Nate:

No ghana - we looking at Pgh data

Elise: 

motivaton, description of listen tutor...

outline...

white oak - fort pitt

test
R C 
test
C R
test

test batteries

april:

Fall-Group1 / Fall-Group2: better Fall/Tutor, Fall/Ctrl...

dropout for 4th graders

can't use nonparticipants as ctrls because participants were selected
for low achievement... white oak also selected better students to do
tutor in fall... there is a rich-get-richer effect...

Nate...

looxv didn't include tutor very often but it tended to be positive...

zoom the graphs...

Elise...

Power analysis to detect 8 word difference in oral reading fluency (or
roughly 1 month of progress...)... forget sample size... (roughly 90...)

April...

ASRE RCT workshop recs

REcs for Kane...

  - seasonal vs tutor effects

  - comparable groups

  - usage measures




* Check assumptions on t test... (esp for power, might want to
  bootstrap or simulate...)

* what about the simple ancova for gain = pretest + tx,   (fall only)

  (they did some all-subsets thing instead...)