The main thing is the little black lines – the “confidence interval” – a statistical measure of uncertainty that can be used when showing the average value of data from a survey (or other type of research).
And what this means, which I have confirmed with the ABS, is that the reading rates are statistically the same for males and females within all generations with the exception of gen X.
Is this correct? I haven’t studied statistics since high school so I am completely clueless, but it doesn’t make sense based on my rudimentary understanding of what a confidence interval is supposed to do. The confidence intervals overlap, but they are not identical. Doesn’t that mean that reading rates could be statistically the same, but not that they are statistically the same?
Anyway, I also found it interesting that men read more magazines than women now too, considering it was historically the other way around and that many men actually believed its existence as a societal norm was an example of their superior rational minds.
Yes, it means could be the same, not are the same. It does mean they are confident (95% confident, I assume, I’m not clicking through to the study) that the rates are different for men and women in Gen X
It does mean they are confident that the rates are different for men and women in Gen X
Umm, surely not? If the confidence intervals overlap it means that they are not confident that the rates are different, doesn’t it? Of course, it also does not mean that they can say they are confident that the reading rates are the same.
So the statistically sound way of saying it is that the null hypothesis is that reading rates are the same, and their study has failed to reject the null hypothesis.
If you want to be precise, overlapping intervals mean that we lack evidence to assert that the means are statistically different for our chosen confidence level. This is often simplified to the statement that they are statistically the same.
Is this correct? I haven’t studied statistics since high school so I am completely clueless, but it doesn’t make sense based on my rudimentary understanding of what a confidence interval is supposed to do. The confidence intervals overlap, but they are not identical. Doesn’t that mean that reading rates could be statistically the same, but not that they are statistically the same?
Anyway, I also found it interesting that men read more magazines than women now too, considering it was historically the other way around and that many men actually believed its existence as a societal norm was an example of their superior rational minds.
Yes, it means could be the same, not are the same. It does mean they are confident (95% confident, I assume, I’m not clicking through to the study) that the rates are different for men and women in Gen X
Umm, surely not? If the confidence intervals overlap it means that they are not confident that the rates are different, doesn’t it? Of course, it also does not mean that they can say they are confident that the reading rates are the same.
So the statistically sound way of saying it is that the null hypothesis is that reading rates are the same, and their study has failed to reject the null hypothesis.
Gen X is the only category for which the CIs don’t overlap at all
Oh right I see, sorry.
If you want to be precise, overlapping intervals mean that we lack evidence to assert that the means are statistically different for our chosen confidence level. This is often simplified to the statement that they are statistically the same.