EFTA02667128

Document Text

From: Peter Fenwick Sent: Thursday, May 10, 2018 3:27 PM To: jeffrey E. Subject: FW: Sunday experiment Results of the Sunday =xperiment analysed by Fabio. 4 Subjects 18 trials and the =orrelation matrix. We will try and do another this Saturday. =ou mentioned a dual pen recorder to confirm timing. That I will leave =o the main experiment as we will have plenty of timings available in =hat. This is a simple experiment, is it Cult or is it =nteresting? Looks interesting! Let me know if you want =nything else for Saturday. Peter From: Fabio Babiloni =mailto Sent: 07 May 2018 =0:14 To: Peter iuseppe Bersani •‘>; Corazza, Ornella =1 Andreas loannides Viirre, Erik rget Subject: R: =undayexperiment ° Dear Peter, thank you for sending such numbers to me. Indeed, it seems =hat something is moving there © I have to state several =hings: 1. If the number of correlation tests performed is 4 (e.g. =lain against the four students) then the level of r to be taken into =onsideration by using Bonferroni correction is as follows (r = =,57) From this point of view we are quite close the same to the =ignificance for some of them. EFTA_R1_01918755 EFTA02667128 2. However, Bonferroni =orrection is something that is quite harsh but it is indeed appropriate =or multiple execution of unilateral tests. Maybe Holm will be more =ppropriate (resulting maybe in a lower threshold). 3. If you run the =orrelation everybody as everybody we obtained e.g a correlation also