Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:00,004 --> 00:00:02,006
- [Instructor] When you gather data about your business,
2
00:00:02,006 --> 00:00:05,000
you will often analyze samples
3
00:00:05,000 --> 00:00:07,003
instead of an entire data set.
4
00:00:07,003 --> 00:00:10,005
And that's because gathering data can be very expensive
5
00:00:10,005 --> 00:00:13,004
both in terms of money and time.
6
00:00:13,004 --> 00:00:16,004
That said, you should gather as large a sample as you can,
7
00:00:16,004 --> 00:00:19,007
interview as many customers, review as many products,
8
00:00:19,007 --> 00:00:22,009
to make sure that you have all the data that you need.
9
00:00:22,009 --> 00:00:24,008
More is better.
10
00:00:24,008 --> 00:00:27,009
From there, you can estimate the data population's
11
00:00:27,009 --> 00:00:31,000
standard deviation based on that sample.
12
00:00:31,000 --> 00:00:33,008
And then you determine your desired confidence level.
13
00:00:33,008 --> 00:00:36,007
Do you want to be 90% confident in your result?
14
00:00:36,007 --> 00:00:40,006
Do you want to be 95% confident, 98, 99?
15
00:00:40,006 --> 00:00:43,009
It all depends on how precise you want to be.
16
00:00:43,009 --> 00:00:46,004
From there, you can calculate your margin of error
17
00:00:46,004 --> 00:00:49,007
and that will tell you, at your given confidence level,
18
00:00:49,007 --> 00:00:53,006
how much above or below your values will be.
19
00:00:53,006 --> 00:00:56,002
So what is margin of error?
20
00:00:56,002 --> 00:01:00,000
Well, let's say that each bottle contains 12 ounces
21
00:01:00,000 --> 00:01:04,000
of olive oil with a margin of error of plus or minus
22
00:01:04,000 --> 00:01:08,008
0.3 ounces at a 95% level of confidence.
23
00:01:08,008 --> 00:01:13,008
So you want to be 95% confident that the true margin
24
00:01:13,008 --> 00:01:16,009
of error is plus or minus 0.3 ounces
25
00:01:16,009 --> 00:01:19,003
based on the mean of 12.
26
00:01:19,003 --> 00:01:22,003
You rarely see the confidence level stated explicitly
27
00:01:22,003 --> 00:01:25,009
in political polls or in business analysis
28
00:01:25,009 --> 00:01:29,008
but it is always there and it is useful information.
29
00:01:29,008 --> 00:01:31,005
You calculate the margin of error,
30
00:01:31,005 --> 00:01:34,002
first by calculating your standard error.
31
00:01:34,002 --> 00:01:37,009
And that is the standard deviation of the data set,
32
00:01:37,009 --> 00:01:40,005
in this case the sample that you've taken,
33
00:01:40,005 --> 00:01:42,006
and dividing that by the square root
34
00:01:42,006 --> 00:01:45,001
of the number of measurements.
35
00:01:45,001 --> 00:01:49,003
Margin of error is then standard error times your Z-score.
36
00:01:49,003 --> 00:01:52,006
And the Z-score is the number of standard deviations
37
00:01:52,006 --> 00:01:55,008
that your confidence level is from the mean.
38
00:01:55,008 --> 00:01:59,004
A Z-score of one includes about 68% of values.
39
00:01:59,004 --> 00:02:04,004
But being 68% certain about something is not reliable.
40
00:02:04,004 --> 00:02:06,007
I recommend going for higher values.
41
00:02:06,007 --> 00:02:10,005
And I have those listed in the chart on this slide.
42
00:02:10,005 --> 00:02:12,006
Here we have some of the commonly used Z-scores.
43
00:02:12,006 --> 00:02:15,008
And you can see the confidence level and the Z-score
44
00:02:15,008 --> 00:02:18,006
that is associated with them.
45
00:02:18,006 --> 00:02:22,005
Most analysis in business operates at the 90%
46
00:02:22,005 --> 00:02:27,004
or 95% confidence levels so Z-scores of 1.645
47
00:02:27,004 --> 00:02:31,000
and 1.96 respectively.
48
00:02:31,000 --> 00:02:34,005
Getting enough samples to reach the 98%
49
00:02:34,005 --> 00:02:38,005
and 99% confidence levels can be expensive.
50
00:02:38,005 --> 00:02:41,007
If you can do it, great, but 90 and 95%
51
00:02:41,007 --> 00:02:43,008
are much more common.
52
00:02:43,008 --> 00:02:46,002
With all this in mind, let's switch over to Excel
53
00:02:46,002 --> 00:02:51,003
and calculate the margin of error for a data sample.
54
00:02:51,003 --> 00:02:54,001
I've switched over to Excel and my sample work
55
00:02:54,001 --> 00:02:57,002
because 01_05, Margin of Error.
56
00:02:57,002 --> 00:02:59,002
And you can find it in the chapter one folder
57
00:02:59,002 --> 00:03:01,006
of the exercise files collection.
58
00:03:01,006 --> 00:03:05,004
And I have set up the problem that I described earlier.
59
00:03:05,004 --> 00:03:07,006
So in cell B3, you see that we have
60
00:03:07,006 --> 00:03:10,006
a standard deviation of 0.1.
61
00:03:10,006 --> 00:03:14,004
And again, this is the standard deviation of olive oil
62
00:03:14,004 --> 00:03:16,005
put into bottles at a factory.
63
00:03:16,005 --> 00:03:19,002
So we have a standard deviation of 0.1,
64
00:03:19,002 --> 00:03:21,006
a certainty level of 95%.
65
00:03:21,006 --> 00:03:25,004
And then we can put in the Z-score for 95%.
66
00:03:25,004 --> 00:03:28,003
I have those listed over here on the right.
67
00:03:28,003 --> 00:03:31,006
So I see that 95% is 1.96.
68
00:03:31,006 --> 00:03:35,005
So I'll click back in cell B5, 1.96.
69
00:03:35,005 --> 00:03:37,003
I could create a formula to do a look up
70
00:03:37,003 --> 00:03:41,000
but I don't think that's necessary here.
71
00:03:41,000 --> 00:03:44,002
And finally, we have our number of measurements.
72
00:03:44,002 --> 00:03:47,007
For the standard error, I need to divide
73
00:03:47,007 --> 00:03:50,004
the standard deviation by the square root
74
00:03:50,004 --> 00:03:52,001
of the number of measurements.
75
00:03:52,001 --> 00:03:54,001
So in B8 I'll type equal.
76
00:03:54,001 --> 00:03:57,003
And our standard deviation is in B3.
77
00:03:57,003 --> 00:04:00,000
And I will divide that by the square root,
78
00:04:00,000 --> 00:04:04,004
and that function is SQRT, of the number of measurements,
79
00:04:04,004 --> 00:04:06,003
which is in B6.
80
00:04:06,003 --> 00:04:10,009
Right parentheses and Enter. And I get 0.016.
81
00:04:10,009 --> 00:04:15,001
Okay, now I can look at the margin of error.
82
00:04:15,001 --> 00:04:17,006
And that, as the formula on the right reminds us,
83
00:04:17,006 --> 00:04:22,001
is the Z-score, which is in B5,
84
00:04:22,001 --> 00:04:27,000
multiplied by the standard error in B8 and Enter.
85
00:04:27,000 --> 00:04:32,005
And we have a margin of error of 0.031, which is rounded
86
00:04:32,005 --> 00:04:37,000
to two digits, the 0.03 that we saw earlier in the movie.
6954
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.