abdullah's picture
Add files using upload-large-folder tool
50deadf verified
raw
history blame
45 kB
1
00:00:04,670 --> 00:00:08,210
Today, Inshallah, we are going to start Chapter 7.
2
00:00:09,830 --> 00:00:14,910
Chapter 7 talks about sampling and sampling
3
00:00:14,910 --> 00:00:22,690
distributions. The objectives for this chapter are
4
00:00:22,690 --> 00:00:27,610
number one, we have different methods, actually we
5
00:00:27,610 --> 00:00:31,330
have two methods: probability and non-probability
6
00:00:31,330 --> 00:00:34,750
samples, and we are going to distinguish between
7
00:00:35,420 --> 00:00:40,700
these two sampling methods. So again, in this
8
00:00:40,700 --> 00:00:44,980
chapter, we will talk about two different sampling
9
00:00:44,980 --> 00:00:49,480
methods. One is called probability sampling and
10
00:00:49,480 --> 00:00:52,940
the other is non-probability sampling. Our goal is
11
00:00:52,940 --> 00:00:56,520
to distinguish between these two different
12
00:00:56,520 --> 00:00:59,280
sampling methods. The other learning objective
13
00:00:59,280 --> 00:01:04,400
will be, We'll talk about the concept of the
14
00:01:04,400 --> 00:01:06,700
sampling distribution. That will be next time,
15
00:01:06,800 --> 00:01:09,960
inshallah. The third objective is compute
16
00:01:09,960 --> 00:01:15,480
probabilities related to sample mean. In addition
17
00:01:15,480 --> 00:01:18,160
to that, we'll talk about how can we compute
18
00:01:18,160 --> 00:01:22,920
probabilities regarding the sample proportion. And
19
00:01:22,920 --> 00:01:27,130
as I mentioned last time, There are two types of
20
00:01:27,130 --> 00:01:30,270
data. One is called the numerical data. In this
21
00:01:30,270 --> 00:01:33,470
case, we can use the sample mean. The other type
22
00:01:33,470 --> 00:01:36,630
is called qualitative data. And in this case, we
23
00:01:36,630 --> 00:01:39,330
have to use the sample proportion. So for this
24
00:01:39,330 --> 00:01:41,690
chapter, we are going to discuss how can we
25
00:01:41,690 --> 00:01:46,370
compute the probabilities for each one, either the
26
00:01:46,370 --> 00:01:50,090
sample mean or the sample proportion. The last
27
00:01:50,090 --> 00:01:55,770
objective of this chapter is to use the central
28
00:01:55,770 --> 00:01:58,190
limit theorem which is the famous one of the most
29
00:01:58,190 --> 00:02:02,130
famous theorem in this book which is called again
30
00:02:02,130 --> 00:02:05,690
CLT, central limit theorem, and we are going to show
31
00:02:05,690 --> 00:02:09,310
what are the, what is the importance of this
32
00:02:09,310 --> 00:02:11,930
theorem, so these are the mainly the four
33
00:02:11,930 --> 00:02:16,610
objectives for this chapter. Now let's see why we
34
00:02:16,610 --> 00:02:20,270
are talking about sampling. In other words, most
35
00:02:20,270 --> 00:02:23,850
of the time when we are doing study, we are using
36
00:02:23,850 --> 00:02:27,700
a sample instead of using the entire population.
37
00:02:28,640 --> 00:02:32,080
Now there are many reasons behind that. One of
38
00:02:32,080 --> 00:02:37,840
these reasons is selecting a sample is less time
39
00:02:37,840 --> 00:02:40,940
consuming than selecting every item in the
40
00:02:40,940 --> 00:02:44,060
population. I think it makes sense that suppose we
41
00:02:44,060 --> 00:02:46,560
have a huge population, that population consists
42
00:02:46,560 --> 00:02:53,140
of thousands of items. So that will take more time
43
00:02:54,440 --> 00:03:00,220
If you select 100 of their population. So time
44
00:03:00,220 --> 00:03:02,140
consuming is very important. So number one,
45
00:03:03,000 --> 00:03:05,780
selecting sample is less time consuming than using
46
00:03:05,780 --> 00:03:10,280
all the entire population. The second reason,
47
00:03:10,880 --> 00:03:14,640
selecting samples is less costly than selecting a
48
00:03:14,640 --> 00:03:17,280
variety of population. Because if we have large
49
00:03:17,280 --> 00:03:19,560
population, in this case you have to spend more
50
00:03:19,560 --> 00:03:23,540
money in order to get the data or the information
51
00:03:23,540 --> 00:03:27,940
from that population. So it's better to use these
52
00:03:27,940 --> 00:03:33,300
samples. The other reason is the analysis. Our
53
00:03:33,300 --> 00:03:37,260
sample is less cumbersome and more practical than
54
00:03:37,260 --> 00:03:40,880
analysis of all items in the population. For these
55
00:03:40,880 --> 00:03:45,820
reasons, we have to use a sample. For this reason,
56
00:03:45,880 --> 00:03:53,080
we have to talk about sampling methods. Let's
57
00:03:53,080 --> 00:03:58,540
start with sampling process. That begins with a
58
00:03:58,540 --> 00:04:05,320
sampling frame. Now suppose my goal is to know the
59
00:04:05,320 --> 00:04:13,960
opinion of IUG students about a certain subject.
60
00:04:16,260 --> 00:04:24,120
So my population consists of all IUG students. So
61
00:04:24,120 --> 00:04:27,370
that's the entire population. And you know that,
62
00:04:27,590 --> 00:04:31,750
for example, suppose our usual students is around,
63
00:04:32,430 --> 00:04:39,890
for example, 20,000 students. 20,000 students is a
64
00:04:39,890 --> 00:04:45,490
big number. So it's better to select a sample from
65
00:04:45,490 --> 00:04:49,270
that population. Now, the first step in this
66
00:04:49,270 --> 00:04:55,700
process, we have to determine the frame of that
67
00:04:55,700 --> 00:05:01,320
population. So my frame consists of all IU
68
00:05:01,320 --> 00:05:04,740
students, which has maybe males and females. So my
69
00:05:04,740 --> 00:05:09,560
frame in this case is all items, I mean all
70
00:05:09,560 --> 00:05:15,380
students at IUG. So that's the frame. So my frame
71
00:05:15,380 --> 00:05:18,720
consists
72
00:05:18,720 --> 00:05:22,220
of all students.
73
00:05:27,630 --> 00:05:32,370
So the definition of
74
00:05:32,370 --> 00:05:36,010
the sampling frame is a listing of items that make
75
00:05:36,010 --> 00:05:39,350
up the population. The items could be individual,
76
00:05:40,170 --> 00:05:44,490
could be students, could be things, animals, and
77
00:05:44,490 --> 00:05:49,650
so on. So frames are data sources such as a
78
00:05:49,650 --> 00:05:54,840
population list. Suppose we have the names of IUDs
79
00:05:54,840 --> 00:05:58,840
humans. So that's my population list. Or
80
00:05:58,840 --> 00:06:02,160
directories, or maps, and so on. So that's the
81
00:06:02,160 --> 00:06:05,520
frame, we have to know about the population we are
82
00:06:05,520 --> 00:06:10,900
interested in. Inaccurate or biased results can
83
00:06:10,900 --> 00:06:16,460
result if frame excludes certain portions of the
84
00:06:16,460 --> 00:06:20,620
population. For example, suppose here, as I
85
00:06:20,620 --> 00:06:24,180
mentioned, we are interested in IUG students, so
86
00:06:24,180 --> 00:06:29,280
my frame and all IU students. And I know there are
87
00:06:29,280 --> 00:06:35,900
students, either males or females. Suppose for
88
00:06:35,900 --> 00:06:40,880
some reasons, we ignore males, and just my sample
89
00:06:40,880 --> 00:06:45,080
focused on females. In this case, females.
90
00:06:48,700 --> 00:06:51,900
don't represent the entire population. For this
91
00:06:51,900 --> 00:06:57,720
reason, you will get inaccurate or biased results
92
00:06:57,720 --> 00:07:02,000
if you ignore a certain portion. Because here
93
00:07:02,000 --> 00:07:08,580
males, for example, maybe consists of 40% of the
94
00:07:08,580 --> 00:07:12,960
IG students. So it makes sense that this number or
95
00:07:12,960 --> 00:07:16,980
this percentage is a big number. So ignoring this
96
00:07:16,980 --> 00:07:21,160
portion, may lead to misleading results or
97
00:07:21,160 --> 00:07:26,160
inaccurate results or biased results. So you have
98
00:07:26,160 --> 00:07:29,600
to keep in mind that you have to choose all the
99
00:07:29,600 --> 00:07:33,740
portions of that frame. So inaccurate or biased
100
00:07:33,740 --> 00:07:38,700
results can result if a frame excludes certain
101
00:07:38,700 --> 00:07:43,180
portions of a population. Another example, suppose
102
00:07:43,180 --> 00:07:48,680
we took males and females. But here for females,
103
00:07:49,240 --> 00:07:56,020
females have, for example, four levels: Level one,
104
00:07:56,400 --> 00:07:59,980
level two, level three, and level four. And we
105
00:07:59,980 --> 00:08:05,560
ignored, for example, level one. I mean, the new
106
00:08:05,560 --> 00:08:09,520
students. We ignored this portion. Maybe this
107
00:08:09,520 --> 00:08:12,860
portion is very important one, but by mistake we
108
00:08:12,860 --> 00:08:18,690
ignored this one. The remaining three levels will
109
00:08:18,690 --> 00:08:22,430
not represent the entire female population. For
110
00:08:22,430 --> 00:08:25,330
this reason, you will get inaccurate or biased
111
00:08:25,330 --> 00:08:31,290
results. So you have to select all the portions of
112
00:08:31,290 --> 00:08:36,610
the frames. Using different frames to generate
113
00:08:36,610 --> 00:08:40,110
data can lead to dissimilar conclusions. For
114
00:08:40,110 --> 00:08:46,020
example, Suppose again I am interested in IEG
115
00:08:46,020 --> 00:08:46,720
students.
116
00:08:49,440 --> 00:08:59,460
And I took the frame that has all students at
117
00:08:59,460 --> 00:09:04,060
the University of Gaza, the Universities of Gaza.
118
00:09:09,250 --> 00:09:12,110
And as we know that Gaza has three universities,
119
00:09:12,350 --> 00:09:15,530
big universities: Islamic University, Lazar
120
00:09:15,530 --> 00:09:18,030
University, and Al-Aqsa University. So we have
121
00:09:18,030 --> 00:09:23,310
three universities. And my frame here, suppose I
122
00:09:23,310 --> 00:09:27,410
took all students at these universities, but my
123
00:09:27,410 --> 00:09:32,470
study focused on IU students. So my frame, the
124
00:09:32,470 --> 00:09:38,250
true one, is all students at IUG. But I taught all
125
00:09:38,250 --> 00:09:42,170
students at universities in Gaza. So now we have
126
00:09:42,170 --> 00:09:44,690
different frames.
127
00:09:48,610 --> 00:09:54,590
And you want to know what are the opinions of the
128
00:09:54,590 --> 00:09:59,910
smokers about smoking. So my population now is
129
00:09:59,910 --> 00:10:00,530
just...
130
00:10:14,030 --> 00:10:19,390
So that's my thing.
131
00:10:21,010 --> 00:10:32,410
I suppose I talk to a field that has one atom.
132
00:10:40,780 --> 00:10:46,040
Oh my goodness. They are very different things.
133
00:10:47,700 --> 00:10:53,720
The first one consists of only smokers. They are
134
00:10:53,720 --> 00:10:58,100
very interested in you. The other one consists
135
00:10:58,100 --> 00:11:06,560
of... anonymous. I thought maybe... smoker or non
136
00:11:06,560 --> 00:11:10,460
-smokers. For this reason, you will get...
137
00:11:17,410 --> 00:11:19,350
Conclusion, different results.
138
00:11:22,090 --> 00:11:28,850
So now,
139
00:11:29,190 --> 00:11:33,610
the sampling frame is a listing of items that make
140
00:11:33,610 --> 00:11:39,510
up the entire population. Let's move to the types
141
00:11:39,510 --> 00:11:44,910
of samples. Mainly there are two types of
142
00:11:44,910 --> 00:11:49,070
sampling: One is called non-probability samples.
143
00:11:50,370 --> 00:11:54,650
The other one is called probability samples. The
144
00:11:54,650 --> 00:11:59,790
non-probability samples can be divided into two
145
00:11:59,790 --> 00:12:04,030
segments: One is called judgment, and the other
146
00:12:04,030 --> 00:12:08,710
convenience. So we have judgment and convenience
147
00:12:08,710 --> 00:12:13,140
non-probability samples. The other type which is
148
00:12:13,140 --> 00:12:17,560
random probability samples, has four segments or
149
00:12:17,560 --> 00:12:21,680
four parts: The first one is called simple random
150
00:12:21,680 --> 00:12:25,680
sample. The other one is systematic. The second
151
00:12:25,680 --> 00:12:28,680
one is systematic random sample. The third one is
152
00:12:28,680 --> 00:12:32,940
stratified. The fourth one, cluster random sample.
153
00:12:33,460 --> 00:12:37,770
So there are two types of sampling: probability
154
00:12:37,770 --> 00:12:41,490
and non-probability. Non-probability has four
155
00:12:41,490 --> 00:12:45,350
methods here: simple random samples, systematic,
156
00:12:45,530 --> 00:12:48,530
stratified, and cluster. And the non-probability
157
00:12:48,530 --> 00:12:53,090
samples has two types: judgment and convenience.
158
00:12:53,670 --> 00:12:58,490
Let's see the definition of each type of samples.
159
00:12:59,190 --> 00:13:03,720
Let's start with non-probability sample. In non
160
00:13:03,720 --> 00:13:07,000
-probability sample, items included or chosen
161
00:13:07,000 --> 00:13:10,800
without regard to their probability of occurrence.
162
00:13:11,760 --> 00:13:14,740
So that's the definition of non-probability. For
163
00:13:14,740 --> 00:13:15,100
example.
164
00:13:23,660 --> 00:13:26,480
So again, non-probability sample, it means you
165
00:13:26,480 --> 00:13:29,580
select items without regard to their probability
166
00:13:29,580 --> 00:13:34,030
of occurrence. For example, suppose females
167
00:13:34,030 --> 00:13:42,430
consist of 70% of IUG students and males, the
168
00:13:42,430 --> 00:13:49,930
remaining percent is 30%. And suppose I decided to
169
00:13:49,930 --> 00:13:56,610
select a sample of 100 or 1000 students from IUG.
170
00:13:58,620 --> 00:14:07,980
Suddenly, I have a sample that has 650 males and
171
00:14:07,980 --> 00:14:14,780
350 females. Now, this sample, which has these
172
00:14:14,780 --> 00:14:19,260
numbers, for sure does not represent the entire
173
00:14:19,260 --> 00:14:25,240
population. Because females has 70%, and I took a
174
00:14:25,240 --> 00:14:30,890
random sample or a sample of size 350. So this
175
00:14:30,890 --> 00:14:35,830
sample is chosen without regard to the probability
176
00:14:35,830 --> 00:14:40,370
here. Because in this case, I should choose males
177
00:14:40,370 --> 00:14:44,110
with respect to their probability, which is 30%.
178
00:14:44,110 --> 00:14:49,330
But in this case, I just choose different
179
00:14:49,330 --> 00:14:54,990
proportions. Another example. Suppose
180
00:14:57,260 --> 00:14:59,920
again, I am talking about smoking.
181
00:15:05,080 --> 00:15:10,120
And I know that some people are smoking and I just
182
00:15:10,120 --> 00:15:14,040
took this sample. So I took this sample based on
183
00:15:14,040 --> 00:15:18,600
my knowledge. So it's without regard to their
184
00:15:18,600 --> 00:15:23,340
probability. Maybe suppose I am talking about
185
00:15:23,340 --> 00:15:28,330
political opinions about something. And I just
186
00:15:28,330 --> 00:15:36,330
took the experts of that subject. So my sample is
187
00:15:36,330 --> 00:15:42,070
not a probability sample. And this one has, as we
188
00:15:42,070 --> 00:15:44,230
mentioned, has two types: One is called
189
00:15:44,230 --> 00:15:49,010
convenience sampling. In this case, items are
190
00:15:49,010 --> 00:15:51,710
selected based only on the fact that they are
191
00:15:51,710 --> 00:15:55,590
easy. So I choose that sample because it's easy.
192
00:15:57,090 --> 00:15:57,690
Inexpensive,
193
00:16:02,190 --> 00:16:09,790
inexpensive, or convenient to sample. If I choose
194
00:16:09,790 --> 00:16:13,430
my sample because it is easy or inexpensive, I
195
00:16:13,430 --> 00:16:18,480
think it doesn't make any sense, because easy is
196
00:16:18,480 --> 00:16:23,780
not a reason to select that sample
223
00:18:17,050 --> 00:18:20,970
segment and so on. But the convenient sample means
224
00:18:20,970 --> 00:18:24,690
that you select a sample maybe that is easy for
225
00:18:24,690 --> 00:18:29,430
you, or less expensive, or that sample is
226
00:18:29,430 --> 00:18:32,980
convenient. For this reason, it's called non
227
00:18:32,980 --> 00:18:36,300
-probability sample because we choose that sample
228
00:18:36,300 --> 00:18:39,540
without regard to their probability of occurrence.
229
00:18:41,080 --> 00:18:48,620
The other type is called probability samples. In
230
00:18:48,620 --> 00:18:54,200
this case, items are chosen on the basis of non
231
00:18:54,200 --> 00:18:58,600
-probabilities. For example, here, if males
232
00:19:02,500 --> 00:19:11,060
has or represent 30%, and females represent 70%,
233
00:19:11,060 --> 00:19:14,840
and the same size has a thousand. So in this case,
234
00:19:14,920 --> 00:19:19,340
you have to choose females with respect to their
235
00:19:19,340 --> 00:19:24,260
probability. Now 70% for females, so I have to
236
00:19:24,260 --> 00:19:29,430
choose 700 for females and the remaining 300 for
237
00:19:29,430 --> 00:19:34,010
males. So in this case, I choose the items, I mean
238
00:19:34,010 --> 00:19:37,970
I choose my samples regarding to their
239
00:19:37,970 --> 00:19:39,050
probability.
240
00:19:41,010 --> 00:19:45,190
So in probability sample items and the sample are
241
00:19:45,190 --> 00:19:48,610
chosen on the basis of known probabilities. And
242
00:19:48,610 --> 00:19:52,360
again, there are two types. of probability
243
00:19:52,360 --> 00:19:55,580
samples, simple random sample, systematic,
244
00:19:56,120 --> 00:19:59,660
stratified, and cluster. Let's talk about each one
245
00:19:59,660 --> 00:20:05,040
in details. The first type is called a probability
246
00:20:05,040 --> 00:20:11,720
sample. Simple random sample. The first type of
247
00:20:11,720 --> 00:20:16,200
probability sample is the easiest one. Simple
248
00:20:16,200 --> 00:20:23,780
random sample. Generally is denoted by SRS, Simple
249
00:20:23,780 --> 00:20:30,660
Random Sample. Let's see how can we choose a
250
00:20:30,660 --> 00:20:35,120
sample that is random. What do you mean by random?
251
00:20:36,020 --> 00:20:41,780
In this case, every individual or item from the
252
00:20:41,780 --> 00:20:47,620
frame has an equal chance of being selected. For
253
00:20:47,620 --> 00:20:52,530
example, suppose number of students in this class
254
00:20:52,530 --> 00:21:04,010
number of students is 52 so
255
00:21:04,010 --> 00:21:11,890
each one, I mean each student from
256
00:21:11,890 --> 00:21:17,380
1 up to 52 has the same probability of being
257
00:21:17,380 --> 00:21:23,860
selected. 1 by 52. 1 by 52. 1 divided by 52. So
258
00:21:23,860 --> 00:21:27,980
each one has this probability. So the first one
259
00:21:27,980 --> 00:21:31,820
has the same because if I want to select for
260
00:21:31,820 --> 00:21:37,680
example 10 out of you. So the first one has each
261
00:21:37,680 --> 00:21:42,400
one has probability of 1 out of 52. That's the
262
00:21:42,400 --> 00:21:47,160
meaning of Each item from the frame has an equal
263
00:21:47,160 --> 00:21:54,800
chance of being selected. Selection may be with
264
00:21:54,800 --> 00:21:58,800
replacement. With replacement means selected
265
00:21:58,800 --> 00:22:02,040
individuals is returned to the frame for
266
00:22:02,040 --> 00:22:04,880
possibility selection, or without replacement
267
00:22:04,880 --> 00:22:08,600
means selected individuals or item is not returned
268
00:22:08,600 --> 00:22:10,820
to the frame. So we have two types of selection,
269
00:22:11,000 --> 00:22:14,360
either with... So with replacement means item is
270
00:22:14,360 --> 00:22:18,080
returned back to the frame, or without population,
271
00:22:18,320 --> 00:22:21,400
the item is not returned back to the frame. So
272
00:22:21,400 --> 00:22:26,490
that's the two types of selection. Now how can we
273
00:22:26,490 --> 00:22:29,810
obtain the sample? Sample obtained from something
274
00:22:29,810 --> 00:22:33,470
called table of random numbers. In a minute I will
275
00:22:33,470 --> 00:22:36,430
show you the table of random numbers. And other
276
00:22:36,430 --> 00:22:40,130
method of selecting a sample by using computer
277
00:22:40,130 --> 00:22:44,890
random number generators. So there are two methods
278
00:22:44,890 --> 00:22:48,310
for selecting a random number. Either by using the
279
00:22:48,310 --> 00:22:51,950
table that you have at the end of your book or by
280
00:22:51,950 --> 00:22:56,550
using a computer. I will show one of these and in
281
00:22:56,550 --> 00:22:59,650
the SPSS course you will see another one which is
282
00:22:59,650 --> 00:23:03,690
by using a computer. So let's see how can we
283
00:23:03,690 --> 00:23:11,730
obtain a sample from table of
284
00:23:11,730 --> 00:23:12,590
random number.
285
00:23:16,950 --> 00:23:22,090
I have maybe different table here. But the same
286
00:23:22,090 --> 00:23:28,090
idea to use that table. Let's see how can we
287
00:23:28,090 --> 00:23:34,990
choose a sample by using a random number.
288
00:23:42,490 --> 00:23:47,370
Now, for example, suppose in this class As I
289
00:23:47,370 --> 00:23:51,090
mentioned, there are 52 students.
290
00:23:55,110 --> 00:23:58,650
So each one has a number, ID number one, two, up
291
00:23:58,650 --> 00:24:05,110
to 52. So the numbers are 01, 02, all the way up
292
00:24:05,110 --> 00:24:10,790
to 52. So the maximum digits here, two, two
293
00:24:10,790 --> 00:24:11,110
digits.
294
00:24:15,150 --> 00:24:18,330
1, 2, 3, up to 5, 2, 2, so you have two digits.
295
00:24:19,470 --> 00:24:23,710
Now suppose I decided to take a random sample of
296
00:24:23,710 --> 00:24:28,550
size, for example, N instead. How can I select N
297
00:24:28,550 --> 00:24:32,570
out of U? In this case, each one has the same
298
00:24:32,570 --> 00:24:36,790
chance of being selected. Now based on this table,
299
00:24:37,190 --> 00:24:44,230
you can pick any row or any column. Randomly. For
300
00:24:44,230 --> 00:24:51,630
example, suppose I select the first row. Now, the
301
00:24:51,630 --> 00:24:56,570
first student will be selected as student number
302
00:24:56,570 --> 00:25:03,650
to take two digits. We have to take how many
303
00:25:03,650 --> 00:25:08,770
digits? Because students have ID card that
304
00:25:08,770 --> 00:25:13,930
consists of two digits, 0102 up to 52. So, what's
305
00:25:13,930 --> 00:25:17,010
the first number students will be selected based
306
00:25:17,010 --> 00:25:22,130
on this table? Forget about the line 101.
307
00:25:26,270 --> 00:25:27,770
Start with this number.
308
00:25:42,100 --> 00:25:50,900
So the first one, 19. The second, 22. The third
309
00:25:50,900 --> 00:25:51,360
student,
310
00:25:54,960 --> 00:26:04,000
19, 22. The third, 9. The third, 9. I'm taking the
311
00:26:04,000 --> 00:26:16,510
first row. Then fifth. 34 student
312
00:26:16,510 --> 00:26:18,710
number 05
313
00:26:24,340 --> 00:26:29,500
Now, what's about seventy-five? Seventy-five is
314
00:26:29,500 --> 00:26:33,660
not selected because the maximum I have is fifty
315
00:26:33,660 --> 00:26:46,180
-two. Next. Sixty-two is not selected. Eighty
316
00:26:46,180 --> 00:26:53,000
-seven. It's not selected. 13. 13. It's okay.
317
00:26:53,420 --> 00:27:01,740
Next. 96. 96. Not selected. 14. 14 is okay. 91.
318
00:27:02,140 --> 00:27:12,080
91. 91. Not selected. 95. 91. 45. 85. 31. 31.
319
00:27:15,240 --> 00:27:21,900
So that's 10. So students numbers are 19, 22, 39,
320
00:27:22,140 --> 00:27:26,980
50, 34, 5, 13, 4, 25 and take one will be
321
00:27:26,980 --> 00:27:30,940
selected. So these are the ID numbers will be
322
00:27:30,940 --> 00:27:35,480
selected in order to get a sample of 10. You
323
00:27:35,480 --> 00:27:40,500
exclude
324
00:27:40,500 --> 00:27:43,440
that one. If the number is repeated, you have to
325
00:27:43,440 --> 00:27:44,340
exclude that one.
326
00:27:51,370 --> 00:27:57,270
is repeated, then excluded.
327
00:28:02,370 --> 00:28:07,370
So the returned number must be excluded from the
328
00:28:07,370 --> 00:28:14,030
sample. Let's imagine that we have not 52
329
00:28:14,030 --> 00:28:19,130
students. We have 520 students.
330
00:28:25,740 --> 00:28:32,520
Now, I have large number, 52, 520 instead of 52
331
00:28:32,520 --> 00:28:36,080
students. And again, my goal is to select just 10
332
00:28:36,080 --> 00:28:42,220
students out of 120. So each one has ID with
333
00:28:42,220 --> 00:28:46,220
number one, two, all the way up to 520. So the
334
00:28:46,220 --> 00:28:53,160
first one, 001. 002 all the way up to 520 now in
335
00:28:53,160 --> 00:28:56,480
this case you have to choose three digits start
336
00:28:56,480 --> 00:29:00,060
for example you don't have actually to start with
337
00:29:00,060 --> 00:29:03,060
row number one maybe column number one or row
338
00:29:03,060 --> 00:29:06,140
number two whatever is fine so let's start with
339
00:29:06,140 --> 00:29:10,460
row number two for example row number 76
340
00:29:14,870 --> 00:29:19,950
It's not selected. Because the maximum number I
341
00:29:19,950 --> 00:29:25,110
have is 5 to 20. So, 746 shouldn't be selected.
342
00:29:26,130 --> 00:29:29,430
The next one, 764.
343
00:29:31,770 --> 00:29:38,750
Again, it's not selected. 764, 715. Not selected.
344
00:29:38,910 --> 00:29:42,310
Next one is 715.
345
00:29:44,880 --> 00:29:52,200
099 should be 0 that's
346
00:29:52,200 --> 00:29:54,940
the way how can we use the random table for using
347
00:29:54,940 --> 00:29:58,800
or for selecting simple random symbols so in this
348
00:29:58,800 --> 00:30:03,480
case you can choose any row or any column then you
349
00:30:03,480 --> 00:30:06,620
have to decide how many digits you have to select
350
00:30:06,620 --> 00:30:10,500
it depends on the number you have I mean the
351
00:30:10,500 --> 00:30:16,510
population size If for example Suppose I am
352
00:30:16,510 --> 00:30:20,270
talking about IUPUI students and for example, we
353
00:30:20,270 --> 00:30:26,530
have 30,000 students at this school And again, I
354
00:30:26,530 --> 00:30:28,570
want to select a random sample of size 10 for
355
00:30:28,570 --> 00:30:35,190
example So how many digits should I use? 20,000
356
00:30:35,190 --> 00:30:42,620
Five digits And each one, each student has ID
357
00:30:42,620 --> 00:30:51,760
from, starts from the first one up to twenty
358
00:30:51,760 --> 00:30:56,680
thousand. So now, start with, for example, the
359
00:30:56,680 --> 00:30:59,240
last row you have.
360
00:31:03,120 --> 00:31:08,480
The first number 54000 is not. 81 is not. None of
361
00:31:08,480 --> 00:31:08,740
these.
362
00:31:12,420 --> 00:31:17,760
Look at the next one. 71000 is not selected. Now
363
00:31:17,760 --> 00:31:22,180
9001. So the first number I have to select is
364
00:31:22,180 --> 00:31:27,200
9001. None of the rest. Go back.
365
00:31:30,180 --> 00:31:37,790
Go to the next one. The second number, 12149
366
00:31:37,790 --> 00:31:45,790
and so on. Next will be 18000 and so on. Next row,
367
00:31:46,470 --> 00:31:55,530
we can select the second one, then 16, then 14000,
368
00:31:55,890 --> 00:32:00,850
6500 and so on. So this is the way how can we use
369
00:32:00,850 --> 00:32:08,110
the random table. It seems to be that tons of work
370
00:32:08,110 --> 00:32:13,450
if you have large sample. Because in this case,
371
00:32:13,530 --> 00:32:16,430
you have to choose, for example, suppose I am
372
00:32:16,430 --> 00:32:22,390
interested to take a random sample of 10,000. Now,
373
00:32:22,510 --> 00:32:28,370
to use this table to select 10,000 items takes
374
00:32:28,370 --> 00:32:33,030
time and effort and maybe will never finish. So
375
00:32:33,030 --> 00:32:33,950
it's better to use
376
00:32:38,020 --> 00:32:42,100
better to use computer
377
00:32:42,100 --> 00:32:47,140
random number generators. So that's the way if we,
378
00:32:47,580 --> 00:32:51,880
now we can use the random table only if the sample
379
00:32:51,880 --> 00:32:57,780
size is limited. I mean up to 100 maybe you can
380
00:32:57,780 --> 00:33:03,160
use the random table, but after that I think it's
381
00:33:03,160 --> 00:33:08,670
just you are losing your time. Another example
382
00:33:08,670 --> 00:33:14,390
here. Now suppose my sampling frame for population
383
00:33:14,390 --> 00:33:23,230
has 850 students. So the numbers are 001, 002, all
384
00:33:23,230 --> 00:33:28,490
the way up to 850. And suppose for example we are
385
00:33:28,490 --> 00:33:33,610
going to select five items randomly from that
386
00:33:33,610 --> 00:33:39,610
population. So you have to choose three digits and
387
00:33:39,610 --> 00:33:44,990
imagine that this is my portion of that table.
388
00:33:45,850 --> 00:33:51,570
Now, take three digits. The first three digits are
389
00:33:51,570 --> 00:34:00,330
492. So the first item chosen should be item
390
00:34:00,330 --> 00:34:10,540
number 492. should be selected next one 800 808
391
00:34:10,540 --> 00:34:17,020
doesn't select because the maximum it's much
392
00:34:17,020 --> 00:34:21,100
selected because the maximum here is 850 now next
393
00:34:21,100 --> 00:34:26,360
one 892 this
394
00:34:26,360 --> 00:34:32,140
one is not selected next
395
00:34:32,140 --> 00:34:43,030
item four three five selected now
396
00:34:43,030 --> 00:34:50,710
seven seven nine should be selected finally zeros
397
00:34:50,710 --> 00:34:53,130
two should be selected so these are the five
398
00:34:53,130 --> 00:34:58,090
numbers in my sample by using selected by using
399
00:34:58,090 --> 00:35:01,190
the random sample any questions?
400
00:35:04,160 --> 00:35:07,780
Let's move to another part.
401
00:35:17,600 --> 00:35:22,380
The next type of samples is called systematic
402
00:35:22,380 --> 00:35:25,260
samples.
403
00:35:29,120 --> 00:35:35,780
Now suppose N represents the sample size, capital
404
00:35:35,780 --> 00:35:40,520
N represents
405
00:35:40,520 --> 00:35:42,220
the population size.
406
00:35:46,660 --> 00:35:49,900
And let's see how can we choose a systematic
407
00:35:49,900 --> 00:35:54,040
random sample from that population. For example,
408
00:35:55,260 --> 00:35:57,180
suppose
409
00:3
445
00:39:27,800 --> 00:39:31,780
15, 25, 35, and so on if we have more than that.
446
00:39:33,230 --> 00:39:37,730
Okay, so that's for, in this example, he chose
447
00:39:37,730 --> 00:39:42,790
item number seven. Random selection, number seven.
448
00:39:43,230 --> 00:39:50,010
So next should be 17, 27, 37, and so on. Let's do
449
00:39:50,010 --> 00:39:50,710
another example.
450
00:39:58,590 --> 00:40:06,540
Suppose there are In this class, there are 50
451
00:40:06,540 --> 00:40:12,400
students. So the total is 50.
452
00:40:15,320 --> 00:40:26,780
10 students out of 50. So my sample is 10. Now
453
00:40:26,780 --> 00:40:30,260
still, 50 divided by 10 is 50.
454
00:40:33,630 --> 00:40:39,650
So there are five items or five students in a
455
00:40:39,650 --> 00:40:45,370
group. So we have five in
456
00:40:45,370 --> 00:40:51,490
the first group and then five in the next one and
457
00:40:51,490 --> 00:40:56,130
so on. So we have how many groups? Ten groups.
458
00:40:59,530 --> 00:41:04,330
So first step, you have to find a step. Still it
459
00:41:04,330 --> 00:41:07,930
means number of items or number of students in a
460
00:41:07,930 --> 00:41:16,170
group. Next step, select student at random from
461
00:41:16,170 --> 00:41:22,010
the first group, so random selection. Now, here
462
00:41:22,010 --> 00:41:28,610
there are five students, so 01, I'm sorry, not 01,
463
00:41:29,150 --> 00:41:35,080
1, 2, 3, 4, 5, so one digit. Only one digit.
464
00:41:35,800 --> 00:41:39,420
Because I have maximum number is five. So it's
465
00:41:39,420 --> 00:41:42,920
only one digit. So go again to the random table
466
00:41:42,920 --> 00:41:48,220
and take one digit. One. So my first item, six,
467
00:41:48,760 --> 00:41:52,580
eleven, sixteen, twenty-one, twenty-one, all the
468
00:41:52,580 --> 00:41:55,500
way up to ten items.
469
00:42:13,130 --> 00:42:18,170
So I choose student number one, then skip five,
470
00:42:19,050 --> 00:42:22,230
choose number six, and so on. It's called
471
00:42:22,230 --> 00:42:26,130
systematic. Because if you know the first item,
472
00:42:28,550 --> 00:42:32,690
and the step you can know the rest of these.
473
00:42:37,310 --> 00:42:41,150
Imagine that you want to select 10 students who
474
00:42:41,150 --> 00:42:48,010
entered the cafe shop or restaurant. You can pick
475
00:42:48,010 --> 00:42:54,790
one of them. So suppose I'm taking number three
476
00:42:54,790 --> 00:43:00,550
and my step is six. So three, then nine, and so
477
00:43:00,550 --> 00:43:00,790
on.
478
00:43:05,830 --> 00:43:13,310
So that's systematic assembly. Questions? So
479
00:43:13,310 --> 00:43:20,710
that's about random samples and systematic. What
480
00:43:20,710 --> 00:43:23,550
do you mean by stratified groups?
481
00:43:28,000 --> 00:43:33,080
Let's use a definition and an example of a
482
00:43:33,080 --> 00:43:34,120
stratified family.
483
00:43:58,810 --> 00:44:05,790
step one. So again imagine we have IUG population
484
00:44:05,790 --> 00:44:11,490
into two or more subgroups. So there are two or
485
00:44:11,490 --> 00:44:16,010
more. It depends on the characteristic you are
486
00:44:16,010 --> 00:44:19,690
using. So divide population into two or more
487
00:44:19,690 --> 00:44:24,210
subgroups according to some common characteristic.
488
00:44:24,730 --> 00:44:30,280
For example suppose I want to divide the student
489
00:44:30,280 --> 00:44:32,080
into gender.
490
00:44:34,100 --> 00:44:38,840
So males or females. So I have two strata. One is
491
00:44:38,840 --> 00:44:43,000
called males and the other is females. Now suppose
492
00:44:43,000 --> 00:44:47,460
the characteristic I am going to use is the levels
493
00:44:47,460 --> 00:44:51,500
of a student. First level, second, third, fourth,
494
00:44:51,800 --> 00:44:56,280
and so on. So number of strata here depends on
495
00:44:56,280 --> 00:45:00,380
actually the characteristic you are interested in.
496
00:45:00,780 --> 00:45:04,860
Let's use the simple one that is gender. So here
497
00:45:04,860 --> 00:45:12,360
we have females. So IUV students divided into two
498
00:45:12,360 --> 00:45:18,560
types, strata, or two groups, females and males.
499
00:45:19,200 --> 00:45:22,870
So this is the first step. So at least you should
500
00:45:22,870 --> 00:45:26,750
have two groups or two subgroups. So we have IELTS
501
00:45:26,750 --> 00:45:29,630
student, the entire population, and that
502
00:45:29,630 --> 00:45:34,370
population divided into two subgroups. Next,
503
00:45:35,650 --> 00:45:39,730
assemble random samples. Keep careful here with
504
00:45:39,730 --> 00:45:45,770
sample sizes proportional to strata sizes. That
505
00:45:45,770 --> 00:45:57,890
means suppose I know that Female consists
506
00:45:57,890 --> 00:46:02,470
of
507
00:46:02,470 --> 00:46:09,770
70% of Irish students and
508
00:46:09,770 --> 00:46:11,490
males 30%.
509
00:46:15,410 --> 00:46:17,950
the sample size we are talking about here is for
510
00:46:17,950 --> 00:46:21,550
example is a thousand so I want to select a sample
511
00:46:21,550 --> 00:46:24,990
of a thousand seed from the registration office or
512
00:46:24,990 --> 00:46:31,190
my information about that is males represent 30%
513
00:46:31,190 --> 00:46:37,650
females represent 70% so in this case your sample
514
00:46:37,650 --> 00:46:43,650
structure should be 70% times
515
00:46:50,090 --> 00:46:59,090
So the first
516
00:46:59,090 --> 00:47:03,750
group should have 700 items of students and the
517
00:47:03,750 --> 00:47:06,490
other one is 300,000.
518
00:47:09,230 --> 00:47:11,650
So this is the second step.
519
00:47:14,420 --> 00:47:17,740
Sample sizes are determined in step number two.
520
00:47:18,540 --> 00:47:22,200
Now, how can you select the 700 females here?
521
00:47:23,660 --> 00:47:26,180
Again, you have to go back to the random table.
522
00:47:27,480 --> 00:47:31,660
Samples from subgroups are compiled into one. Then
523
00:47:31,660 --> 00:47:39,600
you can use symbol random sample. So here, 700. I
524
00:47:39,600 --> 00:47:45,190
have, for example, 70% females. And I know that I
525
00:47:45,190 --> 00:47:51,370
use student help. I have ideas numbers from 1 up
526
00:47:51,370 --> 00:47:59,070
to 7, 14. Then by using simple random, simple
527
00:47:59,070 --> 00:48:01,070
random table, you can.
528
00:48:09,490 --> 00:48:15,190
So if you go back to the table, the first item,
529
00:48:16,650 --> 00:48:23,130
now look at five digits. Nineteen is not selected.
530
00:48:24,830 --> 00:48:27,510
Nineteen. I have, the maximum is fourteen
531
00:48:27,510 --> 00:48:31,890
thousand. So skip one and two. The first item is
532
00:48:31,890 --> 00:48:37,850
seven hundred and fifty-six. The first item. Next
533
00:48:37,850 --> 00:48:43,480
is not chosen. Next is not chosen. Number six.
534
00:48:43,740 --> 00:48:44,580
Twelve.
535
00:48:47,420 --> 00:48:50,620
Zero. Unsure.
536
00:48:52,880 --> 00:48:58,940
So here we divide the population into two groups
537
00:48:58,940 --> 00:49:03,440
or two subgroups, females and males. And we select
538
00:49:03,440 --> 00:49:07,020
a random sample of size 700 based on the
539
00:49:07,020 --> 00:49:10,850
proportion of this subgroup. Then we are using the
540
00:49:10,850 --> 00:49:16,750
simple random table to take the 700 females.
541
00:49:22,090 --> 00:49:29,810
Now for this example, there are 16 items or 16
542
00:49:29,810 --> 00:49:35,030
students in each group. And he select randomly
543
00:49:35,030 --> 00:49:40,700
number three, number 9, number 13, and so on. So
544
00:49:40,700 --> 00:49:44,140
it's a random selection. Another example.
545
00:49:46,820 --> 00:49:52,420
Suppose again we are talking about all IUVs.
546
00:50:02,780 --> 00:50:09,360
Here I divided the population according to the
547
00:50:09,360 --> 00:50:17,680
students' levels. Level one, level two, three
548
00:50:17,680 --> 00:50:18,240
levels.
549
00:50:25,960 --> 00:50:28,300
One, two, three and four.
550
00:50:32,240 --> 00:50:39,710
So I divide the population into four subgroups
551
00:50:39,710 --> 00:50:43,170
according to the student levels. So one, two,
552
00:50:43,290 --> 00:50:48,030
three, and four. Now, a simple random sample is
553
00:50:48,030 --> 00:50:52,070
selected from each subgroup with sample sizes
554
00:50:52,070 --> 00:50:57,670
proportional to strata size. Imagine that level
555
00:50:57,670 --> 00:51:04,950
number one represents 40% of the students. Level
556
00:51:04,950 --> 00:51:17,630
2, 20%. Level 3, 30%. Just
557
00:51:17,630 --> 00:51:22,850
an example. To make more sense?
558
00:51:34,990 --> 00:51:36,070
My sample size?
559
00:51:38,750 --> 00:51:39,910
3,
560
00:51:41,910 --> 00:51:46,430
9, 15, 4, sorry.
561
00:51:53,290 --> 00:52:00,470
So here, there are four levels. And the
562
00:52:00,470 --> 00:52:04,370
proportions are 48
563
00:52:06,670 --> 00:52:17,190
sample size is 500 so the sample for each strata
564
00:52:17,190 --> 00:52:31,190
will be number 1 40% times 500 gives 200 the next
565
00:52:31,190 --> 00:52:32,950
150
566
00:52:36,200 --> 00:52:42,380
And so on. Now, how can we choose the 200 from
567
00:52:42,380 --> 00:52:46,280
level number one? Again, we have to choose the
568
00:52:46,280 --> 00:52:55,540
random table. Now, 40% from this number, it means
569
00:52:55,540 --> 00:52:59,620
5
570
00:52:59,620 --> 00:53:06,400
,000. This one has 5,000. 600 females students.
571
00:53:07,720 --> 00:53:13,480
Because 40% of females in level 1. And I know that
572
00:53:13,480 --> 00:53:17,780
the total number of females is 14,000. So number
573
00:53:17,780 --> 00:53:23,420
of females in the first level is 5600. How many
574
00:53:23,420 --> 00:53:28,040
digits we have? Four digits. The first one, 001,
575
00:53:28,160 --> 00:53:34,460
all the way up to 560. If you go back, into a
576
00:53:34,460 --> 00:53:39,520
random table, take five, four digits. So the first
577
00:53:39,520 --> 00:53:43,340
number is 1922.
578
00:53:43,980 --> 00:53:48,000
Next is 3950.
579
00:53:50,140 --> 00:53:54,760
And so on. So that's the way how can we choose
580
00:53:54,760 --> 00:53:58,640
stratified samples.
581
00:54:02,360 --> 00:54:08,240
Next, the last one is called clusters. And let's
582
00:54:08,240 --> 00:54:11,400
see now what's the difference between stratified
583
00:54:11,400 --> 00:54:16,500
and cluster. Step one.
584
00:54:25,300 --> 00:54:31,720
Population is divided into some clusters.
585
00:54:35,000 --> 00:54:41,160
Step two, assemble one by assembling clusters
586
00:54:41,160 --> 00:54:42,740
selective.
587
00:54:46,100 --> 00:54:48,640
Here suppose how many clusters?
588
00:54:53,560 --> 00:54:58,080
16 clusters. So there are, so the population has
589
00:55:19,310 --> 00:55:25,820
Step two, you have to choose a simple random
590
00:55:25,820 --> 00:55:31,440
number of clusters out of 16. Suppose I decided to
591
00:55:31,440 --> 00:55:38,300
choose three among these. So we have 16 clusters.
592
00:55:45,340 --> 00:55:49,780
For example, I chose cluster number 411.
593
00:55:51,640 --> 00:56:01,030
So I choose these clusters. Next, all items in the
594
00:56:01,030 --> 00:56:02,910
selected clusters can be used.
595
00:56:09,130 --> 00:56:15,770
Or items
596
00:56:15,770 --> 00:56:18,910
can be chosen from a cluster using another
597
00:56:18,910 --> 00:56:21,130
probability sampling technique. For example,
598
00:56:23,190 --> 00:56:28,840
imagine that We are talking about students who
599
00:56:28,840 --> 00:56:31,460
registered for accounting.
600
00:56:45,880 --> 00:56:50,540
Imagine that we have six sections for accounting.
601
00:56:55,850 --> 00:56:56,650
six sections.
602
00:57:00,310 --> 00:57:05,210
And I just choose two of these, cluster number one
603
00:57:05,210 --> 00:57:08,910
or section number one and the last one. So my
604
00:57:08,910 --> 00:57:12,590
chosen clusters are number one and six, one and
605
00:57:12,590 --> 00:57:19,090
six. Or you can use the one we just talked about,
606
00:57:19,590 --> 00:57:23,340
stratified random sample. instead of using all for
607
00:57:23,340 --> 00:57:29,140
example suppose there are in this section there
608
00:57:29,140 --> 00:57:36,180
are 73 models and the other one there are 80
609
00:57:36,180 --> 00:57:42,300
models and
610
00:57:42,300 --> 00:57:46,720
the sample size here I am going to use case 20
611
00:57:50,900 --> 00:57:56,520
So you can use 10 here and 10 in the other one, or
612
00:57:56,520 --> 00:58:03,060
it depends on the proportions. Now, 70 represents
613
00:58:03,060 --> 00:58:09,580
70 out of 150, because there are 150 students in
614
00:58:09,580 --> 00:58:14,060
these two clusters. Now, the entire population is
615
00:58:14,060 --> 00:58:17,300
not the number for each of all of these clusters,
616
00:58:17,560 --> 00:58:22,310
just number one sixth. So there are 150 students
617
00:58:22,310 --> 00:58:25,090
in these two selected clusters. So the population
618
00:58:25,090 --> 00:58:30,030
size is 150. Make sense? Then the proportion here
619
00:58:30,030 --> 00:58:33,210
is 700 divided by 150 times 20.
620
00:58:35,970 --> 00:58:41,610
The other one, 80 divided by 150 times 20.
621
00:58:51,680 --> 00:58:55,960
So again, all items in the selected clusters can
622
00:58:55,960 --> 00:58:59,400
be used or items can be chosen from the cluster
623
00:58:59,400 --> 00:59:01,500
using another probability technique as we
624
00:59:01,500 --> 00:59:06,640
mentioned. Let's see how can we use another
625
00:59:06,640 --> 00:59:10,860
example. Let's talk about again AUG students.
626
00:59:28,400 --> 00:59:31,800
I choose suppose level number 2 and level number
627
00:59:31,800 --> 00:59:37,680
4, two levels, 2 and 4. Then you can take either
628
00:59:37,680 --> 00:59:43,380
all the students here or just assemble size
629
00:59:43,380 --> 00:59:46,460
proportion to the
630
00:59:50,310 --> 00:59:54,130
For example, this one represents 20%, and my
631
00:59:54,130 --> 00:59:56,730
sample size is 1000, so in this case you have to
632
00:59:56,730 --> 01:00:00,310
take 200 and 800 from that one.
633
01:00:03,050 --> 01:00:04,050
Any questions?