How to set up the errors order matrix

To make the practice rows:

1-16 : same for all CB, present all 16 stims

17-18 use CounterBal.xls to make locations

19, 50, 81, 11, 143, 174, 250 – vary based on CB

Use special headings/constants as outlined in Matlab Code for phase 99

Contone, dimone, decone, conttwo, dimtwo, dectwo

Presents same context box twice, each with the dimension boxes

20-49: write context # x30

dim1# x15

dim2# x15

dec1# x15

dec2# x15

221-227: practice each set of contexts, no labels, indefinite

List each 99 row again

228-283: practice trials identical to test, see test set up

284: redo practice?

285: starts test

Test setup

InitialErr.xls has all the trial types listed

Copying those 24 3x = 1 block/run

Copying that 1 run of 72 x8 = entire experiment

Leave 3 spaces at the end of each run for 3 catch trials

Catch trials

Under other 24 in initial err

Copy them 6 times, then randomly sort

Insert into last 3 spaces of each run

Randomly sort within each run

Stims: the list of all 16 stims is at the end of InitialErr

Copy it and paste down to the end of the 8 runs

Shuffle randomly

Remove rand column, i.e. pair stims with trials

Match items

Number 1-600 on left end of columns

Sort columns by “DEC” then “SHAPE” then “RESP”

If decision = 1, this means shape

If resp = 1, this means match

Depending on CB, text could be in one of the four (LT, LB, RT, RB) locations. Find out which, then in the corresponding column type:

IF(RESP=1, cell under SHAPE, 2)

You’re saying if the trial is a match, make the shape location stimulus the same as the trial shape stim

Drag this down the column until Dec changes

When resp changes to 2, change the “2” in the if statement to a 1

Repeat this for orient, texture, size and respective columns

Then type 1,2/3,4/5,6/7,8 into 4 columns an even number of times for all the DEC trials

Put rand() next to each column and sort

Insert the columns into the 3 remaining location columns for all the DEC1 trials

You won’t need to do ¼ because that’s your match location

Repeat this for all 4 decision types, resorting each time

Sort all the columns based on the far left numbers

Insert switch columns to measure the amounts of switching

Image label meanings: Texture, Orientation, Shape, Size

e.g. 1111, Texture = 1, Orientation = 1, etc

Texture: 1 = pebbled, 2 = metal

Orientation: 1 = up, 2 = down

Shape: 1 = lighthouse, 2 = top

Size: 1 = large, 2 = small

For corners,

1 = pebbled

2 = metal

3 = up

4 = down

5 = top

6 = lighthouse

7 = large

8 = small

It can’t be 1 and 2 because the four corners change based on counterbalance number. If it were just 1 and 2, then the fprintf code would read “if 1 print ‘pebbled’” and “if 1 print ‘up’” which obviously doesn’t work