00:02 Understanding confounding and effect modulation in medical studies. 01:55 Understanding confounding vs. effect modulation 04:01 Smoking can affect the outcome, exposure, or both, impacting the study findings. 06:08 Confounding affects both exposure and outcome in research studies. 08:19 Understanding key methods for comparing means in research studies 10:11 Chi-square is used to compare proportional data 12:05 Understanding analysis of variance and t-tests 14:06 Understanding effect modulation vs. confounding 15:59 Understanding Number Needed to Treat (NNT) 17:38 Understanding of absolute reduction and number needed to treat 19:21 Effect modulation vs. confounding in research.
Hi everyone, indeed dr randy neil did come down from heaven ! May God bless him on this Friday where ever he is now ...his unique flavor in teaching touched many dr's preparing for the usmle ...
I will post my situation here , as I'm preparing for my usmle , I'm also applying for jobs in clinical research , residency is going be competitive, your thoughts dr randy neil is appreciated as always ...
Thanks Randy. Just to clarify (min 13:30): if the CI does not include zero is significant for a mean value. If CI does not include 1, is significant for OR or RR. : )
This is exactly what confused me! Thanks for pointing it out. Aren't we talking about the RR here as well? So how do we day its significant even though it includes 1? 🤔
Thank you Dr Neil. Im in my first semester of my post grad studies Epidemiology and Biostatistics. Am currently learning Biostastics online (due to the pandemic) and it was just terrible. I couldn't nor understand grasp the calculation concepts at all. I came on UA-cam to look into Biostastics for dummies but came across your channels. Aren't I glad I did! I have been online all day going through your videos and I can say am confident that I can pass my biostat exams. Within a day am now very conversant with how to handle questions starting form the question part, before going through any long prose. I now know what formula to apply in a ,majority of the questions be it NNT, PPV, LR, AR%, when to draw the chi square among others. Thank you so so much. Your videos are extremely helpful. If you could do some videos on Correlation coefficients, confidence intervals and hypotheses, that would be great. So far, am grateful. I have a few questions to share with you for the purpose of helping me and the rest here learn better too. I'm not sure how to do so though. TIA
Dr. Neil is key to success. Thank you for everything you do and for posting these videos. No test prep company or any professor ever explained this to me. You are a true gift to the world. Thank you for sharing your knowledge.
I always watch your stats videos and genetics video before y FRCOPH exams. They are such immense help. I cant thank you enough. sending you lots of virtual flowers and wishes.
First aid has NNT listed as NNT = 1/ ARR with ARR = (c/ (c + d)) - (a/ (a + b)). When I solved the problem using the formula from first aid, I got around 4.2 for my answer. I'm a bit confused on the formula you used to tackle the last problem.
Going through your videos my step 3 is Just a week way.. your videos have made these calculations so easy to understand.. thank you very much and i hope i score well
if only you would be teaching pharmacist for their board exam, they would have passed on a first try with flying scores. thank you for putting time and creating these videos to help us better understand confusing concepts .
this is an amazing video! I kept getting these chi-square questions wrong over and over because I wanted to do it the "proper" school way. Dr. Randy you're a godsend! thank you thank you. Step 1 this Wed!
Michelle, You are nervous because it means something to you... STAY POSITIVE, STAY STRONG... Everything ha already been determined LONG AGO so NO NEED TO WORRY...
In the NNT problem, I believe you had labeled the graph correctly. If we are calculating EVENT RATE, then that means we are calculating who had hair loss in the control group minus who had hair loss in the treatment group. Please correct me if my concept is wrong though my step is in a few days LOL
hi dr. can you please do videos on... a. linear regression vs logistic regression b. kaplan meyer curve c. forest plot d. population pyramid p.s thankyou for your videos:)
Always hated biostats, was never any good at it. Thank you SO much for these simplified videos! Made it super easy! Can't thank you enough! I'm still can't get what's a good confidence interval value and why.
For the most part...think...the Confidence interval that doesn't contain the number "1" is good... 0.55-0.86 is good.... 1.2-3.5 is good... 0.9-1.1 is NO good...
Wow , great explanations and made so easy to understand the concept in ridiculously short time , Dear Dr kindly upload a Neuro video of high yield topics for step 1. Thanking you in anticipation 😊
A better explanation of chi squared would be that it's comparing two data sets where both the independent and dependent variable are categorical (vs categorical independents and measured dependent variables)
wow ! .. actually it should be woowooowoowooowooowo ! u made it like super easy within 22 min video ! I wana be your life long student , Sir . u r really great , hero of biostat I must say . i think everyone will agree with me too.
Smoking affects exposure AND outcome (both of these variables) which is by definition CONFOUNDING BIAS. People who drink alcohol may also smoke (distorting exposure values) and people who smoke may also get pancreatic CA (distorting outcome values). This cumulatively distorts the relationship bridge between smoking and pancreatic CA and confounds the results. Smoking is an effect modulator with DVT as it only affects the outcome which is by definition EFFECT MODULATION. Think about it...smokers are vasculopaths who are inherently in a hypercoagulable state and thrombose easily) but there is NO clear association between the NMDA drug and smoking.
is there a mistake ? For the question with smoking and prostate ca, if the p values were not statistically significant due to the p-vaules doesn't it mean the results are not valid in the stratification and that there is NO difference, no effect mod, and No confounding? thx
I was thinking the same that when one study is so statitically insignificant, how can we base our assumptions on that. Rather, it might be that the researcher had the predilection towards proving the association between the drug and DVTs, so he carried out an insignificant study to support his belief. In that case, the right answer would be 'confirmation bias'.
Hi Dr. Neil. one more question : A 26 year old medical student comes to the physician with a 3 week history of night sweats and myalgias. During this time, he also had a 3.6 kg weight loss. He returned from a month tropical medicine rotation in Cambodia 1 month ago. A chest x ray shows reticulonodular opacities suggestive of TB. The student is curious about his likelihood of having active TB. He reads a study that compares sputum testing results between 2800 patients with likely active TB on a basis of history, clinical symptoms and CXR pattern and 2,400 controls. The results are: Active TB likely on basis of History, clinical and CXR Active TB not likely on basis of History, clinical and CXR Sputum Test positive 700 300 Sputum Test Negative 2100 2100 What is the probability that a patient with a diagnosis of active TB on the basis of history, clinical symptoms, and CXR pattern actually had active TB? The actual answer is 0.25. But my calculation is 0.70. Would you please explain this? Thank you
Hi, the given table is the confusing aspect. They gave you the flipped table. The top heading should be the gold standard (Sputum test) and the side heading should be for the test conducted ( History, clinical and CXR) . You should rearrange and construct the chart like this: Sputum Test positive Sputum Test Negative Active TB likely on basis of History, clinical and CXR 700 2100 Active TB not likely on basis of History, clinical and CXR 300 2100 Then, PPV is 700/(700+2100) = .25
i dont understand the 2nd question on confounding and effect modulation. in the second half of the question statement where they corelated the effects of smoking with both alcohol consumption and pancreatic cancer the test results are of no significant value. both are more than acceptable 0.05. meaning that smoking is not related to either of them.
Hi Dr Neil, I have a question regarding Q2, in second part of the question the P value is high so wouldnt that means it is not significant and hence being a smoker or a non smoker doesnt really make a difference? am I missing out something here or am I not getting something and I have spoken something beyond stupidity? Would be nice if you shed a light on this. Thank you so much for your videos your work is wonderful!
your videos have been so amazing thank you for your hard work! hoping you have time to upload a short one on population pyramid. happy holidays doctor Randy !
Amazing! Thank you so much, specially the ANOVA, T test, and chi square!. I had a question about Q2 however; if they re-examnined the findings, and found that smoking/nonsmoking was only associated with a increase/decrease in smoking, wouldnt that be an effect modification as it only affects the outcome (of pancreatic cancer), and not the exposure (alcohol)? I am a little confused on that one
Saba, Thank you for the question... The take home point is that we know there is an association with BOTH exposure and outcome (for Confounding)...and only with OUTCOME for effect modulation. Everything else is "smoke and mirrors". Just take a step back and see it "big picture"... not sure if I am making sense...but for now, just know principle between the two.
Hi Dr.Neil, can you do a video that involves the large 2 page pharm ads and how to navigate that? I get so intimidated. I'm trying to take the step III asap....my confidence needs the most work.
Thanks so much Would you plz go over this question The new diagnostic test was developed to aid in diagnostic GERD. Of 500 patient with GERD , the new diagnostic test identified 300 true positive. Of 500 patient without GERD , the new diagnostic test identified 215 true negative what is specificity ? I had FP as 500-215 = 285 So my answer was 215/215+285 But the explanation answer had FP as 500-300=200 Can you plz plz explain? Also would you have any lecture on different kind of test ? Cost effectiveness, minimization, utility , and benefit ?
Poneh, Based on your information, your work looks right and you know where the numbers go... It's possible there was an error in whomever made the question. It's possible...
Hi Dr. Neil, thank you for the videos. Would you please explain this problem: A recent study was conducted to assess the intelligence of students enrolled in an alternative high school program. The results showed the IQs of the students distributed according to the normal curve, with a mean of 115 and a SD of 10. Based on this information it is most reasonable to conclude that a) 50% of the students will have the IQ below the standard mean of 100 b) 50% of the students will have the IQ below 105 c) Students with IQs of 125 are the 84th percentile d) 2.5% of the students will have IQs greater than 125 The answer is b. Would you please explain where did 84th percentile come from? Thank you!
That is a rule of thumb: Mean +- 2SD = 95% confidence interval Mean +- 1SD = 66,6% CI In a Normal Distribution, mean = median = p50 An IQ of 125 is the mean + 1SD. So it is p50 + (66,6/2) = 83.3% (roughly 84%). We divide 66.6 by 2 because we are only adding 1SD, not adding and subtracting, get it? Remember: - 2SD = p5 - 1SD = p17 Mean = p50 + 1SD = p83 + 2SD = p95 + 3 SD = p99
These videos are awesome ❤️👏🏻 i just wanna ask u if we compare two mean values , when will it be regression analysis and when will be it two sample t test?? Because its so confusing
00:02 Understanding confounding and effect modulation in medical studies.
01:55 Understanding confounding vs. effect modulation
04:01 Smoking can affect the outcome, exposure, or both, impacting the study findings.
06:08 Confounding affects both exposure and outcome in research studies.
08:19 Understanding key methods for comparing means in research studies
10:11 Chi-square is used to compare proportional data
12:05 Understanding analysis of variance and t-tests
14:06 Understanding effect modulation vs. confounding
15:59 Understanding Number Needed to Treat (NNT)
17:38 Understanding of absolute reduction and number needed to treat
19:21 Effect modulation vs. confounding in research.
Thank you...
@@RandyNeilMD no problem sir
Man, these videos are gifts from the divine. Thank you, Dr. Neil!!
Sky, You are most welcome. Glad it was helpful.
How is there not millions of views on this. I tried EVERYTHING to understand and remember biostats. THIS DID IT FINALLY.
Wohoo!! Perfection!!
Thank you, SIR!!!!
future doctors owe it to you.
Help the next guy in line...
Hi everyone, indeed dr randy neil did come down from heaven ! May God bless him on this Friday where ever he is now ...his unique flavor in teaching touched many dr's preparing for the usmle ...
Thanks Jameel... Appreciate the kind words! KEEP STUDYING HARD!!
By any chance , is it OK to share your email, any case any dr want to reach out and seek your advice on any matter relating to their journey!
I will post my situation here , as I'm preparing for my usmle , I'm also applying for jobs in clinical research , residency is going be competitive, your thoughts dr randy neil is appreciated as always ...
@@jameelashkar9244 sure... randy.neil8@gmail.com
Thanks, Dr. Neil for your dedication and help to all of us!
Thanks Randy.
Just to clarify (min 13:30):
if the CI does not include zero is significant for a mean value.
If CI does not include 1, is significant for OR or RR. : )
This is exactly what confused me! Thanks for pointing it out. Aren't we talking about the RR here as well? So how do we day its significant even though it includes 1? 🤔
Thank you so much. Tying to beat this beast of an exam is not easy, so these videos are a lifesaver!
Glad it has been helpful.
Thank you Dr Neil. Im in my first semester of my post grad studies Epidemiology and Biostatistics. Am currently learning Biostastics online (due to the pandemic) and it was just terrible. I couldn't nor understand grasp the calculation concepts at all. I came on UA-cam to look into Biostastics for dummies but came across your channels. Aren't I glad I did! I have been online all day going through your videos and I can say am confident that I can pass my biostat exams. Within a day am now very conversant with how to handle questions starting form the question part, before going through any long prose. I now know what formula to apply in a ,majority of the questions be it NNT, PPV, LR, AR%, when to draw the chi square among others. Thank you so so much. Your videos are extremely helpful. If you could do some videos on Correlation coefficients, confidence intervals and hypotheses, that would be great. So far, am grateful. I have a few questions to share with you for the purpose of helping me and the rest here learn better too. I'm not sure how to do so though. TIA
Clara, Glad it was helpful...
Gift from God. Thank you for making these concepts simple!!!
Thank you for the kind words Yash
Dr. Neil is key to success. Thank you for everything you do and for posting these videos. No test prep company or any professor ever explained this to me. You are a true gift to the world. Thank you for sharing your knowledge.
Thank you for the kind words... KEEP MOVNG FORWARD. Keep me posted...
Great, short and succinct. Thank you, Professor Neil!
Amazing Video for Step 2 CK!
You are welcome.
Amazing video. Your dedication to medical field is greatly appropriated!
Princess, Thank you for the comment.
I always watch your stats videos and genetics video before y FRCOPH exams. They are such immense help. I cant thank you enough. sending you lots of virtual flowers and wishes.
Thank you Kiva for the kind words.
The best way in learning is to solve questions! Thanks for making it easy
Asmaa, Glad it was helpful.
just found ur videos one month b4 my exam and stats were my weakest subject and its the oppsite now
PERFECT!!
Thanks Dr.Neil. I have tried alot sources but yours was best one.
Thank you for the comment.... Glad it has been helpful.
Great job Dr.Neil ,simply amazing.Huge respect for u
Mish, Thank you for the comment
Thank you so much another extraordinary video on biostatatistic, you are remarkable Dr.neil.
Musawair, Thank you for your support. Much appreciated.
Your amazing i feel so much confident about what I've learned , so very greatful.
Thank you for commenting
First aid has NNT listed as NNT = 1/ ARR with ARR = (c/ (c + d)) - (a/ (a + b)). When I solved the problem using the formula from first aid, I got around 4.2 for my answer. I'm a bit confused on the formula you used to tackle the last problem.
Hey, just wondering if there is any follow up on your query?
Going through your videos my step 3 is Just a week way.. your videos have made these calculations so easy to understand.. thank you very much and i hope i score well
Went well?
You are the best out there, Dr Neil. Stay awesome ⭐️
Thank you... I do appreciate the feedback.
if only you would be teaching pharmacist for their board exam, they would have passed on a first try with flying scores. thank you for putting time and creating these videos to help us better understand confusing concepts .
Par Rez, Thank you for taking time to comment. I am humbled by your comments.
this is an amazing video! I kept getting these chi-square questions wrong over and over because I wanted to do it the "proper" school way. Dr. Randy you're a godsend! thank you thank you. Step 1 this Wed!
I needed this!!! Taking Step 1 soon and Im super nervous and anxious about it
Thank you for this videos!!
Michelle, You are nervous because it means something to you... STAY POSITIVE, STAY STRONG... Everything ha already been determined LONG AGO so NO NEED TO WORRY...
Ur videos have been of great help to me in my biostatistics/epidemiology 👍
Ikechukwu, Thank you for taking time to comment. Glad it was helpful...
Thanks a lot Dr. Neil!! I am very much relieved now
Perfect.
This video is super helpful Dr Randy! Thank you so much.
Anusha, You are most welcome.
Your videos should be an A plus resource for biostats in the First Aid section for biostats :)
Ali, Thank you for the kind words.
Thank you 😊.I was so confused about confounding,this was helpful.
Glad it was helpful!
In the NNT problem, I believe you had labeled the graph correctly. If we are calculating EVENT RATE, then that means we are calculating who had hair loss in the control group minus who had hair loss in the treatment group. Please correct me if my concept is wrong though my step is in a few days LOL
I'm confused as well
You are Correct👍🏼
Legend
Dr. Neil these videos are amazing. Thanks for the great explanations!
Glad it was helpful.
Great video agan. Got a lot of these in the USMLEs
PERFECT. It WILL be out strength.
Dr Neil No words can describe about your teaching :)
Thank you Chetan... Hope you are staying focused and studying HARD!!
why for the last question when the equation was control - treatment, you put the new treatment data - the placebo ?
Thank you Dr. Neil!
You are most welcome.
hi dr. can you please do videos on...
a. linear regression vs logistic regression
b. kaplan meyer curve
c. forest plot
d. population pyramid
p.s thankyou for your videos:)
Thanks for the topics... I will definitely look to see if I have those in the library... You are putting me to work.
ua-cam.com/video/ulk_JWckJ78/v-deo.html
Thank you for being so thorough. Your work is amazing as expected 😉
Thanks Jolie, You are always very kind with your comments.
This video saves my life. Thank you so much sir.
Glad it was helpful.
Thank you so much, Dr. Neil!
These are super clear to learn. Thank you very much
Chinthaka, You are most welcome.
So grateful for you Prof!
My pleasure!
Always hated biostats, was never any good at it. Thank you SO much for these simplified videos! Made it super easy! Can't thank you enough! I'm still can't get what's a good confidence interval value and why.
For the most part...think...the Confidence interval that doesn't contain the number "1" is good... 0.55-0.86 is good.... 1.2-3.5 is good... 0.9-1.1 is NO good...
@@RandyNeilMD Oh okay, thanks alot!
Nice concept explanation thank you
You are welcome
Thank you Dr. Neil
You are most welcome Amrinder.
Wow , great explanations and made so easy to understand the concept in ridiculously short time , Dear Dr kindly upload a Neuro video of high yield topics for step 1.
Thanking you in anticipation 😊
Ahtesham, Thank you for commenting. I will eventually make my way to NEURO... I'll have to look back and see what I have. Appreciate the comment.
A better explanation of chi squared would be that it's comparing two data sets where both the independent and dependent variable are categorical (vs categorical independents and measured dependent variables)
wow ! .. actually it should be woowooowoowooowooowo ! u made it like super easy within 22 min video ! I wana be your life long student , Sir . u r really great , hero of biostat I must say . i think everyone will agree with me too.
Taz, As long as we are all moving towards getting through these exams (STEP 1, 2, 3) then we will make whatever video topic is needed.
Thorough as always! Thank you!
Bork Boi, Thank you for your continued support.
Hi Randy, great videos. In the second question I still don’t understand why smoking is a confounder bias rather than effect modulation.
Smoking affects exposure AND outcome (both of these variables) which is by definition CONFOUNDING BIAS. People who drink alcohol may also smoke (distorting exposure values) and people who smoke may also get pancreatic CA (distorting outcome values). This cumulatively distorts the relationship bridge between smoking and pancreatic CA and confounds the results.
Smoking is an effect modulator with DVT as it only affects the outcome which is by definition EFFECT MODULATION. Think about it...smokers are vasculopaths who are inherently in a hypercoagulable state and thrombose easily) but there is NO clear association between the NMDA drug and smoking.
is there a mistake ? For the question with smoking and prostate ca, if the p values were not statistically significant due to the p-vaules doesn't it mean the results are not valid in the stratification and that there is NO difference, no effect mod, and No confounding? thx
I was thinking the same that when one study is so statitically insignificant, how can we base our assumptions on that. Rather, it might be that the researcher had the predilection towards proving the association between the drug and DVTs, so he carried out an insignificant study to support his belief. In that case, the right answer would be 'confirmation bias'.
You were probably talking about Q#2, but yeah same point in Q#1.
Fantastic video! Thank you!
You are most welcome.
Awesome Dr Neil!
Thank you Obinna
Hi Dr. Neil. one more question :
A 26 year old medical student comes to the physician with a 3 week history of night sweats and myalgias. During this time, he also had a 3.6 kg weight loss. He returned from a month tropical medicine rotation in Cambodia 1 month ago. A chest x ray shows reticulonodular opacities suggestive of TB. The student is curious about his likelihood of having active TB. He reads a study that compares sputum testing results between 2800 patients with likely active TB on a basis of history, clinical symptoms and CXR pattern and 2,400 controls. The results are:
Active TB likely on basis of History, clinical and CXR Active TB not likely on basis of History, clinical and CXR
Sputum Test positive 700 300
Sputum Test Negative 2100 2100
What is the probability that a patient with a diagnosis of active TB on the basis of history, clinical symptoms, and CXR pattern actually had active TB? The actual answer is 0.25. But my calculation is 0.70. Would you please explain this? Thank you
Hi, the given table is the confusing aspect. They gave you the flipped table. The top heading should be the gold standard (Sputum test) and the side heading should be for the test conducted ( History, clinical and CXR) . You should rearrange and construct the chart like this:
Sputum Test positive Sputum Test Negative
Active TB likely on basis of History, clinical and CXR 700 2100
Active TB not likely on basis of History, clinical and CXR 300 2100
Then, PPV is 700/(700+2100) = .25
These videos should be on “Top videos” on UA-cam.
You are too kind...
The table in the NNT calculation is flipped. Exposure(new med) should be at the left side(rows)
Thank you Dr. Please can you make quick vedio for the p value
Sos un salvavidas, Dr Neil.
i dont understand the 2nd question on confounding and effect modulation. in the second half of the question statement where they corelated the effects of smoking with both alcohol consumption and pancreatic cancer the test results are of no significant value. both are more than acceptable 0.05. meaning that smoking is not related to either of them.
I have the same doubt
Every video is so amazing 🔥🎉
Thank you WW
You are a blessing.🙏👏
Thank you for the kind words.
thanksssssssssss a TON dr Neil,
Syed, You are MOST welcome...
Randy you have great hands
I work out...
thankyouuuu dr Neil!!!
You are welcome Zuha.
awesome simply awesome
Thank you Simran.
Hi Dr Neil, I have a question regarding Q2, in second part of the question the P value is high so wouldnt that means it is not significant and hence being a smoker or a non smoker doesnt really make a difference? am I missing out something here or am I not getting something and I have spoken something beyond stupidity? Would be nice if you shed a light on this. Thank you so much for your videos your work is wonderful!
I have been having the same doubt. Hope Dr Neil replies!
your videos have been so amazing thank you for your hard work! hoping you have time to upload a short one on population pyramid. happy holidays doctor Randy !
Michael. You are putting me to work...
Thank u sooooo much .. no words 😶 to say....this is very very good....
USMLE Goal, Thank you for the comment
Thanks for the video. God bless you immensely. Please can you do a video on hazard ratio
Genesis, Thank you for the comment... Yeah, I should have put a question about HAZARD RATIO... let me see what i can find.
I'm giving my Step 3 exam in 1 month!! I hope I pass!! I'll come back once i get my results.
STAY FOCUSED, STAY POSITIVE... YES, keep us posted.
YOU'RE THE BEST! Thank you!
Thanks Ryan, Appreciate the comment.
Thank you so much 😊
Menekse, You are most welcome.
Thank you so much, God bless you
You are most welcome.
Shouldn’t we put hair loss/ no hair loss on top of the chart and tx/no tx aside in NNT question? Thank you.
Thank you so much..
Matanat, You are most welcome.
Amazing! Thank you so much, specially the ANOVA, T test, and chi square!. I had a question about Q2 however; if they re-examnined the findings, and found that smoking/nonsmoking was only associated with a increase/decrease in smoking, wouldnt that be an effect modification as it only affects the outcome (of pancreatic cancer), and not the exposure (alcohol)? I am a little confused on that one
Saba, Thank you for the question... The take home point is that we know there is an association with BOTH exposure and outcome (for Confounding)...and only with OUTCOME for effect modulation. Everything else is "smoke and mirrors". Just take a step back and see it "big picture"... not sure if I am making sense...but for now, just know principle between the two.
@@RandyNeilMD thank you!
Hi Dr.Neil, can you do a video that involves the large 2 page pharm ads and how to navigate that? I get so intimidated. I'm trying to take the step III asap....my confidence needs the most work.
Thanks a alot for the video.
If data is significant, can a question ask to choose between p value of 0.05 vs 0.03?
Only if they would say which was better... less chance for error... you'd choose 0.03
THANKYOU for another amazing video :)
Sarah, Thank you for the comment. You are most welcome.
Hey Randy could you please explain relative risk reduction RRR=1-RR but I got lost with the words..can you just explain with an example?
it was really helpful,thank you!
You are most welcome.
So Randy if any time i see a confidence interval containing 0 decimals something i will cross that CI because of no significance?
Awesome!!
Thank you Suryasnata
Hey Randy! Which textbook do I use for bio stats ?
Thank you
Arimeto, Not sure I know of one...
You are the best ❤
Thank you
Thanks always sir!
Appreciate your comments.
hello Dr.Randy Neil, how can I get notes for your stats lectures?
jejeje checking this exactly 1 year later 8/13/2021
Good to see you back... As long as there is a trail of success!!
Very well explained
Shalvin, Thank you for the feedback... much apprecaited.
Thank you 🙏
Sangay, You are most welcome. Hope you are studying HARD!!
love every vid u made
Thank you for taking time to comment.
Thanks so much
Would you plz go over this question
The new diagnostic test was developed to aid in diagnostic GERD. Of 500 patient with GERD , the new diagnostic test identified 300 true positive. Of 500 patient without GERD , the new diagnostic test identified 215 true negative what is specificity ?
I had FP as 500-215 = 285
So my answer was 215/215+285
But the explanation answer had FP as 500-300=200
Can you plz plz explain?
Also would you have any lecture on different kind of test ? Cost effectiveness, minimization, utility , and benefit ?
Poneh, Based on your information, your work looks right and you know where the numbers go... It's possible there was an error in whomever made the question. It's possible...
Randy Neil, MD
Thanks so much
The explanation answer is referring to FN rate, but question asked for FP rate.
Hi Dr. Neil, thank you for the videos. Would you please explain this problem:
A recent study was conducted to assess the intelligence of students enrolled in an alternative high school program. The results showed the IQs of the students distributed according to the normal curve, with a mean of 115 and a SD of 10. Based on this information it is most reasonable to conclude that
a) 50% of the students will have the IQ below the standard mean of 100
b) 50% of the students will have the IQ below 105
c) Students with IQs of 125 are the 84th percentile
d) 2.5% of the students will have IQs greater than 125
The answer is b. Would you please explain where did 84th percentile come from? Thank you!
That is a rule of thumb:
Mean +- 2SD = 95% confidence interval
Mean +- 1SD = 66,6% CI
In a Normal Distribution, mean = median = p50
An IQ of 125 is the mean + 1SD. So it is p50 + (66,6/2) = 83.3% (roughly 84%). We divide 66.6 by 2 because we are only adding 1SD, not adding and subtracting, get it?
Remember:
- 2SD = p5
- 1SD = p17
Mean = p50
+ 1SD = p83
+ 2SD = p95
+ 3 SD = p99
You're amazing!
Thanks.
I feel like uworld explanations will mess me up, (after i get the Q right )after watching and learning everything i need to from Dr Randy.
Sameen, Thank you for the kind words.
I try to go back and like as many videos as I can on this channel. You are an angle. I got a 72 percent on my block and all Ive done are your videos!!
@Dr. Neil: Where have you been all my life?
For several years I watched UA-cam preparing for STEP 1... :)
Thanks this is helpful
Thanks
You are welcome.
These videos are awesome ❤️👏🏻 i just wanna ask u if we compare two mean values , when will it be regression analysis and when will be it two sample t test?? Because its so confusing