Please wait while the transcript is being prepared...
0:03
I constantly use the
term data distribution.
Here is a density plot
of weight in pounds.
Immediately if you're acquainted
with graphics in biostatistics,
notice that I'm using density
plot and not histogram.
Density plots are good for
representing proportion.
Histograms are common
for actual counts.
This is a density plot,
because look at the
scale on the y-axis
to the left.
Notice we're not
showing whole numbers.
This is density distribution,
and the weight in pounds
shows up on the x-axis.
If you're acquainted with the
concept of normal distribution
and what we call the
bell-shaped curve,
then just by looking at this
and I'm going to use some
value laden terms here,
I would say that
this distribution
approximates normal
distribution.
But, I would not say
with any confidence
that it represents true
normal distribution,
a perfect bell-shaped curve.
It would be the rare phenomenon
in biostatistics that does.
But this gives you certainly
a good sense of
data distribution.
For the next slide
before I show it,
let me mention the term facet,
F-A-C-E-T.
We're looking at density
plot of weight in pounds.
1:27
Not a histogram.
It's a density plot.
But I'm breaking
it out by gender.
By inserting just
one line of syntax
in the programming associated
with density plot,
I'm able to present
it this way showing
breakout between
female and male.
They're both on the same scale
so I can make a good comparison
from 100-250 pounds
and the density
distribution values
are on the left, the y-axis.
You have a good sense of
data distribution and values
using this density plot broken
out by the two genders.
A facet F-A-C-E-T approach is
often good for three, four,
maybe up to five
breakouts in a group.
Beyond that it gets a
little hard to know
where to begin and where to end
to discriminate between
one and the other.
Certainly for two, three
and four it's fine.
