This entire video series on statistics has been fantastic! You guys do an amazing job at easily explaining concepts that undergraduate courses tend to over-complicate.
Maxine Lavigne's comment below is correct. The interpretation of the confidence interval in this video is incorrect. A 95% interval is a procedure which will result in an interval that contains the true parameter 95% of the time. However, once a sample is collected and an interval constructed, that interval either does or does not contain the true parameter Unless you are a Bayesian, you cannot make probability statements about the location of the true parameter.
Hi Laurence. Thanks for sharing your insights. In helping public health decision makers interpret 95% CIs, we have used a more pragmatic interpretation. According to the Cochrane handbook, “A 95% confidence interval is often interpreted as indicating a range within which we can be 95% certain that the true effect lies. This statement is a loose interpretation, but is useful as a rough guide. The strictly-correct interpretation of a confidence interval is based on the hypothetical notion of considering the results that would be obtained if the study were repeated many times. If a study were repeated infinitely often, and on each occasion a 95% confidence interval calculated, then 95% of these intervals would contain the true effect.” We have also consulted with a statistician, who has agreed that the video provides an acceptable interpretation of the 95% CI to support program planning decisions.
if I may ask, is it relatively possible to determine the effectiveness of an intervention of research when it does not mention OR or RR, but is based solely on the confidence interval and p values?
Thank you for reaching out. If a study reports a p-value, then you know if the effect of the intervention was statistically significant, but you have no sense of how large the effect was. The confidence interval (CI) is the crucial information required for decision making because it gives us the range within which we can expect the intervention effect to be. A decision should be based on considering ‘if the effect is at the lowest end of the CI, is it still worth the investment?’ In other words, when we consider the smallest expected effect for an intervention would we still choose to implement it? This is what the CI gives us. An RR or OR only tells us the magnitude of effect for the participants in that study. It does not tell us in the effect is statistically significant, or what the expected range of effect would be in the real world. On its own the RR or OR is not sufficient for decision making. In addition, we only need the confidence interval to determine if the intervention effect is statistically significant. So, if we had to choose only one of the OR or RR, the CI, or the p-value, we would want the CI.
Thanks for reaching out. In the first example, the confidence interval ranges from 30% above to 30% below the relative risk. In the second example, the confidence interval ranges from 50% above to 50% below the relative risk.
No words can express my gratitude 🌷
This entire video series on statistics has been fantastic! You guys do an amazing job at easily explaining concepts that undergraduate courses tend to over-complicate.
We’re happy you found it useful! If you have any questions, feel free to drop us a line or check out our glossary: www.nccmt.ca/glossary.
so true!
A 40% reduction , but still no effect! This is the problem of interpreting CI as a p value.
At last statistics simplified ! Thankyou
Maxine Lavigne's comment below is correct. The interpretation of the confidence interval in this video is incorrect. A 95% interval is a procedure which will result in an interval that contains the true parameter 95% of the time. However, once a sample is collected and an interval constructed, that interval either does or does not contain the true parameter Unless you are a Bayesian, you cannot make probability statements about the location of the true parameter.
Hi Laurence. Thanks for sharing your insights. In helping public health decision makers interpret 95% CIs, we have used a more pragmatic interpretation. According to the Cochrane handbook, “A 95% confidence interval is often interpreted as indicating a range within which we can be 95% certain that the true effect lies. This statement is a loose interpretation, but is useful as a rough guide. The strictly-correct interpretation of a confidence interval is based on the hypothetical notion of considering the results that would be obtained if the study were repeated many times. If a study were repeated infinitely often, and on each occasion a 95% confidence interval calculated, then 95% of these intervals would contain the true effect.” We have also consulted with a statistician, who has agreed that the video provides an acceptable interpretation of the 95% CI to support program planning decisions.
Great and georgious presenter.
thank you for these videos! I have been really struggling and ,my teachers are no help at explaining!
We’re glad you find these videos helpful!
if I may ask, is it relatively possible to determine the effectiveness of an intervention of research when it does not mention OR or RR, but is based solely on the confidence interval and p values?
Thank you for reaching out. If a study reports a p-value, then you know if the effect of the intervention was statistically significant, but you have no sense of how large the effect was. The confidence interval (CI) is the crucial information required for decision making because it gives us the range within which we can expect the intervention effect to be. A decision should be based on considering ‘if the effect is at the lowest end of the CI, is it still worth the investment?’ In other words, when we consider the smallest expected effect for an intervention would we still choose to implement it? This is what the CI gives us. An RR or OR only tells us the magnitude of effect for the participants in that study. It does not tell us in the effect is statistically significant, or what the expected range of effect would be in the real world. On its own the RR or OR is not sufficient for decision making. In addition, we only need the confidence interval to determine if the intervention effect is statistically significant. So, if we had to choose only one of the OR or RR, the CI, or the p-value, we would want the CI.
@@nccmt Wow, thank you so much for the response, it was very helpful!
so just 45% above and under the mean is the confidence interval?
Thanks for reaching out. In the first example, the confidence interval ranges from 30% above to 30% below the relative risk. In the second example, the confidence interval ranges from 50% above to 50% below the relative risk.
I like your other videos, but that interpretation is erroneous. You gave the interpretation of the credible interval and not a confidence interval.
Thank you for your comment, we are looking into the matter.