WEBVTT
1
00:00:01.139 --> 00:00:04.240 A:middle L:90%
this question asks Win, Is it appropriate to use
2
00:00:04.240 --> 00:00:09.019 A:middle L:90%
the median as a measure of central tendency? Well
3
00:00:09.029 --> 00:00:11.630 A:middle L:90%
, we knew that we've got to measures of central
4
00:00:11.630 --> 00:00:14.589 A:middle L:90%
tendency right now, mean and media, when you
5
00:00:14.589 --> 00:00:19.250 A:middle L:90%
have a distribution, a range of numbers that has
6
00:00:19.670 --> 00:00:22.440 A:middle L:90%
looks like this, a lot of numbers closer to
7
00:00:22.440 --> 00:00:27.820 A:middle L:90%
the middle and fewer numbers in the outside's the median
8
00:00:28.129 --> 00:00:31.239 A:middle L:90%
, which I'll call em, and the Mean X
9
00:00:31.239 --> 00:00:33.789 A:middle L:90%
bar are going to be about the same when it's
10
00:00:33.799 --> 00:00:36.350 A:middle L:90%
symmetrical like this. We call this a symmetrical distribution
11
00:00:37.939 --> 00:00:40.850 A:middle L:90%
, but consider what happens when we have a distribution
12
00:00:41.439 --> 00:00:46.520 A:middle L:90%
that looks more like this where we have most of
13
00:00:46.520 --> 00:00:51.170 A:middle L:90%
our values here, but a couple extreme values way
14
00:00:51.170 --> 00:00:55.149 A:middle L:90%
out here. What happens then? Well, the
15
00:00:55.149 --> 00:00:58.990 A:middle L:90%
median we know is just the middle, right?
16
00:00:58.990 --> 00:01:02.090 A:middle L:90%
So it doesn't really matter where the high values are
17
00:01:02.090 --> 00:01:03.650 A:middle L:90%
and where the low values are. The median will
18
00:01:03.650 --> 00:01:08.890 A:middle L:90%
always stay the same, so the media will still
19
00:01:08.890 --> 00:01:12.609 A:middle L:90%
be in the middle of that big bump. But
20
00:01:12.609 --> 00:01:17.829 A:middle L:90%
the mean does depend on extreme values. In fact
21
00:01:17.870 --> 00:01:21.540 A:middle L:90%
, extreme values, uh, haven't even greater effect
22
00:01:21.540 --> 00:01:23.969 A:middle L:90%
on the mean. Then average values do so when
23
00:01:23.969 --> 00:01:26.379 A:middle L:90%
you have values that are way out here in the
24
00:01:26.390 --> 00:01:30.439 A:middle L:90%
tails of the distribution is going to affect the mean
25
00:01:30.579 --> 00:01:36.189 A:middle L:90%
and bring it closer to the tails. That means
26
00:01:36.189 --> 00:01:38.459 A:middle L:90%
when we have distributions like this with extreme values,
27
00:01:38.469 --> 00:01:42.269 A:middle L:90%
the median is a better measure of central tendency.