Kappa Value Calculation | Reliability

Enroll in our online course: bit.ly/PTMSK DOWNLOAD OUR APP:
📱 iPhone/iPad: goo.gl/eUuF7w
🤖 Android: goo.gl/3NKzJX GET OUR ASSESSMENT BOOK ▶︎▶︎ bit.ly/GETPT ◀︎◀︎
This is not medical advice. The content is intended as educational content for health care professionals and students. If you are a patient, seek care of a health care professional. In the ideal scenario, when two or more investigators conduct the same test, they will agree 100%. Though that's rarely ever the case. And if so, what is the chance that they simply got lucky in their agreement. That's where a Kappa statistic comes into play. The Kappa value rates how good the agreement is whilst eliminating the chance of luck. Kai shows how to calculate it.
Article: www.ncbi.nlm.nih.gov/pubmed/1... (Viera et al. 2005)
Support us on Patreon: / physiotutors
Visit our Website: bit.ly/web_PT
Like us on Facebook: bit.ly/like_PT
Follow on Instagram: bit.ly/IG_PT
Follow on Twitter: bit.ly/Tweet_PT
Snapchat: bit.ly/Snap_PT

[@keyoordeshpande533]

Great explanation and great biceps! Motivated me for learning and going to gym :D

[@Physiotutors]

Great comment! And great initiative 💪🏼

[@AbdullahGriffin]

Man, you are a master! I had a lot of trouble understanding that concept and you explained it very well in 3 minutes! Thanks, sensei.

[@jonasakakpo5757]

Great, very clear, and precise explanation.Thank you!.

[@willemblois]

Great, clear explanation! I'm in med, not physio, but these videos are still an incredible help. Thanks!

[@Physiotutors]

+Willem Blois glad you liked it! We've got quite a lot of med students especially when prepping for the OSCE !

[@roderickspender8298]

Incredibly helpful! Thank you

[@amywilderson85]

I have read so many things to understand kappa. But I could not before this video. Thank you very much!

[@uthpalagunasekara1493]

Clearly explained. Keep up the good work

[@gabrielsanzvillegas1707]

Very concise. Thank you!

[@Physiotutors]

Thanks for the compliment and for following Gabriel! You're welcome!

[@hughconboy7330]

I'm doing psych but still found this helpful. :)

[@baselalyafi4022]

Simple and straight to the point. Thanks👏👏

[@Physiotutors]

Thanks Basel!

[@abeeraljohar67]

Thank you so much for a great explanation and take me on the test

[@Physiotutors]

+abeer aljohar you're welcome!

[@Rhyno012345]

This helped a lot. Thank you!

[@Physiotutors]

Happy to help Ryan!

[@sibaprasad5461]

this is very good video. please explain more details with more examples.

[@oyeyemiooyelade8989]

Its so clear and explicit. Thanks

[@Physiotutors]

Happy you like it Oyeyemi!

[@sattamdatta8921]

Thank you soooooo much. You save my life

[@drewlopez9725]

Wished you explained more where you got the values in ReD. Overall great video and simple to follow.

[@Physiotutors]

Thanks, it's just the cases in which the raters do not agree.

[@krupanidhi912]

Tx for the nice explanation

[@banafshagul2945]

nicely explained

[@mrd8300]

Great video, could you help when comparing more than 2 variables?

[@chandrakanthmalyala1942]

Very good explanation

[@oz4876]

Thank you, I was reading a 65 page academic paper on kappa value and still had no clue what it was talking about - you clarified it perfectly in a little over 3 minutes

[@Physiotutors]

That’s the goal 😀

[@lomoyang3034]

Thanks for your explanation!

[@Physiotutors]

Ur welcome Lomo!

[@derrickwilson7858]

This was great and now I understand perfectly.

[@Physiotutors]

Awesome, that's the goal of the video!

[@arunvaibhav5059]

Neatly Explained! Thanks a lot!

[@Physiotutors]

We are glad you could profit from the video!

[@hibye8619]

Thanks! I’m doing nursing and I found this so helpful!

[@Physiotutors]

Glad it was helpful!

[@alighorbani9437]

Thanks a lot mate

[@SoniaB5]

You guys are great at explaining things!

[@Physiotutors]

+Sonia Borges Happy to hear! The kappa value is not an easy topic so I hope it became clearer to you!

[@SoniaB5]

It did...thank you!

[@lopamudragoswami7891]

Hi, thank you for the quick tutorial. Here's my question. In case of large data sets when the primary investigator has completed all the ratings (for example) and now to ensure inter-rater reliability gets another investigator (external) involved to rate, does this new investigator need to go through all of the data set or is there a magic number (minimum percentage of data set needed to be evaluated by both) in terms of establishing this reliability? I hope my question is clear. Thank you!

[@rajeshmahida2905]

Awesome explaintion

[@Physiotutors]

+Rajesh Mahida thanks 🙏🏼

[@wafaajamal1658]

Thank you 🙏🏼

[@dulce654]

Thanks!!

[@shereenjadallh8867]

It was very clear thanks

[@Physiotutors]

Good to hear Shereen! Thx!

[@huzeyfebilge2863]

Hi! Thanks for the video. I have a question about Kappa. Let's assume that two raters evaluate more than one thing. For instance, two teachers evaluate whether ten many learning outcomes are completely about reading skill or partially about, or not about (three categories for ten learning outcomes, and two raters). In this case, how can I use Kappa? How can I calculate it? Thanks for your help.

[@vuyelwavutela6427]

this was great, thank you

[@Physiotutors]

You're very welcome!

[@sak8485]

the video was short, sweet and spot on. :)

[@Physiotutors]

Thanks 🙏🏼

[@msnageshrao]

sir, very nicely explained , very easy to understand . thnaks a lot- nagesh

[@Physiotutors]

Thanks a lot for the compliment and for following us nagesh !

[@naomibarranzuela2320]

Nice video...but do you know how to calculate kappa for each category? would be awsome if you could explain that.... thank you!

[@Virchov]

I just got one question: Can I use this test to compare two diagnostic tests? I am comparing a rapid test with cultures and other methods. Is Kappa test useful in this case to determine if the rapid tests are useful?

[@lubelihlegwebu6155]

clearly explained

[@Physiotutors]

+LUBELIHLE GWEBU thanks, glad you could comprehend it!

[@DnyaneshwarPanchaldsp]

if their 3 row and 3 column e.g. positive, negative and neutral then how to calculate

[@christopherjamescj1331]

Someone give this guy a medal❤❤❤❤❤

[@dramina2326]

Fantastic!

[@Physiotutors]

Thanks Amina!

[@Shekarg76]

thank you sir

[@Physiotutors]

You’re welcome 😊

[@mmd6866]

what happens when pe = 1? In the case when they agree everywhere same.

[@Nakitende256]

Thanks

[@Physiotutors]

Ur welcome!

[@turinmormegil7715]

Whenever we don't have the table with the randomic distribuition, eve nthough we can use the % to reach the chance agreement, I wonder why we are told, in order to fill the cells with the random distribuition and obtain the random values, to multiply marginal total percentages of each cell (ex: cell A - multiply % of the total + from physio 1 and from physio 2 to find the % that A represents from the total). Why does it work? Why should I deduce that the percentage the total of + from P1 in relation to the total, multiplied by the percentage the total of + from P2 (that % also in relation to the total) will give the % o A in relation to the sum of all four cells?

[@AwaisSpeakOfficial]

Your lecture was awsome. But what we should do when we have multimonial data ?

[@Physiotutors]

This is how deep we are into this topic with our Master's degrees. Think other pages can help to get more into depth.

[@lenahudson6397]

Great video, can you please tell diff. between Kappa and Chi-sq

[@Physiotutors]

To our knowledge Chi-sq is used to test for a relationship between two variables, so if they are dependent or not. That is something completely different from the Kappa value, which displays the reliability between repeated measures of one examiner - or more often - the reliability of two examiners. Reliability is the degree to which study objects can be distinguished from each other (in case of Kappa "without chance")

[@lenahudson6397]

Thank-you....... :)

[@allademuth9366]

great!

[@Physiotutors]

Thanks a lot, happy you like our video!

[@Yasirkhan-fg7zd]

Hey! U r doing a great job can you please tell me how to find kappa statistics for multiple classes.

[@Physiotutors]

Thanks Buddy, this is a bit where our knowledge ends as well. Might look into it in the future

[@Yasirkhan-fg7zd]

@@Physiotutors never mind, but I did it by my own..... And your video is quite helpful to me.

[@digitalworld_7883]

How to know the threshold value from maximum kappa and how does it used to compare the different model prediction

[@Physiotutors]

Not exactly sure what you mean. Maximum kappa is 1. Kappa is not about prediction, but about reliability.

[@wickedsunny8462]

Very well explained Please make videos on six sigma case study in detail

[@Physiotutors]

Don't have any knowledge on six sigma. Can only remember that it was about quality control back in the days when i studies economics.

[@May-vm3uc]

You didn't explain how the 10 and 15 written in red were gotten please

[@getahunnegash6513]

greatly appreciated go head

[@ManojM-hy2rk]

In Pe value calculation 100 is Total sample or it is used to calculate the percentage ?

[@Physiotutors]

total sample

[@mrxzxcxz6906]

Nice explanation! Can you give just a simple example rather than hypo mobility?

[@mrxzxcxz6906]

Just a simple example for a student

[@Physiotutors]

This is as simple as it gets with the kappa value

[@ariftanurjanah]

It clear. But i'm still confused about if we have data of the population, whether we also must calculate the kappa value?

[@Physiotutors]

Not sure if we understand your question. The Kappa value is calculated if the reliability of a measurement or a test is to be assessed by one rater (intra-rater reliability) or between 2 or more raters (inter-rater reliability).

[@rifkyakhmadf7872]

is there just only 2 rater/observer in Kappa cohen coeficient?

[@Physiotutors]

With more than 2 raters it's called the Fleiss' Kappa as far as we know

[@girmatekle5927]

it's short and precise

[@Physiotutors]

Thanks 🙏🏼

[@RTReview.]

Well very well explained! Thank you for helping me out. The Kappa Value is 0,375. I was wrong :-)

[@Physiotutors]

Where do you see a mistake in the calculation or how do you end up with 0.3939?

[@Physiotutors]

Just calculated it again. Don't see where we are rounding. But happy you like our explanation!

[@RTReview.]

Physiotutors iam really sorry. That was my bad. I round the values wrong. Thank you for correcting me! Really wanna thank you for the explanation!

[@sarathj]

This is good; In the other video, you use the term "reliability' in a loose incorrect meaning. The better word I think is precision. Reliability is a probability in time domain.

[@Physiotutors]

Thanks. Not sure what you mean as the Kappa value is about reliability.

[@GauravGiri.]

Can you please explain the table at 0:50 in a bit more detail?

[@Physiotutors]

What exactly is unclear? Then we can answer more specifically.

[@GauravGiri.]

@@Physiotutors values written in Red..

[@Physiotutors]

The values in red are the cases when there was no agreement between rater 1 and 2.

[@ManojKumar-gk9sj]

Nice 👍👍

[@Physiotutors]

Thank you! Cheers!

[@jellybean5626]

few days to go and it'll be board exam day, hoping this video would save me. T.T

[@Physiotutors]

Good luck 🍀👍🏼

[@LetisiaChelly]

Why those numbers & why are they calculated that way?

[@Physiotutors]

not sure I understand what you're asking. Could you rephrase your question?

[@MrJamieb147]

Very stupid question, but what does 1 - 0.6 actually mean? So 0.75 - 0.6 = 0.15, do I then calculate 1 - 06, which equals 0.4? So the final calculation is 0.15 - 0.4? Where am I going wrong? 😂

[@anitarana5385]

Wow ❤

[@mohammadal-dabbas9481]

4Head

[@albertbarsaga8042]

What kappa

[@Physiotutors]

A measure of inter-rater reliability that corrects for chance

[@putrabhin1099]

Omg his hot 😍

[@adammadsen873]

Amazingly good video. Is there also a formular for calculating the intra-tester reliability?

[@Physiotutors]

Happy you like it Adam. The Kappa value can also be calculated for intra-tester reliability. It's exactly the same formula except that in the 2x2 table you will have your "1st measurement" and "2nd measurement" instread of "rater 1" and "rater 2".