Am I able to use the Bland-Altman plot for non-parametric data? If so would I need to use a Wilcoxon signed rank test instead of a one sample T-test to assess difference prior?
Hi, thanks for the nice video. We have this problem: what happens if at the very beginning the mean difference is signifacantly different from 0? Is there a way to go further?
Very helpful txs. Upper and Lower LOA clear now how to calculate en show in SPSS. 1 question: How can we make the lines thicker in the graphs if we look at the three lines on the Y ax??
Thank-you so much this was incredibly helpful!! I followed it the whole way through with my data and at the end I found there was proportional bias. Do you have a reference of something I can follow with what to do now? Thanks again, this has really helped me!
why do you use T-test instead of normality test of Kolmogorov-Smirnov or Saphiro-Wilk to see whether there's significant difference between first measurement method and second measurement method?
I have a question: Can you explain why you did the one-sample t-test before you did the Bland-Altman plot? So you were testing to see if the values differed from 0. If the significant level (2-tailed) is lower than 0.025 (2.5%), then you would've done the plot. But if it was higher, then wouldn't that mean the values differed enough that you didn't need to do the plot? Your value was 0.199.
@@tacappaert can you provide a reference to this that could be used for publication? i.e., there is no need of plotting B&A plot for significant variables
Thank You for the brillliant video and information. Can I use Bland and Altman plot in an non normal sample? If the sample is skewed is possible to use it? Thank You
Thanks for a great video, I've found it really helpful. Just a quick question, why do we not use the 95% upper and lower confidence limits that were generated by the T test?
It is the year of 2022. I'm writing my thesis at the moment and have the same question... This is a great and helpfull video, but this is the part that confuses me as well.
Also, when you do the linear regression in the end, what would be a statistically significant t-score? For some reason, I keep getting 0.000 for my Sig. coefficients and -11.203 and 11.200 for my t-scores.
This is great, thanks. Can I ask when working out confidence limits my upper confidence limit is outside the parameters of the graph. I.E my difference on the Y axis is =2 but my confidence limit 2.1. Does that mean I must have calculated wrong or do I just need to adjust the graph?
When you are in the Chart Editor window, double-click on the name of the Y-axis and a new window will open. Then you should have a tab that allows you to customize the scale to better represent your data.
Thank you for this very informative video. I have read some scientific papers that used this method to assess the level of agreement between two methods (or two observers). In these manuscripts there is also a "concordance index" with 95% CI and my question is: can we measure it with SPSS? Thank you again.
Very helpful. However, I cannot understand the last phrases, because you speak too fast. My question is, I found proportional bias when I did this on my data, what do I do next to report it to my results. Could you please make a video about that or reply to my comment. I am looking forward to hearing from you soon. Best regards, Faraidoon
This is brillliant. Thank you. I am doing some relaibility analysis and I initially intended to use ICC but then I read somewhere that ICC only works for parametric data. Is this true? I ask because my data is skewed. Many thanks again.
It is more accurate when the data is normally distributed as is true for most correlations. I would suggest transforming your data and then running the ICC.
![Bland-Altman Plot [Simply explained]](http://i.ytimg.com/vi/hHsfn7jE79M/mqdefault.jpg)
Thank you for taking us through this invaluable method.
Excellent ! many thanks for guiding !
You are welcome!
Thank you very much, I greatly appreciate this video, it is greatly useful, good luck with best regard, Iraqi physician
thank you. very useful video!
Do you have any video for the next step after assuming there's proportional bias?
Am I able to use the Bland-Altman plot for non-parametric data? If so would I need to use a Wilcoxon signed rank test instead of a one sample T-test to assess difference prior?
Hi, thanks for the nice video. We have this problem: what happens if at the very beginning the mean difference is signifacantly different from 0? Is there a way to go further?
Thank you. it is very helpful.
You are welcome!
Very valuable information, thanks
do you have a lecture of Bland Altmant Analysis for non-parametric data?
Very helpful txs. Upper and Lower LOA clear now how to calculate en show in SPSS. 1 question: How can we make the lines thicker in the graphs if we look at the three lines on the Y ax??
Great video! Thank you so much!
Can we use this plot for agreement between 2 repeated measures with same instrument technique for example Elisa
Thanks for this vid found it really helpful.
Why do you multiply the standard deviation with 1.96 in order to get the confidence limits?
Tessoo 1.96 is obtained from the z-table for a significance level alpha=0.05
Very helpful, thx! Do you have any videos on calculating and inserting confidence intervals for the limits of agreements?
+Tobias Lyngeraa I'm sorry, I do not.
Thank-you so much this was incredibly helpful!! I followed it the whole way through with my data and at the end I found there was proportional bias. Do you have a reference of something I can follow with what to do now?
Thanks again, this has really helped me!
If a paired t-test show significant differences between the two methods of measurement, is a Bland-Altman plot still necessary? Thank you.
why do you use T-test instead of normality test of Kolmogorov-Smirnov or Saphiro-Wilk to see whether there's significant difference between first measurement method and second measurement method?
Thank you so much! Bless you
You are so welcome
Hi, with a sig. (2 tailed) = 0,913, you say that the difference is not significally? what is the refference level? thanks
I have a question: Can you explain why you did the one-sample t-test before you did the Bland-Altman plot? So you were testing to see if the values differed from 0. If the significant level (2-tailed) is lower than 0.025 (2.5%), then you would've done the plot. But if it was higher, then wouldn't that mean the values differed enough that you didn't need to do the plot? Your value was 0.199.
+LiNingAir That's correct. If they are significantly different from one another, there isn't agreement between them.
@@tacappaert can you provide a reference to this that could be used for publication? i.e., there is no need of plotting B&A plot for significant variables
@@nidahr8719 Did you find anything in that regard? I'm looking for such a reference myself.
Thank You for the brillliant video and information. Can I use Bland and Altman plot in an non normal sample? If the sample is skewed is possible to use it? Thank You
+Maria Sofia R. Alves Pimenta It is possible but the result should be interpreted carefully.
When calculating the UCL and LCL , what is 1.96 that you multiply by std. dev (minute 8 - 9 in video) ? Thanks
boston_boy i also want to know what the 1,96 is .
fatima there are some really good videos out there which explain the entire process.. I'll try to find out the links
If you want to
let me know if this helps
boston_boy i cant find the link.
Thanks for a great video, I've found it really helpful. Just a quick question, why do we not use the 95% upper and lower confidence limits that were generated by the T test?
It is the year of 2022. I'm writing my thesis at the moment and have the same question... This is a great and helpfull video, but this is the part that confuses me as well.
Also, when you do the linear regression in the end, what would be a statistically significant t-score? For some reason, I keep getting 0.000 for my Sig. coefficients and -11.203 and 11.200 for my t-scores.
+LiNingAir You would use the significance value. If this value is less than 0.05, then the regression model is statistically significant.
This is great, thanks. Can I ask when working out confidence limits my upper confidence limit is outside the parameters of the graph. I.E my difference on the Y axis is =2 but my confidence limit 2.1. Does that mean I must have calculated wrong or do I just need to adjust the graph?
When you are in the Chart Editor window, double-click on the name of the Y-axis and a new window will open. Then you should have a tab that allows you to customize the scale to better represent your data.
Thank you for the video!!!
Thank you for this very informative video.
I have read some scientific papers that used this method to assess the level of agreement between two methods (or two observers). In these manuscripts there is also a "concordance index" with 95% CI and my question is: can we measure it with SPSS?
Thank you again.
You could also use Linn's Concordance Coefficient (How to Use SPSS-Lin's Concordance Correlation Coefficient).
Very helpful. However, I cannot understand the last phrases, because you speak too fast. My question is, I found proportional bias when I did this on my data, what do I do next to report it to my results. Could you please make a video about that or reply to my comment.
I am looking forward to hearing from you soon.
Best regards,
Faraidoon
This is brillliant. Thank you. I am doing some relaibility analysis and I initially intended to use ICC but then I read somewhere that ICC only works for parametric data. Is this true? I ask because my data is skewed. Many thanks again.
It is more accurate when the data is normally distributed as is true for most correlations. I would suggest transforming your data and then running the ICC.
you rock dude! Kudos
Thanks
Welcome
thank you