you get -1.25?
776-780
_________
16/√25
= -1.25
How? 😭
776 - 780 = -4. 16/sqrt(25) = 16/5 = 3.2. -4/3.2 = -1.25. Or, if you prefer: -4/(16/sqrt(25)) = -4 * (5/16) = -5/4 = -1.25.
Thank you sir 😊 I got it
that was heaps good