You should never run out of ways to drill yourself on the old material. After you have done a test, go through each and every question and look at each correct and incorrect answer choice. Under or beside each answer (correct or incorrect), record the reason for each answer's correctness or incorrectness using keywords, expressions and acronyms like these:
too strong/extreme (i.e. "ans. overreaches...not inferable"),
beyscope or
scopeshift (i.e. "ans. wrong b/c beyond scope"),
revans (i.e. "reverse answer"),
oppans ("i.e. ans. wrong b/c gives opposite"),
suff v. nec (i.e. "mistakes one for the other"),
correlcause (i.e. "mistakes correlation for causation"),
str (i.e. "ans. strengthens when asked to weaken"),
wkns (i.e. "ans. weakens when asked to strengthen"),
noeffect (i.e. "ans. neither strengthens nor weakens"),
circular (i.e. "ans. correct; flawed b/c stimulus employs circular reasoning"),
cause w/o effect (i.e. "ans. weakens b/c shows that the cause doesn't lead to stated effect"),
eff w/o cause (i.e. "ans. correct b/c shows that the effect can have other causes"),
altcause (i.e. "ans. weakens b/c presents possibility of an alternate cause"),
revrel (i.e. ans. weakens by presenting possibility of reverse relationship between stated cause and effect"),
statprob (i.e. "ans. weakens b/c points to statistical problem in argument"),
c>e (i.e. " ans. strengthens by increasing likelihood that the cause leads to stated effect"),
e<c (i.e. " ans. strengthens by increasing likelihood that the effect results from cause"),
elim. alt cause (i.e. "ans. strengthens by reducing/eliminating possibility of an alternate cause"),
addrstatprob (i.e. "ans. strengthens by eliminating statistical problem"),
elimrevrel(i.e. "ans. strengthens by making cause-effect relationship more likely...eliminates possibility of reversal"),
addrscopeshift (i.e. "is correct assumption b/c it addresses scope-shift"),
more paradoxical (i.e. "ans. makes stimulus MORE paradoxical"),
irrelevant (i.e. "ans. irrelevant")
notinferable/NI (i.e. "ans. could be true but incorrect b/c doesn't have to be true given the stimulus info...thus not a correct inference"),
unflawed (i.e. "ans. incorrect b/c not flawed"),
flawdiff (i.e. "ans. provides flaw different from that in stimulus"),
patt-reas (i.e. "ans. gives different pattern of reasoning from stimulus...suff-nec v. causal, etc.").
There are more types of right and wrong answers, but you get my point.
After you have done the answers from all 66 tests, go back and take them all timed again. See how friggin' ready you are then. 180 dawg!