All about tim­ing accuracy

Alex Irvine, MRC Cog­ni­tion and Brain Sci­ences Unit, Cambridge
@alexanderirvine

By load­ing the video, you agree to YouTube’s pri­va­cy pol­i­cy.
Learn more

Load video

Always unblock YouTube

My talk at BeOnline 2020 will be on the accu­ra­cy and con­sis­ten­cy of visu­al pre­sen­ta­tion and record­ing of key­board respons­es online. So if you are run­ning research that involves pre­sent­ing images for short dura­tions, or looks at small dif­fer­ences (less than a few hun­dred mil­lisec­onds) in reac­tion times, then this talk will be of inter­est to you. The preprint paper I am pre­sent­ing looks at dif­fer­ent plat­forms — Goril­la, jsPsych, PsychoPy/Pavlovia and lab.js on 4 dif­fer­ent web browsers and 4 dif­fer­ent device cat­e­gories. It is split up into two parts. In the first, we used a pho­to­di­ode to detect onset of stim­uli on the screen, and a robot­ic actu­a­tor to press the key­board at spe­cif­ic times. This allowed us to look at the key met­rics of: accu­ra­cy (the dif­fer­ence between expect­ed and actu­al tim­ing), and pre­ci­sion (the vari­abil­i­ty of these dif­fer­ences). The phi­los­o­phy behind how we set up these test­ing rigs, was to rep­re­sent as close­ly as pos­si­ble the devices that par­tic­i­pants would have in their own homes — so we cre­at­ed the tasks on each plat­form as a nor­mal user would, and used stan­dard key­boards. In the sec­ond part of the paper, we look at the demo­graph­ics of a large sam­ple of Goril­la par­tic­i­pants (200,000 peo­ple), where they are, what plat­forms were used to recruit them, and what their devices looked like. This allows us to put the tim­ing results in the first part into con­text. The results show that for most plat­forms and devices, tim­ing was very good. It also showed that the most com­mon brows­er and device com­bi­na­tion of PCs and Chrome, showed suf­fi­cient­ly good tim­ing. If you want to see what the results are in full, then you should vir­tu­al­ly attend my talk, and if you are real­ly keen go and find the preprint on PsyArx­iv before hand. Thanks, and hope to see you there.