URL of experiment: https://gitlab.pavlovia.org/sheremat/asymprecision
Description of the problem: After running many participants I realized that the experiment was not controlling for the difference in screen dimensions between PCs, Linux machines, and Macs. The location of stimuli are along the circumference of an imaginary circle and there is a color wheel.
Is there a fix for this?
This isn’t anything to do with the operating systems per se, but just the natural variability across all computers when running online.
You need to describe what you actually want to control. There is not enough detail here to give you a useful answer.
Sorry for not being more clear.
In the experiment, Participants are clicking on a location along a color wheel to indicate the color of a previously presented stimulus. Each polar angle of the color wheel represents a color. Therefore, I need the color wheel to be presented as a circle, not an elongated oval.The screen width/height ratios on Mac screens is 1.6/1 but on the PCs and Linux machines that have been used are all 1.78/1. I am using normalized screen dimensions, so I thought it would account for the difference in screen resolutions. However, from the data it is clear that there is a difference between Mac users’ data and non-Mac users’ data. I realize that the operating system can be used on different machines, but these are undergraduates typically using a typical Mac with a Mac operating system or a PC with a windows operating system.
Also, running it on a typical pc with a windows operating system the color wheel is indeed an elongated oval.
You need to read the following page. norm units are not appropriate here. height will probably do what you want. But make sure you test carefully before running an experiment.
https://www.psychopy.org/general/units.html
Only on laptops (and even then, the 11 inch is 16:9). iMacs are 16:9.
That can certainly vary too. As above, this isn’t an OS issue. For example, Intel Macs can boot into Windows and Linux as well.
Computer displays come in all sorts of shapes and sizes. Hence why you need to think very carefully about the choice of units before running experiments online.
Thank you for the information, but using height does not resolve the problem. I had been using height for the units of the experiment, but norm for the color wheel. I changed the color wheel to height but the problem is different. The color wheel comes from an image and if we use height it is stretched along the x dimension. It seems like my options are to use normalized units and have a window that is not full screen (not ideal) or throw out participants whose data is collected on a computer with a different screen resolution (also not ideal).
As mentioned above, I do realize that computers can boot different operating systems, but that is not the case in our experiment. I have the resolution of the screen as well as the operating system and can confirm that the issue is due to the screen resolution, not the operating system, but in this case it is linked.
It can. You should show us exactly what you are doing (i.e. how you specify the size of your colour wheel image), and what you want to achieve. Plead provide screenshots of the current settings, and post the image you are using.
You have one option: present your stimuli in a way which is independent of screen aspect ratio. This is easily achievable by using height units. If you provide the details above, we can show how.