1
00:00:02,746 --> 00:00:06,506
In today's class, I want to show
you how to not run an experiment.

2
00:00:07,596 --> 00:00:10,276
In the previous class, I showed you the results

3
00:00:10,276 --> 00:00:13,286
from a very simple experiment
that was run correctly.

4
00:00:14,146 --> 00:00:20,026
We considered our outcome to be the profit we
made in the store and there were two factors.

5
00:00:20,466 --> 00:00:22,986
The first was the amount of light in the store

6
00:00:23,426 --> 00:00:25,976
and the second was the price
of the product we are selling.

7
00:00:27,006 --> 00:00:29,366
The outcome was the profit we could make.

8
00:00:30,686 --> 00:00:34,436
You'll recall from those experiments
we had two continuous factors.

9
00:00:34,436 --> 00:00:40,176
The first factor was the amount of light:
either at 50% or 75% on the dimmer.

10
00:00:40,176 --> 00:00:49,546
The second factor was the price of
the product: either $7.79 or $8.49.

11
00:00:51,746 --> 00:00:55,436
Here's how NOT to run the experiments.

12
00:00:55,436 --> 00:00:58,296
Many people will run only 3 experiments.

13
00:00:58,676 --> 00:01:00,666
They consider the first experiment here

14
00:01:00,666 --> 00:01:03,276
in the bottom left corner to
the be their starting point.

15
00:01:03,276 --> 00:01:07,836
This is the experiment with
low light and low prices.

16
00:01:09,116 --> 00:01:14,046
Then, they figure they need to move
across and only increase the light

17
00:01:14,506 --> 00:01:16,426
and run their second experiment over here.

18
00:01:16,426 --> 00:01:23,306
After that, they go back to the starting point,
and move up from there and do a third experiment

19
00:01:23,306 --> 00:01:27,486
over here, increasing only the price.

20
00:01:27,486 --> 00:01:30,926
Many people assume this is the
correct way to do experiments,

21
00:01:31,216 --> 00:01:34,836
because every time we are
only changing one thing.

22
00:01:35,376 --> 00:01:40,456
You've been taught that in school and in
university: "only change one thing at a time".

23
00:01:40,456 --> 00:01:48,176
Now if you did these 3 experiments, what
you notice is that you only get ONE estimate

24
00:01:48,176 --> 00:01:54,186
of the effect of lighting, and only ONE
estimate of the effect of changing price.

25
00:01:55,366 --> 00:02:00,206
Recall, in the previous videos I
said you should do 4 experiments.

26
00:02:00,206 --> 00:02:04,696
With just that one extra experiment,
up here in the top right hand side,

27
00:02:05,356 --> 00:02:09,006
we can actually estimate the effects twice.

28
00:02:09,006 --> 00:02:14,266
We have have two estimates of lighting changes
and we have two estimates of price changes.

29
00:02:14,266 --> 00:02:17,796
So by adding that one extra experiment,

30
00:02:18,086 --> 00:02:22,216
what we have essentially done is doubled
the amount of information we have achieved.

31
00:02:23,296 --> 00:02:28,306
In many cases this is well worth the
extra work of doing that extra experiment.

32
00:02:29,856 --> 00:02:32,536
Now in the next few modules I'm going to
show you there's another good reason why

33
00:02:32,536 --> 00:02:34,736
that 4th experiment is so crucial.

34
00:02:34,736 --> 00:02:37,896
So there's lots to look forward
to in this course.

35
00:02:38,066 --> 00:02:42,466
Finally, we'd like to challenge you to think
about your own experiments that you can work

36
00:02:42,466 --> 00:02:44,506
on for the duration of this course.

37
00:02:44,506 --> 00:02:47,776
This makes your learning far more applicable.

38
00:02:47,776 --> 00:02:51,856
Do you have an objective in mind already?

39
00:02:51,896 --> 00:02:56,326
What are the factors that will influence
the outcome from your experiments?

40
00:02:56,606 --> 00:03:04,136
Maybe the experiments are related
to water treatment, or energy usage,

41
00:03:04,136 --> 00:03:07,356
It might be a recipe that you want to fine-tune.

42
00:03:07,356 --> 00:03:11,096
It might be experiments that
you'd like to run at work,

43
00:03:11,096 --> 00:03:15,696
or experiments that will influence the people
around you: your community and your neighbours.

44
00:03:15,776 --> 00:03:17,636
We'd like you to film a short
video, upload it to YouTube,

45
00:03:17,636 --> 00:03:18,766
and share the video link on the course forums.

46
00:03:18,796 --> 00:03:19,846
Show people around you what you are working on.

47
00:03:19,876 --> 00:03:21,856
Myself, and others, would like to comment
on your work, and give you some feedback.

48
00:03:21,886 --> 00:03:22,606
See you in the coming modules.