WEBVTT

00:00:00.520 --> 00:00:02.820
Let's say I've got
a set of numbers.

00:00:02.820 --> 00:00:08.710
2, say I've got three 3's, I've
got a couple of 4's, and

00:00:08.710 --> 00:00:10.620
I've got a 10 there.

00:00:10.620 --> 00:00:14.210
And what we want to do is find
the middle of these numbers.

00:00:14.210 --> 00:00:17.820
We want to represent these
numbers with the center of the

00:00:17.820 --> 00:00:19.900
numbers, or the middle of the
numbers, just so we have a

00:00:19.900 --> 00:00:24.350
sense of where these numbers
roughly are.

00:00:24.350 --> 00:00:27.360
And this central tendency that
we're going to try to get out

00:00:27.360 --> 00:00:31.020
of these numbers, we're going
to call the average.

00:00:31.020 --> 00:00:34.990
The average of this
set of numbers.

00:00:34.990 --> 00:00:38.840
And you've, I'm sure, heard the
word average before, but

00:00:38.840 --> 00:00:41.110
we're going to get a little
bit more detailed on the

00:00:41.110 --> 00:00:44.270
different types of averages
in this video.

00:00:44.270 --> 00:00:46.890
The one you're probably most
familiar with, although you

00:00:46.890 --> 00:00:49.950
might have not seen it referred
to in this way, is

00:00:49.950 --> 00:00:58.290
the arithmetic mean, which
literally says, look, I, the

00:00:58.290 --> 00:01:01.270
arithmetic mean of this set of
numbers, is literally the sum

00:01:01.270 --> 00:01:03.420
of all of these numbers divided
by the number of

00:01:03.420 --> 00:01:04.209
numbers there are.

00:01:04.209 --> 00:01:07.440
So the arithmetic mean for this
set right here is going

00:01:07.440 --> 00:01:17.070
to be 2 plus 3 plus 3 plus 3
plus 4 plus 4 plus 10, all of

00:01:17.070 --> 00:01:19.230
that over, how many
numbers do I have?

00:01:19.230 --> 00:01:22.870
1, 2, 3, 4, 5, 6, 7.

00:01:22.870 --> 00:01:24.720
All of that over 7.

00:01:24.720 --> 00:01:25.660
And what is this equal to?

00:01:25.660 --> 00:01:34.560
This is 2 plus 9, which is 11,
plus 8, which is 19, plus 10,

00:01:34.560 --> 00:01:35.520
which is 29.

00:01:35.520 --> 00:01:41.290
So this is going to be equal to
29/7, or you could say it's

00:01:41.290 --> 00:01:43.540
equal to 4 and 1/7.

00:01:43.540 --> 00:01:45.240
If I got my calculator out,
we could figure out

00:01:45.240 --> 00:01:46.480
the decimal of this.

00:01:46.480 --> 00:01:50.990
But this is a representation
of the central tendency, or

00:01:50.990 --> 00:01:52.450
the middle of these numbers.

00:01:52.450 --> 00:01:53.670
And it kind of makes sense.

00:01:53.670 --> 00:01:57.670
4 and 1/7, it's a little
bit higher than 4.

00:01:57.670 --> 00:02:01.070
We're kind of close to the
middle of our number range

00:02:01.070 --> 00:02:01.810
right there.

00:02:01.810 --> 00:02:03.590
And you might say, well, it's
a little skewed to the right

00:02:03.590 --> 00:02:04.700
and what caused that?

00:02:04.700 --> 00:02:06.950
And well, gee, 10 is a little
bit larger than all of the

00:02:06.950 --> 00:02:07.840
other numbers.

00:02:07.840 --> 00:02:09.320
It's kind of an outlier.

00:02:09.320 --> 00:02:13.370
Maybe that skewed this average
up, the arithmetic mean.

00:02:13.370 --> 00:02:16.680
So there are other types of
averages, although this is the

00:02:16.680 --> 00:02:19.200
one that, if people just say,
hey, let's take the average of

00:02:19.200 --> 00:02:21.660
these numbers, and they don't
really tell you more, they're

00:02:21.660 --> 00:02:23.940
probably talking about
the arithmetic mean.

00:02:23.940 --> 00:02:30.270
The other forms of average,
though, are the median, and

00:02:30.270 --> 00:02:32.290
this literally is the
middle number.

00:02:35.970 --> 00:02:38.360
If there are two middle numbers,
you actually take the

00:02:38.360 --> 00:02:40.730
arithmetic mean of those
two middle numbers.

00:02:40.730 --> 00:02:43.310
You actually find the number
halfway in between those two

00:02:43.310 --> 00:02:44.330
middle numbers.

00:02:44.330 --> 00:02:47.260
So the median of this set
right here-- let me just

00:02:47.260 --> 00:02:48.350
rewrite them.

00:02:48.350 --> 00:02:57.060
So I have a 2, a 3,
3, 3, 4, 4, 10.

00:02:57.060 --> 00:02:59.690
So, let's see, we have seven
numbers right here.

00:02:59.690 --> 00:03:02.800
The middle number, if
I go 1, 2, 3, to the

00:03:02.800 --> 00:03:04.140
right, we're there.

00:03:04.140 --> 00:03:06.470
If we go 1, 2, 3 to the
left, we're there.

00:03:06.470 --> 00:03:09.850
The middle number is
that 3 right there.

00:03:09.850 --> 00:03:13.360
I just listed them in order, and
I said, well, look, 3, you

00:03:13.360 --> 00:03:16.640
could think of it as the fourth
number from the right,

00:03:16.640 --> 00:03:19.550
and it's also the fourth
number from the left.

00:03:19.550 --> 00:03:21.360
3 is the middle number.

00:03:21.360 --> 00:03:23.940
And this case, it
is the median.

00:03:23.940 --> 00:03:29.280
So in this case, 3, if you use
the median, is our average.

00:03:29.280 --> 00:03:30.360
And that also makes sense.

00:03:30.360 --> 00:03:32.030
I mean, it's literally the
middle number, and if you look

00:03:32.030 --> 00:03:35.670
at this set of numbers, it kind
of does represent the

00:03:35.670 --> 00:03:37.790
central tendency of this set.

00:03:37.790 --> 00:03:40.880
Now just to be clear, it was
very clear what the middle

00:03:40.880 --> 00:03:45.350
number was, because I had an
odd number of numbers.

00:03:45.350 --> 00:03:48.560
I had three on each side of the
three, so it was very easy

00:03:48.560 --> 00:03:50.790
to figure out the median,
the middle number.

00:03:50.790 --> 00:03:53.800
But if I had a situation-- let's
say I have the situation

00:03:53.800 --> 00:03:57.850
where I have 2, 3, 4, and 5.

00:03:57.850 --> 00:04:01.510
Let's say that's my
set of numbers.

00:04:01.510 --> 00:04:04.350
Well, here, there is no
one middle number.

00:04:04.350 --> 00:04:07.460
The 3 is closer to the left
than it is to the right.

00:04:07.460 --> 00:04:10.040
The 4 is closer to the right
than it is to the left.

00:04:10.040 --> 00:04:12.220
There's actually two middle
numbers here.

00:04:12.220 --> 00:04:17.380
The two middle numbers here
are the 3 and the 4.

00:04:17.380 --> 00:04:20.640
And here, when you have two
middle numbers, which occurs

00:04:20.640 --> 00:04:24.430
when you have an even number
in your data set, there the

00:04:24.430 --> 00:04:27.630
median is halfway in between
these two numbers.

00:04:27.630 --> 00:04:31.070
So in this situation, the median
is going to be 3 plus 4

00:04:31.070 --> 00:04:35.670
over 2, which is equal to 3.5.

00:04:35.670 --> 00:04:37.610
And if you look at this data
set, that's not what our

00:04:37.610 --> 00:04:44.430
original problem was, but if
you look at this data set

00:04:44.430 --> 00:04:47.770
right there, you're actually
going to find that the

00:04:47.770 --> 00:04:52.070
arithmetic mean and the median
here is the exact same thing.

00:04:52.070 --> 00:04:53.230
Let's calculate it.

00:04:53.230 --> 00:04:55.470
What 's the arithmetic
mean over here?

00:04:55.470 --> 00:05:00.240
It's going to be 2 plus 3 plus
4 plus 5, which is what?

00:05:00.240 --> 00:05:09.490
5 plus 9, which is equal
to 14, over 4.

00:05:09.490 --> 00:05:10.910
And what's this equal to?

00:05:10.910 --> 00:05:16.740
14/4 is 3 and 2/4, or 3 and
1/2, the exact same thing.

00:05:16.740 --> 00:05:19.310
So for this data set, they
were the same thing.

00:05:19.310 --> 00:05:24.020
For this data set, our median
is a little bit lower.

00:05:24.020 --> 00:05:28.090
It's 3, while our arithmetic
mean is 4 and 1/7.

00:05:28.090 --> 00:05:30.080
And I really want you to think
about why that is.

00:05:30.080 --> 00:05:32.840
And it has a lot to do with this
10 that sits out there.

00:05:32.840 --> 00:05:37.190
All of these other numbers are
pretty close to whichever

00:05:37.190 --> 00:05:39.590
average you want to pick,
whether it's the arithmetic

00:05:39.590 --> 00:05:42.560
mean or it's the median.

00:05:42.560 --> 00:05:47.860
But this 10 is kind of
an outlier, or it

00:05:47.860 --> 00:05:50.650
skews the data set.

00:05:50.650 --> 00:05:53.790
Maybe it's so much larger than
the other numbers, that it

00:05:53.790 --> 00:05:57.680
makes the arithmetic mean seem
larger than maybe is

00:05:57.680 --> 00:05:59.670
representative of
this data set.

00:05:59.670 --> 00:06:01.500
And that's something important
to think about.

00:06:01.500 --> 00:06:07.570
When you're finding the average
for something, most

00:06:07.570 --> 00:06:09.740
people will immediately go
to the arithmetic mean.

00:06:09.740 --> 00:06:13.520
But in a lot of cases, median
will make a lot more sense, if

00:06:13.520 --> 00:06:16.970
you have these really large or
really small numbers that

00:06:16.970 --> 00:06:18.460
could skew the data set.

00:06:18.460 --> 00:06:21.530
I mean, you can imagine, if this
wasn't a 10-- or let's

00:06:21.530 --> 00:06:22.790
imagine adding another
number here.

00:06:22.790 --> 00:06:27.810
If I added the number 1 million,
if I added 1 million

00:06:27.810 --> 00:06:30.390
to this data set, if that was
the eighth number, the

00:06:30.390 --> 00:06:32.260
arithmetic mean is going
to be this huge number.

00:06:32.260 --> 00:06:36.400
It's going to be much larger
than what is representative of

00:06:36.400 --> 00:06:38.470
most of the numbers
in this data set.

00:06:38.470 --> 00:06:40.240
But the median is still
going to work.

00:06:40.240 --> 00:06:43.730
The median is still going to be
about 3 and a half, right?

00:06:43.730 --> 00:06:47.750
If you had 1 million here,
it would be 1, 2, 3, 4.

00:06:47.750 --> 00:06:49.200
The middle two numbers
would be that.

00:06:49.200 --> 00:06:50.580
It would be 3 and 1/2.

00:06:50.580 --> 00:06:53.900
So the median is less sensitive
to one or two

00:06:53.900 --> 00:06:58.430
numbers at the extremes that
otherwise would skew the mean.

00:06:58.430 --> 00:07:01.130
Now, the last form of average
I want to talk

00:07:01.130 --> 00:07:02.840
about is the mode.

00:07:05.400 --> 00:07:08.000
It has nothing to do
with ice cream.

00:07:08.000 --> 00:07:10.910
The mode is literally the
most frequent number.

00:07:16.800 --> 00:07:19.880
And in this data set, it's
pretty clear what the most

00:07:19.880 --> 00:07:21.130
frequent number is.

00:07:21.130 --> 00:07:25.820
I only have one 2, I have three
3's, I have two 4's, I

00:07:25.820 --> 00:07:28.445
have one 10, and even if want to
include the million, I only

00:07:28.445 --> 00:07:29.710
have one million there.

00:07:29.710 --> 00:07:31.990
So here, the number
that occurs most

00:07:31.990 --> 00:07:35.360
frequently is the 3.

00:07:35.360 --> 00:07:38.330
So, once again, the mode seems
like a pretty good measure of

00:07:38.330 --> 00:07:41.260
central tendency or a
pretty good average

00:07:41.260 --> 00:07:43.300
for this data set.

00:07:43.300 --> 00:07:45.950
Now the mode, it's a little
tricky to deal with, and you

00:07:45.950 --> 00:07:49.070
won't see it used that often,
because it becomes a little

00:07:49.070 --> 00:07:52.920
ambiguous when-- you know,
look at this data set:

00:07:52.920 --> 00:07:55.270
2, 3, 4, and 5.

00:07:55.270 --> 00:07:56.370
What is the mode there?

00:07:56.370 --> 00:07:59.200
All of these numbers are
equally frequent.

00:07:59.200 --> 00:08:01.360
So if you have a situation
like this, then you might

00:08:01.360 --> 00:08:03.930
just-- the mode really
loses its meaning.

00:08:03.930 --> 00:08:06.350
It might force you anyway to
take the median or the

00:08:06.350 --> 00:08:07.530
mean in some form.

00:08:07.530 --> 00:08:10.700
But if you really do have
numbers that one shows up a

00:08:10.700 --> 00:08:14.640
lot more than the other, then
the mode starts to make sense.

00:08:14.640 --> 00:08:18.220
So, hopefully, this has given
you a pretty good overview of

00:08:18.220 --> 00:08:25.400
how to represent the central
tendency of a data set.

00:08:25.400 --> 00:08:26.430
Very fancy word.

00:08:26.430 --> 00:08:28.180
But it's just saying, look,
we're trying to represent with

00:08:28.180 --> 00:08:30.200
one number all of this data.

00:08:30.200 --> 00:08:31.810
And you might say, hey, why do
we even worry about that?

00:08:31.810 --> 00:08:34.059
It only has seven numbers here
or eight numbers here.

00:08:34.059 --> 00:08:36.770
But you can imagine if you had 7
million numbers or 7 billion

00:08:36.770 --> 00:08:39.440
numbers, and you don't want to
show someone all of that data.

00:08:39.440 --> 00:08:42.210
You just want to give someone a
sense of what those numbers

00:08:42.210 --> 00:08:45.345
are on average.

00:08:45.345 --> 00:08:49.420
And as we said, the arithmetic
mean is what I see being used

00:08:49.420 --> 00:08:53.170
the most. But in situations
where you might have numbers

00:08:53.170 --> 00:08:55.940
that would skew the arithmetic
mean, because they're so large

00:08:55.940 --> 00:08:58.670
or they're so small,
the median might

00:08:58.670 --> 00:09:00.480
make a lot of sense.