- Видео 19
- Просмотров 416 631
watsonpho1
Добавлен 26 окт 2011
Видео
Crosstabulation for Epidemiology
Просмотров 9452 года назад
Completing a crosstabulation for Epidemiology can usually be done with careful data collection in an excel spreadsheet. However, completing the crosstabulation by hand and through deduction is helpful to understand how different risk group's attack rates help you discern the cause of an outbreak.
Creating a Program Evaluation (Public Health) Sampling Plan
Просмотров 2,1 тыс.3 года назад
This video is the recording of a class given over zoom about creating a sampling plan for public health program evaluation.
Setting up a survey codebook in Excel -- Tutorial
Просмотров 31 тыс.3 года назад
If you are trying to create a survey codebook for the first time, this is the quick tutorial for you! This tutorial shows you how to setup a codebook in Microsoft Excel, code questions with different data types including "mark all that apply" type questions and even how to created a computed variable from your survey responses. See more videos on SPSS or Epidemiology on my youtube home page. ru...
Oakbrook COVID
Просмотров 2894 года назад
Hi Brothers and Sisters. Here is just a little information about COVID-19 as of 11 march 2020 and what you and your family can do to stay healthy.
Epidemiology: Calculating Incidence and Prevalence
Просмотров 28 тыс.6 лет назад
This video is an online tutorial for epidemiology students interested in seeing how to calculate incidence and prevalence
Setting up a support group
Просмотров 9 тыс.10 лет назад
A complimentary lecture based on the University of Kansas Community Toolbox.
SPSS set up tutorial
Просмотров 6 тыс.10 лет назад
This quick tutorial explains how to set up a SPSS file in data view and variable view.
Coding a survey for SPSS or statistics
Просмотров 71 тыс.10 лет назад
This video explains how to create a coding sheet for a survey. Data types and how to code the results of a survey are presented. This information is helpful for research methodology, statistics, or social science courses.
Calculating Positive Predictive Value and Negative Predictive Value
Просмотров 5 тыс.10 лет назад
A summary of the predictive value of tests based on specificity and the prevalence of disease on the tested population
Sensitivity and Specificity explained
Просмотров 3,6 тыс.10 лет назад
Dr Tyler Watson, a public health professor, explains how the sensitivity and specificity of a test are related.
Calculating Sensitivity and Specificity
Просмотров 6 тыс.10 лет назад
Calculating Sensitivity and Specificity
Calculating Matched Pairs Odds Ratio
Просмотров 32 тыс.10 лет назад
Calculating Matched Pairs Odds Ratio
Epidemiology: Calculating Attributable Risk
Просмотров 107 тыс.10 лет назад
Epidemiology: Calculating Attributable Risk
Epidemiology: Calculating Relative Risk
Просмотров 76 тыс.10 лет назад
Epidemiology: Calculating Relative Risk
Epidemiology: Calculating Death Rates
Просмотров 39 тыс.10 лет назад
Epidemiology: Calculating Death Rates
❤
god bless you man, u just saved so many brain cells in my head
I am glad it helped
Thank you alot
THis was great thanks so much
Is the pairing up of case and controls random? That's the only thing I dont understand
Awesome. Great explanation
How did you get 39 for true positive
Starting at 2:26 I am talking about the Negative results. Because 182 negative tests occurred, and because 161 negative tests were ACTUALLY negative, it left 21 False negatives. 60 had UTI - 21 false negatives, that leaves 39 True Positives
That is not attributable risk its attributable fraction
Thank you! This was very thorough. I had no idea what a code book was when it was mentioned.
Excellent Job! Thanks
Explanation is not clear enough and the equation used is wrong
Good video now I know the difference in group vs therapy support.
Very helpful video, thanks!
Thanks for the tutorial! This was so helpful!
improve your handwriting
I do not agree with the ordinal coding of nominal variables. A matrix coding for these would've been much better!
Thank you for this helpful video!
Does this mean causal relationship without control study??
Amazing! I wish text books make things as simple as this video
spectacular video watsonpho1. I broke that thumbs up on your video. Keep on up the terrific work.
@5:30 when calculating the prevalence rate, if a few subjects died before July, then you wouldn't count it in the numerator? And then subtract that death from the total population in the denominator?
Really didn't answer how it is coded into the spreadsheet.
you are correct..
simple to complicated.
could you please do a video for the last question on the survey? how to enter it in spss?
I just did one that talks about that kind of question. It is using excel instead of spss, but the principles still apply. Good luck.ruclips.net/video/y7C3OAnIVPE/видео.html
What if it's a case where you would *have* to be exposed to get a disease
Very clear explanation here, thank you. BUT how what would be the MARGINAL Odds Ratio in this case?
In this case, the odds ratio calculated IS also the marginal OR. That is because we have the entire data set. There is no 3rd or 4th variable that you would have to "colapse" into the margins (or totals for the row and column). If we would not have matched the cases and controls as we did in this example and instead used that matching characteristic as a variable, then we could calculate a conditional and marginal OR to see if these two variables were truly related, or if they were independent and only linked through that third variable.
@@watsonpho1 so if we were to consider a "stack" of, say K, 2x2 tables then any one k'th table would have a Conditional OR however all K tables combined would have a Marginal OR?
this has been really helpful and very clear
I was hoping the last question on the survey would be included in the video because I don't know how to code a similar question that I have 😔
What is the question... maybe I can help
@@watsonpho1 Omg thank you! So my survey question was "If you have received any traffic fines, please check the reason(s) why from the list below. You may check more than one choice and add more. If you have not received any fines, leave the question." Checklist choices were: Speed, No seatbelt, Distracted by cell phone, and Other. Many responders to the survey had multiple things checked. Now I'm trying to code them for SPSS. My question is: should I split the column I had labeled as "Fine reason" to columns representing each choice and code them some thing like "Checked=1", "Not checked=0". I have already started the process of splitting that column but I feel like there has to be a better way since I'm not sure how it's gonna pan out on SPSS. Sorry if that was too much unnecessary information.
I should clarify: I am coding the data on Excel to eventually do data analysis on SPSS.
@@michelleabdullah5469 that is exactly how you code it... in spss or excel. Just be sure to keep track of how you code it in excel because when you move to spss, all that will move is the numbers. Once it is in spss, each column from excel becomes it's own variable which is how you have to analyze this. Another trick of analysis btw is that once it is coded the way you have it, you can mathematically enter a new calculated variable that will now be ratio instead of nominal data. You just ask the computer to add another column and sum all of the choices for each response. Then the new variable will be the total number of reasons they have fines... with that you can do average and other math including t tests. So... great job... you coded this question exactly as I would have
watsonpho1 Thank you sooo much! That was helpful and reassuring. Really appreciate it.
So informative, thank you
This video is not attributable risk. This isn’t attributable risk percentage either. AR% is RR-1/RR x 100. RR would be a/a+b divided by c/c+d and not what you said in the video.
It's confusing that's why I went to comments to find out if maybe I'm not getting the concept. The explanation is long and I don't think 64% is attributable risk
I have a query. If we calculate only the numerator part as percentage. We get 12 point something. How do you interpret that ? Like 64.81 is the percent of exposed for which it can be attributed to the exposure? But what would be the significance of 12?
Thanks for the question Abhishek. The 12% would be the difference in risk between the two groups.
Occasionally, I will get comments that this video is not attributable risk... That it is Attributable risk percent.🤔 The only difference here is I add the denominator of (risk in exposed) to give the proportion... The numerator is exactly the equation you will see in some of the comments below. I do this for 1 main reason; I teach undergraduate students the principles of epidemiology... and so my focus is on making these calculations mean something rather than just formulas to memorize. Let me give you an example. On June 2, 2020, there were 1,440,414 cases of COVID-19 in all the U.S. minus New York (the unexposed group). New York experienced 371,711 cases. The U.S. population estimate is 328,239,523 and of those 19,453,561 live in New York state. If we calculate the straight Attributable risk (what I tend to call difference in risk) we would get --- (371,711/19,453,561)-- (1,440,414/308,785,962). Leaving us with an attributable risk of 1.44%. For most people that seems really insignificant because the rate of disease as a general rule is very low in both groups. However, when you continue the calculation to the %... you would get (.0191-.00466)/.0191 = 75.60% Which tells us, that 75% of the illness in those who were sick in New York resulted from living in New York. THAT is easy to understand... New York is/was a hotbed... and people who lived there were at much greater risk. I hope this helps any confusion.
Thanks G, I was looking for Attributable risk percent anyways.
Thank you very much , this is very helpful
Hi I have a question regarding one calculation. Could you help me considering that I want to learn the fundamentals?
Take this video down. It’s wrong. AR= incidence in exposed - incidence in unexposed. Not your equation
See Shiva, if you just take the numerator of this formula, you will have your simple calculation. Some people call the calculation I perform here attributable risk percent or proportion. I think it is obvious that this is the proportion. In fact, if you read in the description, you will see that. I am not sure why you would feel the need to tell me to take it down, as there is nothing incorrect about the measurement of risk here... In fact, when you add the proportion part of the equation to attributable risk, it is MUCH easier to understand how interventions might impact the disease. So please, if this measurement tutorial is so offensive... visit another site or make your own.
I don’t mean disrespect. But your video needs to be clear that AR= risk in exposed - risk in unexposed. It just needs tone more clear. Putting it in the description may not be the best thing.its better to explicitly state that. I have read comments and they criticize this very same point. It’s only when reviewing another video then checking for my self that I noticed this.
Hi, thanks for this - so technically one could do this for say - google forms?
Thanks a lot for the wonderful explanation!! Please make a video on different statistical methods in epidemiology..
Pls connect with me 🙏
Awesome. Gracias.
Thanks for this great info boss!
Thanks Tyler. Great info!
Bad audio
Thank you for this !
Aren’t you supposed to subtract 110 from denominator because of existing cases for incidence?
No... If you start to make adjustments to the denominator (the population) for a single variable that influences population, you would also have to adjust for all deaths in the community, all births, all move ins, and move outs (called transiency). Instead, most situations you will just use the mid-year population as a static denominator.
@ shayan you're right
Thanks sir.
Bayes' Theorem and Breast Cancer Screening with mammography. Please update and re-upload this excellent informative video using current scientific data. Sensitivity and Specificity of screening digital mammography in a recent extensive Norwegian clinical trial were 54.1% and 94.2% respectively. So:0.01x0.541 0.01x0.541 + 0.99x0.058 or 0.00541 0.00541 + 0.05742 or 0.00541 0.06283 or 0.0861 or 8.61%! If a woman receives a positive mammography result for breast cancer the probability she actually suffers from the disease is only 8.61% Reference pubs.rsna.org/doi/abs/10.1148/radiol.2019182394
Or you could just simply use (A/(A+B)) - (C/(C+D))
Explanation not clear enough
In the first table, the black font colour in the dark blue background is not easy to see!
Haha that intro totally applies to me. OOPS
thank you very much for the video, you mentioned there was a video for odds ratio but `I cannot find it in the channel, can you send me the link please?
This is for Attributable Risk Percent