Specs for Mozer1a and Mozer1b 

 

Subjects will perform Mozer1a on day 1, and then Mozer1b one day later. 

 

Mozer1a 

 

1. Use 36 Face-Name pairs (Noriko has these). 

 

2. Mozer1a runs through the list of 36 pairs 3 times (3 blocks). Usually the subject just sees both, with the face above the name, for 8 seconds. However, sometimes the subject is given the face and forced to guess on the name, whether they have any idea or not. We are varying (across pairs) which of the three blocks the forced test occurs in (1, 2, 3, or never). In a forced test the subject sees the face and has 8 seconds to start typing in the name; if they haven’t started by 8 sec, prompt them again “Please enter an answer without delay”. Once they do respond, show them the correct answer for 4 seconds.  

 

DESIGN 

 

Randomly and independently for each subject: assign 9 of the face-name pairs to each condition 1-4. 1 means “forced test on first block”, …, 3 means “forced test on third block”; 4 means “never gets a forced test”. 

 

Once that’s done, the essence of the program is trivial:

______________________

For block = 1 to 3 do begin

Shuffle p a random list of digits (1-36)

For x = 1 to 36

Present face p[x]; [if the current block number = the condition for that

word then force the subject to guess, else just present the face and    name together]

End;

-------------------------------- 

 

So suppose number 25 on your list is assigned to condition1, ie p[25]=1. Then in block 1, the subject will see the face and have to guess on the name (even though they cannot possibly know the correct answer since it’s the first trial). If word pair number 1 is assigned to condition 4, then the subject will never have a forced test on that word pair. 

 

Mozer1b 

 

This program is a quick test that will be given to the subjects the next day. Just randomize the list of names and forced test the subject on each picture, with no feedback. 

 

Data Storage and Analysis. 

 

As always, save everything. In Mozer1a, the raw data file should have 1 row per trial, listing the trial number, the name, the condition for that name, what was shown to the subject, what would be the correct response, whether it was a forced-test, and if so, how the subject responded. The data file from Mozer1b should contain a list of the names and how the subject responded. 

 

We want to be able to recover the probability of items being correct in Mozer1b as a function of what condition they were tested in in Mozer1a. The data analysis program should run through the raw data files from both, and reconstruct how many faces that were presented in Mozer1a in each condition were responded to correctly in Mozer1b.

*updated Aug15