assalamualaikum ladies and gentlemen today we have two tasks to perform the

0:47
first one would be to try and complete the lecture that I was giving and in the
0:56
second one which was basically on consumer behavior research we need to go into a little bit more depth on the
1:03
instruments and the way you approach the research frame and then at the end of
1:09
the lecture lecture I will start launching into the concept we call segmentation why I'm going to keep these
1:17
two things together because if you try and recall what we discussed all customers to our mind are
1:25
different they behave differently they have different ideas they have different thought processes they think about
1:32
things differently however we as marketers still need to
1:37
make sure that we can find a group which has got more or less similar behaviors
1:43
and similar patterns and perceptions so that we can target them in the shape of
1:49
offering our product now keeping these two points in mind therefore the
1:54
importance of marketing research or consumer behavior research is paramount in trying to figure out for the marketer
2:02
which particular segment to target therefore segmentation as a broad
2:07
concept within the parameter of marketing strategy is critical and then
2:15
we will discuss various issues and the importance and why do we need to go into
2:21
market segmentation and then we will build on that now if you recall we had
2:27
discussed issues surrounding data that would be available to a marketer which
2:35
can be used for analysis of the consumer behavior and I mentioned to you that there are basically two types of data
2:43
that we can collect one is secondary data which means that it is already available in the organization in
2:49
documents in trends which are normally available on the internets from the organization and so
2:56
and so forth but we need to be able to also collect primary data which means it
3:03
is for a particular specific purpose and therefore we try and make sure that we
3:08
have the objectives clearly laid down for the purposes of carrying out research and in that we were discussing
3:15
the approaches or the paradigms that we have to come to this data collection
3:22
procedures two of them that I had mentioned one was quantitative research the other was qualitative research the
3:29
difference we had suggested is basically in the shape of the data that you collect in one case it is more metric
3:38
oriented that means it is numerically oriented and the qualitative data is
3:44
basically more open-ended more interpretive more attuned to
3:52
interpretation but it makes a little difficult to analyze that kind of data on the other side the metric data or the
3:59
quantifiable data there are tools available statistical tools which you
4:05
can apply and I had suggested that there are a number of statistical software is
4:10
available to do that kind of job for you which should be undertaken or would be undertaken in a different frame
4:15
altogether but we need to be very clear as to what kind of requirements are
4:21
there in terms of collecting quantifiable data now for that purpose
4:28
we have to come to a definite kind of a framework in which we will try and work
4:34
out how to collect this data what kind of instruments are there what is called what we call the research design so we
4:43
will be discussing that now just briefly to go back onto it we need to remember again that qualitative data collection
4:52
is more interpretive that means the expert who is going to carry out that
4:57
kind of research has to know how to collect this information and it is
5:03
basically used when we are not very clear as to what the problem really is and therefore we tend to go into the
5:10
depth and try to explore areas which is how we do it and the method we have
5:15
already discussed is through in-depth open-ended interviews or carrying out
5:21
focus groups but on the other side of the picture we have what I call the quantitative data collection methods and
5:27
I had suggested to you that in that frame we have broadly three areas one is
5:33
observation the other was experiment and the third was survey are giving you some examples
5:39
also regarding observation and recording experiment today we will try and focus
5:47
more into the area of survey we should understand that surveys are normally
5:52
done through questionnaire and we can have two types of questionnaire so we
5:57
will work out on the issues as to what the questionnaires should be like and
6:03
what does it entail there's these questionnaires we have suggested that there are methods in
6:10
which you can use these questionnaires to collect data and in that we had the
6:16
person interview type where the person was cutting out the research or the
6:22
interviewer fills the form or the questionnaire for part from the customer
6:28
regarding a purpose or a particular phenomena that is to be studied
6:33
basically it could be anything it could be attitudes it could be behaviors
6:40
perceptions and so on so forth the second one that we discussed was telephonic surveys that means somebody
6:47
would go on the telephone call the respondent ask the questions or the
6:53
other method could be that you go to the mall which is called the mall intercept which means you go to the shops where
6:59
the actual buying behavior is taking place and request the customer who's
7:05
just done some purchase to fill up the form for you or exactly you can do it
7:11
for them or the researcher can do it for them finally we had the male survey and
7:17
the male survey was what we consider to be more cost-effective though the response rate
7:24
seems to be less why because if you are cross a person and you can motivate the
7:29
person to help you fill the form the likelihood is that it will be filled but in a mail survey when you send the
7:37
document off you have to depend on the person's own ability or his motivation
7:43
to fill that form up therefore what has been suggested by research that this
7:49
increase in response rate and motivation to give you the correct answer
7:55
really depends on the type of questions
8:00
asked the layout or the firm or format of the questionnaire the sequence in the
8:07
question in the way the questions are asked and all these things pull together would improve not only to get you the
8:16
correct answers but also improve the response rate so we will be discussing
8:21
certain features about the questionnaires because please remember that the most difficult thing that is
8:30
more critical I think in terms of developing a questionnaire is the wording that the questionnaire must have
8:36
and you have to avoid certain pitfalls that means you cannot just go and write
8:42
a simple question because there are implications behind it and we discussed
8:47
these two implications and the implication was that we need to be very
8:54
clear as to what we are going to ask we need to be very clear as to what would
9:01
be the required response because of validity and reliability issues now validity and reliability if you remember
9:08
we discussed very categorically that validity is whether the instrument will
9:15
get you the kind of response that you are looking for whereas reliability more
9:21
or less suggests that if you ask the same question from the same sample you
9:28
will get the same answer so we have to be very very careful regarding the issue
9:34
of reliability now the questionnaires in a very short format can be both disguised and
9:41
undisguised questionnaire now what does it mean to you to be able to say that these questions or these questionnaires
9:49
are disguised or these questions or questionnaire is undisguised the
9:56
difference between the two is that in one case the first case where is disguised you do not tell the person the
10:03
purpose of the research now some people suggest that this could be unethical but
10:10
other people suggest that this type of methodology or this type of method of
10:16
having an a disguised questionnaire gives you better and truthful responses
10:23
the reason Quartet is very simple the reason quoted is that if the people know
10:29
what kind of what is the purpose of the research then people become inclined to
10:38
give you that answer which they think you are looking for so the responses
10:46
from the respondents if it is let's say undisguised and that means the purpose
10:53
is known it would suggest that people will try and give you the answer that they believe that you are looking for
10:58
and i gave you an example at the point of time when I discussed the Xerox
11:03
photocopying system where people were asked how do they use a photo copiers
11:08
and they try to give the laid out format for a manual now therefore if you hide
11:15
the purpose it is more likely that people will give you the right response now that was the first part the second
11:22
part is the questionnaires can have questions which are two types one is
11:27
what we call the open-ended question and the second is what we call the closed-ended question in the open-ended
11:34
question you ask a question in such a form that it gives the other person the
11:41
ability and the flexibility to give any kind of answer that he thinks is right
11:47
so you give him space right his view his thought his
11:52
perception his attitude whatever that he wants to write in a space that you have
11:58
created for him for example the question can be what do you think about whatever
12:05
you want to know and you allow him to go in his own way and give you information
12:11
now the only trouble in this kind of system is that interpretation is more
12:18
difficult that means collecting or collating the information creating means que Cusco egg
12:25
dosa Samer boot Colonel de Tolosa much collage attire because her aadmi in his own language is
12:32
giving you some information therefore the the researcher has to be very good
12:37
in trying to interpret the information that is available in open-ended systems
12:43
and he has to find KSK Beach mojo curry here Jo in sari Java Bank Acharya BOHICA
12:51
this leads to interpretation but on the other hand if the question is close
12:56
ended so close ended Cafe de yeh that when a person asks you a question
13:02
so who's me there is a numerical scale given and the person has to respond to
13:11
any one of the points along which he believes for his perceptions or his PO
13:17
pinions coincide with that kind of a number for example you can say what I
13:23
mean how closely do you suggest your opinion is from among the following and
13:30
you give a question you put the question on whatever the opinion is and you range
13:35
it between one two seven one two five and if one is very close then five will
13:43
be very far away that means if the person says it takes one which will mean
13:50
that the person's opinion is very close to the opinion that you are asking him but if he does not agree with you on
13:57
that opinion then he will score number five so with this method what
14:02
this that you will are able to collect the numerically attune data the other
14:08
benefit of this closed-ended and using some kind of a scale is that you can
14:15
either independently look at each response and you can then analyze each
14:20
response based on how many people were more close to your suggested opinion or
14:27
how many against that opinion so and so forth or you can make an average or
14:34
create a mean by which it would mean that everything can be collected
14:39
together and averages can be taken out so that you get an idea about the overall population beliefs attitudes
14:47
perception and so on and so forth remember that I had already told you that in the case of qualitative research
14:55
the qualitative research is normally done when you are not sure about the problem therefore you are exploring a
15:02
problem so in the quantitative area what we are trying to do is that the are aware of the problem but what we are
15:08
doing is now trying to get the true picture of the phenomena that we are
15:14
interested in for example we are interested in the attitude how does it exist in the targeted population that we are
15:23
concerned with so the quantitative research is more focused on actually
15:29
describing whatever the phenomena that we want to describe as it exists within
15:35
a population or if I would say the profile is developed within the
15:41
population now I told you that questionnaires the most critical issue
15:46
in the questionnaire is the language now we will go over a few points regarding
15:51
the language and we should remember that whenever we are going to write the questions we should avoid these
15:58
following pitfalls now the number one pitfall is that we should not ask a leading question leading question means
16:05
that the answer is suggested within the question for example if I was to say did
16:12
you like this product and of simple form no what I'm saying is that I'm already trying to suggest to the
16:19
person that I'm asking you about liking the product so it will create a biased
16:26
answer it could be better to ask what do you think about the product in terms of
16:34
whether you like it or dislike it on the other side the second difficult situation that we need to constantly
16:41
avoid is to make sure that the language that we use is the language that is
16:48
understood by a common person we do not want to use certain words that are very
16:54
specific to a particular professional so that if it he asked that kind of a
17:00
question and he does not interpret the answer correctly then what we could have had is again a biased answer now the
17:07
third point that we need to be very sure is that we should ask the question which
17:13
the person will be able to answer for example if I was to ask the question how
17:18
many times did you visit let's say macro in the last month it will be very
17:27
difficult for the person to recall each time that he has gone or macro may be a little more easy to remember because you
17:34
go there once a while but if I was to ask let's say going to Anarkali or some
17:40
other significant place where people go and visit and I ask them how many times
17:45
have you been there during a month I think it'll become a little difficult and particularly if you asked him
17:50
between four three months time it will become more more much more difficult now the other issue that you need to be very
17:58
careful about is that the person should be willing to answer a particular question now how does that happen
18:05
do if you are something very personal the person might be totally reluctant in
18:10
answering that question for example in ethical research in the research in
18:16
ethics we normally try and a consumer ethics particularly we normally have an
18:21
issue we find out people sometimes there
18:27
is a problem or shoplifting now shoplifting an issue which is quite understandable
18:32
that exists in our society not only our society but all over the world so if I
18:39
was to ask you the question directly that have you ever shoplifted I assume
18:45
that I will never get the right answer what can be asked is and on the other
18:52
side if you still want to get this kind of information is to ask if I or that if you have observed anybody doing this now
19:01
the point is that this is more likely to get an answer okay
19:06
therefore we have to be very careful in the question that we ask in terms of the
19:13
ability of the person to answer this question the one thing that we are very
19:18
important that we need to be careful about is to totally avoid question which
19:24
has got two points in sight which means if I was to ask that last time you
19:30
visited macro did you get value and were you satisfied now there are two points
19:37
which are in the same question and this is not a very good idea to do because the person might be that might have got
19:44
the value but might not have been satisfied so if he says yes he has a problem so therefore avoided two points or
19:52
asking two questions in one single question that is what you need to keep
19:57
in mind having defined or described the issues that one faces
20:02
while worrying it we now move on to what we call the research design now research
20:09
design requires three things number one the research instrument number two the
20:16
sampling plan and third the method of data collection how would you collect the data and we
20:22
have already discussed that the data collection procedures but in the design form we need to make sure that we
20:28
understand the base that the instrument or the
20:34
questionnaire can collect the data and that in that area will go to the concept
20:39
of measures that we use now in the case of Mayors we normally use skills now there are
20:47
attitude skills and sometimes we are also interested in customer satisfaction
20:52
so at this point of time therefore we will discuss the question of the
20:57
research design and the first part is the type of scale that you use now in
21:03
terms of scale the number of scales that we find there are typically three or
21:09
four that you most probably are in would be interested in and these are used for specific purposes that has to be kept in
21:16
mind so for the purposes of these skills what I will do is I will show each scale
21:23
on the screen and then explain the frameworks that are required to be
21:28
filled or the method of writing the scale now I've told you that all these scales are absolute scales for example
21:34
Likert scale requires agreement or disagreement with a statement that you
21:40
have given now for example what we will be saying is if you see look at the screen the question would be something
21:46
like this for each of the following statements please record the number that best
21:53
describes the extent to which you agree or disagree with the statement now
22:00
number one would be strongly agree two will be agree and you'll go to number
22:08
five which is strongly disagree now the question that is being asked is basically if you look at it it is fun to
22:16
shop online and there is a dash on it which means that the person is required
22:22
to fill the number to which he wants to write on this point but it could be if
22:27
he agrees it totally he will say it strongly agree number one and if he thinks no he does not enjoy shopping
22:34
online then he will fill in fight or any number which is linked with that system
22:40
now the second scale that we are interested in is the satisfaction measures now again in this what we are
22:47
interested in is to find out whether the customer is satisfied or not satisfied with whatever purchase behavior or
22:55
consumer behavior that is exhibit in this again what we are saying is number one will be very satisfied
23:03
whereas number five will be very dissatisfied and normally we find this
23:08
if you recall we find these questions normally asked in restaurants after the
23:13
service has been provided before the bill is paid people normally bring this kind of a questionnaire and asking and
23:19
that is considered to be satisfaction survey the third one is we call the
23:24
important skills now in a part of scales the idea is to be able to measure the
23:30
importance of a particular concept in terms of its importance now for example
23:35
I would be asking maybe whether we consider among the purchase behaviors
23:41
internet is a very important tool now again in this case we have the five
23:47
scale the first one is extremely important whereas number five is not at
23:53
all important which can also be asked for example if you want to ask whether the package is considered to be very
24:01
important for the product if it is very important or not important the person
24:06
will again be required to circle the same answer that he wants the fourth is
24:11
the differential scale which is basically semantic differential scale what does it mean basically semantic
24:17
differential now semantic differential means in the language there are words
24:22
which have got two extremes for example we can have a word called reliability
24:27
and then no reliability or bad or good
24:33
these are two frames there are two extremes of the same idea now semantic differentials normally try
24:41
and give us a picture or a profile of how people rate features for example if
24:48
there is a product that we want to know for example Toshiba we may ask do we
24:54
believe Toshiba is a reliable or not reliable and in this case again the
24:59
scale is divided into let's say 5 odd numbers or sometimes even numbers and
25:06
again the person has to choose the specific number which is what he believes about the concept now in the
25:14
fifth one we have behavior intention skill now behavior intention skills
25:20
normally relates to the future behavior that the consumer would be exhibiting
25:27
which means we ask would you be able to or would you like to buy this product if
25:33
a price is increased by let's say 5% or decreased by 5% now this buyer intention
25:42
sometimes is very critical for deacon for the marketer to understand given
25:47
that he has to have some kind of a strategy coming into position
25:53
pricing most probably adding a new feature and what you are interested in is trying to figure out whether based on
26:01
these features or these variables that you have added or subtracted would the
26:07
customer still be willing to buy into the future so basically buyer intention
26:13
skills are meant to understand the future buying behavior and again let me
26:18
say what the question could be the question would be how likely are you to
26:24
continue using a or B or C product for
26:29
the next six months for example now the first choice is definitely will continue
26:35
second is probably will continue third is might or might not continue continue
26:41
and the fourth is probably will not continue and the fifth is definitely
26:47
will not continue after this there is one more skill that we are interested in and that is called rank order scale now
26:54
rank order scale actually helps you to complete the picture of the person or
26:59
the concept or the the phenomena that you're interested in in terms of how do
27:06
you rank various products or various advertisements or various whatever is
27:13
the choice that you are looking for and the question normally is used for example given the following prices or
27:22
the following quality of the product will you like to buy
27:27
number one and then you say which would be the one that you buy number two number three number four number five you
27:34
give the various products and you ask the consumer to actually put the numbers in front of the choices that they have
27:41
in terms of prioritizing their frameworks I mean whatever they want to
27:47
buy and how they pro ties their behavior in terms of the products or features or
27:52
whatever you want now therefore this more or less covers the kind of skills
27:58
that we normally use and I've already discussed with you the kind of language
28:04
that you use and I've already shown you certain questions that are used and you
28:11
should be able to pay more attention to these now we come into the third part of
28:18
it and as that is where do you collect this data from now this is what we call
28:23
the sampling design or the sampling design includes these following
28:29
frameworks who do you ask this question from how many people would you be using
28:35
to get these answers so what will be the first question is what we call the
28:40
sampling unit so you want to find it from one person and he or she becomes
28:48
their sampling unit so you need to define the sampling unit very correctly number two is how many would you ask 50
28:55
people 100 people that requires the kind of question that is responded to by the
29:02
sample size and the third is how do you actually select these people so the
29:08
third part in the sampling area or the sampling design is how do you carry out the selection of the sample now first
29:16
before we go into the details let's ask ourselves why do we want to have a sample to define why do we want to have
29:24
a sample we must also look at what is the population of interest now basically
29:29
when a marketer wants to sell a product or once you find an opinion the market
29:36
definition is quite vast they're miss probably be hundreds and thousands of and buyers of the same
29:43
product or buyers who the marketer is interested in therefore the researcher can ask the
29:51
whole population the same question but like I said it will become a little difficult you cannot go to every person
29:58
and ask the same question so what we do is that we try and find a group of
30:04
people or group of the unit sampling unit which totally represents the
30:09
population the idea comes from and let me tell you that normally we think about
30:16
how reliable is this instrument and how valid is the research that we have
30:21
undertaken then one of the major concepts within that is the sampling frame the issue is that if you have
30:29
selected the correct sampling frame from which you will choose the actual people
30:34
then this sampling frame must represent the overall population the idea is that
30:40
what we are trying to do is to measure that one concept we measure that one
30:46
phenomenon measure the one thought or of interest from among the total population
30:53
but how do we do it we use this sample to represent the population we if we can
30:59
accurately target these the units properly then whatever the answer given
31:06
by this the sampling unit we would assume would represent the total
31:12
attitude or the total perception of the overall population please remember that
31:18
the idea is that we need to generalize our research our findings on a
31:25
particular population now generalizing a population or generalizing the idea over
31:31
a population we would like to have a sample which will be exactly representative of the population
31:38
therefore the most important thing before we can actually draw out a sample
31:44
we need to define the population on which we want to were called generalize
31:51
the research in doing that what you need to do is to be able to profile the the the
31:59
population in a very in-depth manner so that when we want to select who would be
32:07
part of that sample or research we are very clear that we are not making a mistake so we are actually asking a
32:13
question from the person who is representing it I'll give you an example here which was a mistake made and later
32:20
on it was considered that it was a mistake long time back there was a
32:27
elections take place in America and people try and predict who will win just
32:33
like it happened in the present time it was considered that Obama will win and
32:38
in doing that people from all over America would not ask the same question
32:45
however the census the people who do the polls they check from a limited number
32:53
of people to ask whether who will win Obama will win or McCain will win and
33:01
ultimately it was concluded the Obama will win and obviously he did win but
33:06
this was done through a sample now long time back again at the beginning of the
33:13
20th century that is somewhere around nineteen hundred and thirty a similar research was done by the same poll
33:21
company which tried to figure out who will be the president at that point of time the only thing was that they asked
33:29
a question from those people who actually owned telephones at that point
33:35
of time because it was simple to reach them on the telephone so they got an answer but unfortunately the person who
33:44
they predicted would win never won the election it was the other person who won
33:49
so later on it was discovered that the sample that was chosen was based on the
33:56
ownership of telephones so the voting was done by people who did not own even
34:02
telephones but like I said the sample was the question was asked from people who
34:08
did own telephones and this gave a wrong research findings and that is why we say
34:14
and remember that at that point of time close to about question were asked from about 70,000 people now 70,000 is a very
34:23
large number I'm not going to quote the exact number but it was approximately around that range but that was a very
34:30
large number so it was assumed that this census will be and this research would
34:36
be accurate however it was not so we come to two points here number one is the sample size what should be the
34:43
sample size number two is the representation of the population by the
34:49
sample according to research it is suggested that if the sample represents
34:55
the population accurately then the likelihood is that the error between the
35:02
findings in the sample and the actual position in the population will be very
35:09
small and can be ignored so therefore the the idea is that when you are going
35:15
to select a sample which is obviously the selection of the sample frame then
35:20
it should be defined in such a manner that it will represent the population
35:26
very very accurately let's take an example here for example you want to do
35:31
research among students if you were to do research among students which you
35:37
think are of the age group between 18 and 25 and studying at the bachelors
35:43
level then if you only went and asked the question from people who are
35:50
studying business because there are limited business schools and you can
35:55
approach these people in a very simplistic manner then your findings are
36:00
going to be wrong because they are not going to be related to all the bachelor students but only specifically to the
36:08
bachelor students who are studying the business to be able to collect information from all for the purposes of
36:15
generalizing on all graduates then you must have the ability key to select the sampling frame from
36:23
all these students who are studying in the a graduate program having said this
36:28
we now must look at how do we select the
36:33
sample from the defined sampling frame now to be able to do that we have two
36:39
methodologies one method is what we call the random sampling method and the
36:44
second one is the non random sampling method now first of all what's the
36:50
difference between the two and obviously what are the outcomes of these two methods it is assumed that if you have
36:58
chosen the sample totally randomly the idea is that every unit I've already
37:07
defined the sampling unit every unit in this sample has the opportunity or the
37:14
probability to be included in this sample and that is what is random
37:19
sampling the logic is that if you ask questions randomly from the sample
37:25
population then you are most likely to get a very generalized view in which you
37:32
are interested in however if you go to non-random sampling then what will
37:37
happen is that you will most probably be inclined towards a particular group of
37:44
people which will bias the results now what is the issue here is that if you
37:51
look at the board I'll show you what is the basic formula that we want to have in terms of regression now we have Y Y
38:01
is the dependent variable which we want to study and X is the independent
38:08
variable and the formula that we must have studied in our basic algebra
38:14
courses Y is equal to a plus BX plus E now what is this e is this error term
38:21
which actually means that the impact that is being brought by the independent
38:28
variable which is X on the dependent variable which is why is not exact in the sense that there
38:36
could be some other factors which are operating which is captured by the term II now what is of our interest as
38:45
researchers is to reduce the impact of this error term II now one method of
38:51
reducing this error term II is that if we can reduce the bias or the noise that
38:57
is created in this relationship between the X and the y to do that we need to
39:05
make sure that this error which is e is only related to what we call the
39:12
sampling error now selling error means that whatever is the difference if we
39:17
keep repeating let's say if we keep repeating this analysis and we keep
39:22
collecting data on the same frame and we are keep analyzing it then this change
39:28
that is taking place in terms of Y based on the term X is going to be related to
39:36
the change in D type of sample that you have collected which means that it is
39:42
random error coming through which means it is not a systemic error or systematic
39:49
error which means it is coming in because you are taking information from
39:55
the sample but from different people within the sample now this is why we
40:02
want to make sure that we have error term to be reduced and we make sure that
40:08
there is less biasness obviously we cannot eliminate the biases totally but we want to reduce the biases
40:14
on the other hand if you look at it if we have a non random sample then number
40:19
one mathematically we cannot we cannot measure this sampling error which means
40:25
we will never know whether this error that is coming in repeatedly is it
40:30
because of the randomness of the sample selection or is it coming out because of
40:35
a bias and that is why we want to make sure that we select the random sampling
40:41
however let me put it this way that there are certain types of researchers which may
40:47
not permit in terms of time expenses cost whatever for you to go into a
40:55
totally random selected sample randomly select a sample but we need to be very
41:01
clear that the best results particularly in quantitative analysis will come when
41:07
the sample is random now again there is a direct relationship between the
41:13
randomness of the sample and the size of the sample now if the sample selection is totally random because like we said
41:20
we are trying to represent the population then if the selection is
41:26
random we believe that it will represent exactly the population however in terms
41:31
of non random sample to reduce this difference between the sample finding
41:38
and the populations actual behaviors we need to increase the sample size for
41:44
example if you are interested in getting information about hundred people then
41:49
from if we just collected the information from ten but the sample is
41:55
totally random we believe that it will be quite accurate in representing the 100 on the other hand if we are going to
42:02
go for a non random sample the representation of the population must be determined by increasing the size of the
42:10
sample which means let's say if we have randomly selected and we only choose ten
42:15
of them then it is possible sorry non randomly selected if we have chosen only ten it is likelihood that we
42:22
are going to have a biased representation but to improve over on that we might like to go to 5060 now if
42:30
we go to 60 then the likelihood of the error between the information that is
42:36
generated out of this sixty and the information that is available or sitting in the hundred will be minimal that the
42:43
difference will be minimal therefore sample size is directly related to also the randomness of the selection
42:50
procedures how do you do it now to the various ways of random sampling first
42:57
and then we will go to the non random sampling simple land sample or selection method is when every
43:04
unit has a chance of being included in the sample of the question there are
43:10
many methods of random sampling many are there which includes using or systems
43:18
that is you put the numbers into a box and shake the box and then take out a
43:24
number and the number is randomly selected and you do not have a bias it's a very simple example of a simple random
43:31
sample the second method is which is called a systematic random sampling
43:36
method and in that what you do is let's say you have a list of a thousand people and you select the first person or the
43:44
first unit per sample unit in terms of using the same method that I've given
43:50
you so you let's say collect number nine and then you decide that every sixth
43:56
person will be included in the sample now this is what they call systematic
44:01
random sampling method which means the first choice is done totally random and then systematically you take a number
44:08
and you go and proceed through that whole system the third method is what we
44:15
call clustering system now cluster random sampling is that I mean for
44:21
example we want to sample the population in Lahore what we can do is we can break
44:27
down this whole population of Lahore into areas blocks like say your town
44:33
then we might have Goldberg and we might have Controlman tanned so on so forth
44:38
and then since we have clustered the various areas in terms of blocks we
44:45
select from each block the sample in the
44:50
random manner in which I've already told you for example we can do it by the number of or given the houses a
44:56
particular number and then choose the house number randomly for each area so
45:04
we get clusters of these days sample based on certain frame finite method
45:11
that we have is what we call static sampling now what happens is and a good
45:16
example again is it's very similar to cluster sampling but stratified is when the create Strutters of the population
45:25
of interest and then pick samples from there let's say we want to study
45:30
students in a university let's take a simple example and we want to study students in a university and we know
45:38
that students are enrolled let's say in the intermediate in the undergraduate
45:44
program then in the graduate program and in the ph.d program or the you know postgraduate program what we do is that
45:53
we identify the various strata so there is a strata which is basically only the
45:59
intermediate students second strata is the undergraduates the third strata is
46:05
the graduates and the final is the postgraduate then what we do is we look at the number of people in the
46:11
intermediate strata and decide that we will select 10% of the intermediate we
46:17
decide that we will select 10% of the undergraduate and 10% of the total
46:23
population in the Graduate and then the postgraduate and within these then we
46:29
select them randomly so we get a list of people in the intermediate level and
46:35
make our choice based on the frames that I've already told you in terms of
46:40
whether you use the random simple random sample or systematic now once we are
46:48
clear on these four methodologies we have on random sampling we now look at the areas of non random sampling now
46:56
there are three major methods available to us in terms of non-random something
47:01
one is called convenient sampling what is basically convenient sample and that means basically that we choose a sample
47:10
which we can approach easily it is on art or the researchers own convenience that he approaches that group of people
47:17
who he believes fall within the population frame who represent it but
47:23
the choice of the individuals are based on the convenience of the researcher and is why I said that this may create a
47:30
bias because the person who is approaching the people who he knows might have similar behavior than similar
47:38
attitudes so they're obviously response would be more similar in that framework
47:43
the second one is called judgmental sampling now again we have to be very
47:48
careful judgmental sampling is a little tricky in the sense that we need to make a judgement about these sample being
47:56
representative of a population who we can make a good judgment it has to be an
48:01
expert who can say yes that this population is represented by this sample
48:07
so judgmental sample is basically choosing a sample which in the opinion
48:12
of the expert is accurately representing the population the final one is called
48:18
the quota sampling now quota sampling is very similar in terms of the cluster or
48:25
the stratified sampling but the only difference is that in quota sampling
48:30
once you have determined the number of people that you will choose from various strata or various clusters what you do
48:38
is that you select them again on a non-random base okay so when you select
48:44
them on the non-random base again we have what we call quotas given to a
48:49
particular discharge a particular cluster and the choice is non randomly selected okay today I had planned to
48:59
move into the lecture on segmentation but since we have taken very in-depth
49:05
view of the area of research and I hope that it has been able to clear a lot of
49:12
areas for you we will take up the segmentation section in the next class
49:18
so before we finish I would like to just touch upon the ethical issues that we
49:23
face while we are carrying out this research the ethical issues are generally very simple to understand but
49:29
we need to be very careful when we are collecting data now one of the most important and significant things in the
49:37
ethical issue is that the data that you collect is very personal and therefore the person you
49:44
are asking the question we call them respondents they must be assured that this information that you
49:50
collect will not be used for any other purpose except for the purpose of the research that you are conducting now
49:56
this is very very important number two there should not be any misrepresentation from the interviewer
50:04
to the respondent I'll just give you a very simple example there are times that
50:09
people have done it and I think that is unethical to do and that is that you
50:14
make a call or some interviewer makes a call to the likely respondent and tells
50:19
him to induce him to give the answer and he tells him it will take only 10 minutes to do so that he will be taking
50:26
only 10 minutes but then what he goes and does is that he carries on the questioning and the research for the
50:33
next 30 to 40 minutes now this is not justified so therefore there are certain
50:38
basic issues which are incorporated or which should be.we be aware of in terms of ethics in research framework now to
50:47
tie up this day's session we have discussed basically the instruments and
50:52
the language that is required to be used and what they should avoid doing while we are framing the question second area
50:58
that we looked in very depth is the kind of skills and measurement skills that we
51:04
have and what are the basic what are called the features of these skills and
51:09
how and where can livi use these skills finally before the finish and I told you
51:15
that in the next class we will go into the area of segmentation which is critical and you will understand the
51:22
linkage between research as well as segmentation procedures and the concept
51:29
as it fits into the marketing strategy okay thank you very much
English (auto-generated)