Presentation on " ANALYSIS OF TED TALK BY MONA CHALABI ON 3 WAYS TO SPOT A BAD STATISTIC" made as a task for the internship on "DATA ANALYTICS WITH MANAGERIAL APPLICATIONS" under Professor Sameer Mathur, IIM Lucknow. Submitted by TARANG JAIN,DTU
5. VARIED
VIEWSON
STATISTICS
SOME SAY:
Statistics are
crucial, we need
them to make sense
of society as a
whole in order to
move beyond
emotional
anecdotes and
measure progress in
an objective way.
WHILE OTHERS SAY:
Statistics are
elitist, maybe even
rigged; they don't
make sense and they
don't really
reflect what's
happening in
people's everyday
lives.
7. 1.CANYOU
SEE
UNCERTAINT
Y
Many a times there may be a lot of variations in
the data set itself.
Thus taking the average of such a data set can
lead to very misleading conclusions,
So whenever we see a statistic, the first step is
to heck for the uncertainty in the data.
8. SMALL
SAMPLESIZE
If the sample size used to predict a result is too
small, then the chances are that the statistic
wrong.
So another way to check for uncertainty is by
observing the size of sample.
9. 2.CAN ISEE
MYSELF IN
THE DATA
Whenever we see a statistic, we shall try to see if
we fit into the data.
Even if we cannot directly fit into the dataset we
shall try to get as much context as possible.
We shall try to zoom out from one point and see
the variation of statistic for various sections of
the dataset.
10. 3. HOWWAS
THE DATA
COLLECTED?
Statisticians many times tend to overlook one
fact over another and that can lead to severely
misleading results.
We shall look into what data was actually
collected in a survey and then also investigate
which factor in the survey was given precedence
over which factor
As the precedence of factors can bend the
statistic towards either way.
12. Whenever a manager is faced with a
statistic, he shall look into variance in the
data set used and see if there is skew in it.
He shall always ask his analysts to never go
by the average of the data, as it can lead to
very misleading conclusions.
14. ENSURETHAT
THESOURCE
IS RELIABLE
ANDGOOD.
He shall ensure that his analysts are given
good data source to create statistics.
A good data set must have the following
properties:-
1.A big sample size
2.Contains all the data the valid
parameters that may affect the result.