Common Statistical Flaws : CR

 Learn to recognize that statistics  , don’t always lead directly to the conclusions that one might try to draw from them. GMAT will  test you on this concept repeatedly.
Commonly tested flaws using statistics include:

• Wordplay—
A statistic does not match directly the conclusion that follows
(e.g., arrests vs. crimes)
 Precision in wording is an oft tested concept on the GMAT. With statistics in particular, the authors of the GMAT are able to take your eye off of the premise and conclusion, and invite
you to focus on the numbers. Know this: When you see statistics in Critical
Reasoning questions, there is almost always a flaw in logic. Read critically and
ensure that the numbers are tied to an oranges to oranges comparison.

• Absolute Number vs. Percentage/Proportion
Questions will often try to blind you with numbers in a situation that requires
percentages or proportions. For example, one could argue that Texas
does not carry its weight in contributing U.S. federal income taxes, because
it contributes only $54 billion per year, while California contributes well over
$300 billion. But that absolute number is misleading: Texas actually
contributes more federal tax money per citizen than any other state. It just
happens to have a relatively low population.

• Unequal Basis Points/“Unweighted Averages”—
Uneven sample sizes are compared to a common third pool (e.g., more people die from DRUG OVERDOSE each year than are killed by great white sharks; therefore it is safer to
swim with sharks than to TAKE OVERDOSE OF A DRUG)
Perhaps the most “statistical” of statistical flaws, the unweighted average is
problematic. Consider the very famous statistic that “most traffic accidents
happen within 10 miles of the victim’s home.” Does that really mean that it’s
more dangerous to drive around your neighborhood than to drive in a remote
area that you do not know? Of course not. It’s just that the first and last 10
miles of nearly every trip you take are within 10 miles of your home. So a
massive percentage of your driving takes place there. Using the basis point “all
accidents” is misleading.

• Incongruent Samples—
Two “equivalent” statistics were not obtained in
the same fashion (e.g., a local, low-cost, part-time MBA program with no
application fee vs. Stanford ; even if the local program’s acceptance rate is
low, are its applicants analogous to Stanford’s?)
Often data is flawed because the statistics are simply not parallel. Even
statistics that are comparable (percentage to percentage) can have intervening
factors in their sample pools or collection procedures that leave them less
than concrete in proving a conclusion. For instance, Harvard has an extensive
(and expensive) application procedure and is known to have extremely high
admissions standards. For many it may simply not be worth the time and
application fee to apply without a high likelihood of success. Other schools
may have lower barriers to application—famously, some undergraduate
schools have accepted applications through Twitter—that attract a high
number of applicants for a low number of seats, creating a low, seemingly
selective, acceptance rate that is in fact really not that selective.
Again, the difference may be clearest in wordplay. One can say that a school
“has a lower acceptance rate than Stanford,” but as soon as one takes the
statistic and concludes something in different terms (“more selective” is not
the same as “lower acceptance rate”) there exists a subtle gap in logic that an
answer choice can exploit.

WE shall discuss some questions based on  confusing statistics tomorrow on this blog...Stay tunes peeps.