New feature idea:
Both vertical and horizontal (R2016b required) error bars can be added with an additional 'SDBar',true Name-Value pair argument. The error would be the within-subject standard deviation of the mean (horizontal) and the difference (vertical).
This is useful for visualisation of the data when repeated measurements are graphed. The amount of variation within a subject becomes apparent. This provides little information about the agreement between methods, but it does show what the methods mean-statistic standard deviations are within subjects. The bias and LOA and their CIs provide the overall agreement information.
Example image (ignore the different lengths in x and y directions):
TODO
[ ] Manage zero standard deviation correctly; how does the error bar look when the error is zero?
[ ] Add 'SDBar' (or 'SDBars') Name-Value pair argument: what do other MATLAB arguments use, singular or plural?
[ ] Detect R2016b and if not available: plot only vertical SD (of the within subject differences).
New feature idea: Both vertical and horizontal (R2016b required) error bars can be added with an additional 'SDBar',true Name-Value pair argument. The error would be the within-subject standard deviation of the mean (horizontal) and the difference (vertical).
This is useful for visualisation of the data when repeated measurements are graphed. The amount of variation within a subject becomes apparent. This provides little information about the agreement between methods, but it does show what the methods mean-statistic standard deviations are within subjects. The bias and LOA and their CIs provide the overall agreement information.
Example image (ignore the different lengths in x and y directions):
TODO