Fantastic video. Clear. Didn't rush through it. Absolutely perfect for a slow individual like myself. One question though, when finding degrees of freedom in the part where you rounded up from 26.75 to 27, my instructor would have rounded down from 26.75 to 26, which I never understood. I forgot his explanation for doing that. Is this just a matter of convention or what?
Rounding down is more safe, because if you round up you might end up with an critical value which is based on a higher sample then you actually have in your question.
Dear Professor Eugene, I love the Yellow Blood Transfusion Ruler as well. I hope I can become a donor very soon too. You are also blood donor who is giving blood (and thereby, life) to our knowledge
Hi medi lectures, The t tables that I use show critical values for both one and two tail tests. Examine the top left of the table at 11:35 in the video and you'll see what I mean. I choose the 0.05 column for the two tail test. Hope this helps, Dr E.
Good work. Please if u have a question to check if the weights differ btw 11 men and their wives, which T test statistic will u employ? Is it a paired t test or unpaired t test and why?
sir thank you for the great vid, it really helped me a lot. I kinda have a dumb question, I read in another source that in reporting the t-test results in apa style there is a "p" in it. However in your video you used "a". Should I change "a" to "p" or just keep it that way? Thanks in advance!
Hi again Prof O'Loughlin, What if the value for the degrees of freedom turned out to be something like 26.25 - closer to the lower bound, 26 and not the upper bound, 27, please? Do we still round 26.25 UP to 27 or DOWN to 26, please
This is probably a personal call, but in the case of fractional degrees of freedom, my general rule is to always round down regardless. This ensures that the calculations are carried out with a slight error on the side of caution, as a lower number of degrees of freedom corresponds to a higher critical value given the same alpha level. So by rounding down, you’re requiring a higher degree of difference to be obtained before you reject the null. You’re being conservative, lowering the chance of wrongly concluding that a difference exists (type I error goes down) in exchange for a higher chance of missing a true effect (type II error goes up). In MOST cases, I’d rather take the chance of missing an effect rather than to make a false claim. But the exact nature of my decision would depend on what was at stake.
Isn't it suppose to be a z distribution when the variances is known instead of working with the standard deviation although the samples is less than 30.
look in the f distribution table..using the degrees of freedom of the two samples. To calculate degrees of freedom , subtract one from the sizes of the samples. In our case, the DF's are 18 and 16.If u look in the table using these values, u get value aprox to 2.25
Eugene O'Loughlin Sir while calculating degree of freedom you have divided by n-1 in denominator of each( variance/n) but in the formula you have divided just by n instead of n-1?
Hi Food Scientist, Check the video at three minutes where I show a diagram illustrating the formulas to use in t-Tests. The "df = n1 + n2 - 2" formula can only be used when the Variances between the two groups are equal. When Variances are not equal, we have to use the more detailed formula to take account of this. It's important to get the right formula. Hope this helps, Dr E.
Hi 000 000 I assume you mean the p < 0.05 at the end? If so, this 0.05 is based on the alpha value set after the hypotheses are stead. You decide what value to use. Dr E.
Thank you so much sir for all doubts clear of independent t test calculation by this video, i am very helpful 😊
Fantastic video. Clear. Didn't rush through it. Absolutely perfect for a slow individual like myself. One question though, when finding degrees of freedom in the part where you rounded up from 26.75 to 27, my instructor would have rounded down from 26.75 to 26, which I never understood. I forgot his explanation for doing that. Is this just a matter of convention or what?
Rounding down is more safe, because if you round up you might end up with an critical value which is based on a higher sample then you actually have in your question.
Dear Professor Eugene, I love the Yellow Blood Transfusion Ruler as well. I hope I can become a donor very soon too. You are also blood donor who is giving blood (and thereby, life) to our knowledge
:-)
Great explanation , helped a lot , thanku sir!!
thank you sir. it helps me alot in my lockdown term..
Thank you for posting this. Is this test the same as the Welch t-test?
Why did you not dividing 0.05 into two for this two tail test during the t crit calculation using the t table
Hi medi lectures,
The t tables that I use show critical values for both one and two tail tests. Examine the top left of the table at 11:35 in the video and you'll see what I mean. I choose the 0.05 column for the two tail test.
Hope this helps,
Dr E.
@@EugeneOLoughlin yeah, Dr E I understand
Thank you SO much for your channel! it is SO helpful to me:)
Very clear and very useful!
Good work. Please if u have a question to check if the weights differ btw 11 men and their wives, which T test statistic will u employ? Is it a paired t test or unpaired t test and why?
This video has helped me so much. Thank you!
thanks for your video... this help me to easy understand
sir thank you for the great vid, it really helped me a lot. I kinda have a dumb question, I read in another source that in reporting the t-test results in apa style there is a "p" in it. However in your video you used "a". Should I change "a" to "p" or just keep it that way? Thanks in advance!
Hi again Prof O'Loughlin,
What if the value for the degrees of freedom turned out to be something like 26.25 - closer to the lower bound, 26 and not the upper bound, 27, please?
Do we still round 26.25 UP to 27 or DOWN to 26, please
This is probably a personal call, but in the case of fractional degrees of freedom, my general rule is to always round down regardless. This ensures that the calculations are carried out with a slight error on the side of caution, as a lower number of degrees of freedom corresponds to a higher critical value given the same alpha level. So by rounding down, you’re requiring a higher degree of difference to be obtained before you reject the null. You’re being conservative, lowering the chance of wrongly concluding that a difference exists (type I error goes down) in exchange for a higher chance of missing a true effect (type II error goes up). In MOST cases, I’d rather take the chance of missing an effect rather than to make a false claim. But the exact nature of my decision would depend on what was at stake.
Isn't it suppose to be a z distribution when the variances is known instead of working with the standard deviation although the samples is less than 30.
How did you get 2.25 for the critical value?
look in the f distribution table..using the degrees of freedom of the two samples. To calculate degrees of freedom , subtract one from the sizes of the samples. In our case, the DF's are 18 and 16.If u look in the table using these values, u get value aprox to 2.25
Hello Sir I think this one is two tailed test so F critical value should be 1.38. Can you please check and confirm
Thank You
please thanks for the video, my question is how did you determine the F critical since you did not have degrees of freedom or alpha level?
Hi Stephen,
I have not shown how to do this in this video - please see a separate video (ruclips.net/video/-ukqBN5sQSI/видео.html).
Dr E.
I was asked a similar question by someone, We should say that we fail to reject H0, but we should never say that we accept H0.
Eugene O'Loughlin Sir while calculating degree of freedom you have divided by n-1 in denominator of each( variance/n) but in the formula you have divided just by n instead of n-1?
@@AnujKatiyal hi, anuj. A very important point made...
Why we didn't calculate the df as n1+n2-2
Plz explain this
Hi Food Scientist,
Check the video at three minutes where I show a diagram illustrating the formulas to use in t-Tests.
The "df = n1 + n2 - 2" formula can only be used when the Variances between the two groups are equal. When Variances are not equal, we have to use the more detailed formula to take account of this. It's important to get the right formula.
Hope this helps,
Dr E.
Null Hypothesis are not "accepted", it is more properly to say that they are just "not rejected"
fail to reject*
how can use Vlookup in subtotal excel file
how do u get p value?
Hi 000 000
I assume you mean the p < 0.05 at the end? If so, this 0.05 is based on the alpha value set after the hypotheses are stead. You decide what value to use.
Dr E.
you beautiful man