I would have a couple issues with the "Confidence intervals and hazard ratio" graph.
Putting confidence intervals side-by-side as they are invites interpretation of whether they overlap or not, which is usually not how you would compare the estimates. (You would look at a confidence interval of the difference of the two quantities, not the difference of the confidence interval.)
Also, the side-by-side dodging when the x-axis has a quantitative meaning is misleading.
Not sure if that's helpful. The graph caught my eye because I remember confidence intervals being one of the hardest concepts to get across during my short stint as a graduate teaching assistant.
I agree about side-by-side confidence intervals allowing for easy misinterpretation. It isn't officially wrong to compare two confidence intervals, but when you do, it is more conservative in that it is more difficult to reject when you should. I should have said more about this in the post.
I'm ok with the x-axis. The gestation age has been grouped, so a categorical interpretation here is appropriate.
I would have a couple issues with the "Confidence intervals and hazard ratio" graph.
Putting confidence intervals side-by-side as they are invites interpretation of whether they overlap or not, which is usually not how you would compare the estimates. (You would look at a confidence interval of the difference of the two quantities, not the difference of the confidence interval.)
Also, the side-by-side dodging when the x-axis has a quantitative meaning is misleading.
Not sure if that's helpful. The graph caught my eye because I remember confidence intervals being one of the hardest concepts to get across during my short stint as a graduate teaching assistant.
I agree about side-by-side confidence intervals allowing for easy misinterpretation. It isn't officially wrong to compare two confidence intervals, but when you do, it is more conservative in that it is more difficult to reject when you should. I should have said more about this in the post.
I'm ok with the x-axis. The gestation age has been grouped, so a categorical interpretation here is appropriate.