Unit 3 Data Collection
Unit 3 Data Collection
(UNIT 3)
BMS 4th Sem
Sugandha Jain
DATA COLLECTION
MEASUREMENT AND SCALING
Classification
of Data
Primary Secondary
Primary Data
• Original
• Problem or project-specific: collected for the specific objectives and
needs spelt out by the researcher.
• The authenticity and relevance is reasonably high
• The monetary and resource implications of this are quite high
• Sometimes a researcher might not have the resources or the time or
both to go ahead with this method. In this case, he/she can look at
alternative sources of data which are economical and authentic
enough to take the study forward – namely the secondary data.
Secondary Data
• Information which is not topical or research- specific
• Collected and compiled by some other researcher or investigative
body.
• The said information is recorded and published in a structured
format, and thus, is quicker to access and manage.
• In most instances, unless it is a data product, it is not too expensive to
collect.
• Also the information collected is contextual; what is primary and
original for one researcher would essentially become secondary and
historical for someone else.
Advantages of Secondary Data
Resource Accessibility of
advantage data
Applicability Accuracy of
of data data
Criteria for evaluation of secondary data
Methodology Accuracy
Check Check
Cost-Benefit
Topical Check
Analysis
Criteria for evaluation of secondary data
• Methodology Check: The process or design used to collect the data so that in case there has been an
element of skewed respondent selection or bias, one can detect it here. Includes sampling considerations,
methodology of data, analytical tools used and subsequent reporting and interpretation of results.
• Accuracy Check: If the study is conducted and the findings compiled by a reputed source, the reliability of
using it as a base for further research is higher, viz., one conducted by a newcomer or on a small scale. In
case information is from such a source, it is advisable to collect similar data from multiple sources and
then collate the findings. A related problem that might occur is when different studies/sources report
contrary findings. In such a case, a short pilot study, supported by an expert opinion survey would help
achieve the right perspective. This is termed as cross-check verification. Another problem of accuracy is
when the data is deliberately manipulated for the purpose of the study. The inaccuracy could also be in
the presentation of the findings, i.e., the scale used might artificially enhance or play down the results.
• Topical Check: There might be a considerable time lag between the earlier reported findings on the
subject and the research being conducted now. If one is looking at the impact of variables such as age and
gender on the purchase patterns of apparels, 5 years is a period where trends and fashions might have
changed and presumptions or hypotheses made on the basis of such a data might be erroneous. To
address these problems, a number of market research firms have started publishing syndicated data which
is periodically updated.
• Cost-Benefit Analysis (Financial Check): Before using secondary data, one needs to measure the cost of
procuring the data, viz., the advantage of the information.
Data Sources
Primary Secondary
Sources Sources
Internal External
INDIRECT Projective
(Disguised) Techniques Expressive
Choice/Ordering
OBSERVATION
• Most appropriate for descriptive research
• Often gets ignored as it appears too simplistic.
• Might be carried out in a naturalistic environment where there are no control
elements or it might be carried out in a simulated environment under certain
controlled conditions.
• The task of the observer-investigator is not to question or discuss with the
individuals whose behaviour is being studied.
• The event being observed might involve a live observation and reporting or it
might involve observing and inferring from a recording of the event.
• Observation involves viewing and recording individuals, groups, organizations or
events in a scientific manner in order to collect valuable data related to the topic
under study.
Observation
Structured Unstructured
Structured Observation
• The nature of content to be recorded and the format and the broad areas
of recording are predetermined.
• Thus, the observer’s bias is reduced and the authenticity and reliability of
the information collected is higher.
• For eg., Fisher Price toys carry out an observational study whenever they
come out with a new toy. The observer is supposed to record the appeal of
the toy for a child, i.e., how often does he/she pick it up from a collection of
the toys available. What is the attention span in terms of how long is it able
to engage the child? Is there any safety issue with the toy? What was the
reaction of the child while/after playing with the toy? Thus, for a clearly
defined information need, in terms of parameters to be noted, it is an
extremely useful and a non-intrusive method. This method is useful for
cross-sectional descriptive studies.
Unstructured Observation
• The observer is supposed to make a note of whatever he understands as relevant for the
research study.
• More useful in exploratory studies where there is a lack of clearly-defined objectives and one is
still trying to identify what parameters need to be investigated.
• Chances of observer’s bias are high as the observer has his/her own presumptions about the
situation being observed. To overcome this, have multiple observers for the same situation in
order to get different perspectives about the same instance.
• For eg., the observation of consumer experiences at a bank to get an insight into the intangible
needs and individual behaviour of service personnel. It could give clear indications of the
elements that might create an unhappy experience or might lead to customer delight. But,
giving clear mandates about what to observe might miss out on important elements of the
service experience which might be critical in delivering a superior value.
• The researcher must have a preconceived plan to capture the observations made. It is not to be
treated as a blank sheet where the observer reports what he sees. The aspects to be observed
might be indicative areas on which the observation is to be made.
Observation
Disguised Undisguised
Disguised Observation
• The observation is done without the respondent’s knowledge who has no idea that
he/she is being observed.
• The investigator is considered to be like a ‘fly on the wall’, there has to be absolutely no
contribution in any way to the situation being observed. He cannot send any verbal or
non-verbal cues to the respondent, which might impact the behaviour being observed.
• The advantage is that since the respondent does not know, one is able to record the
natural manner in which the person behaves and interacts with others in his
environment.
• This may be accomplished by having observers who are a part of the group or the
organization. It is also possible to use other devices like a one way mirror or a hidden
camera or a recorder.
• The only disadvantage is the privacy issue, as this is ethically an intrusion of an
individual’s right to privacy.
Undisguised Observation
• The knowledge that the person is under observation is conveyed to the
respondent.
• Proponents state that the influence of the observer’s presence is brief and does
not really have any effect on the natural way a person behaves. While the other
school of thought is that it distorts an individual’s behaviour pattern drastically.
The decision to choose one over the other depends upon the nature of the study.
• Whenever the objective is to study the latent, subconscious or an intangible
aspect of human behaviour, it is recommended that one opts for disguised
approach. However, when the observation is accepted as non-intrusive as it is a
part of the process the undisguised approach can be used.
Observation
Natural Artificial
environment environment
On the basis of setting of data collection:
Natural or Artificial
• Natural observation, as the name suggests, is carried out in real time locations,
for example the observations of how employees interact with each other during
breaks.
• On the other hand, it could be an artificial or simulated environment in which the
respondent is to be observed. For eg. this is actively done in the armed forces
where stress tests are carried out to measure an individual’s tolerance level.
• Evaluating the reactions of respondents to the phenomena or strategies under
study can be carried out at a smaller scale in a contrived situation, as these would
help predict the behaviour likely to occur, in the actual situation.
• However, when the object is to study true reactions and not the supposed ones,
natural observation is recommended.
Observation
Human Mechanical
observer observer
On the basis of type of Observer:
Human or Mechanical
• Human observation and recording is done by human observers.
• Analytical ability of the recorder makes this mode far superior to mechanical recording. However,
this very advantage might prove to be a negative of the technique as presence of human element
can lead to bias.
• Thus, maintaining objectivity while reporting and inferring is of critical importance. The exact
definition of what are the parameters to be observed in the case of structured observation are
extremely important.
• For example, if we need to observe the level of initiative, then it is essential to define the kind of
behaviour that is part of the job role and which might be construed as initiative. This is critical if
observation is the major data-collection instrument for a descriptive study. This will ensure the
reliability of the findings.
• In mechanical observation, man is replaced by machine. This might or might not involve directives
by human hand. Generally, the recording is done continuously and later subjected to interpretation
and analysis.
Turnstile
Located at the entrance of a
store, mall, office or even
traffic locations to collate data
about individual or vehicular
movement at different times
of the day.
Net Scanners
Observe the net surfing
behaviour in terms of the
time spent, sites visited and
links used are extremely
valuable insights into mapping
consumer interests, as this
helps in designing product
and promotion offering, thus,
catering to the needs and
interests of the potential
users.
Psychogalvanometer
Measures changes in the
electrical resistance of the skin.
Small electrodes are attached to
the individual’s skin and these
electrodes are in turn attached
to a monitor. Any affective
reaction of the individual results
in a higher perspiration which, in
turn, results in a change in the
electrical resistance of the skin.
The respondent could be
exposed to different kinds of
packaging, ads and product
composition, to note his
reaction to them.
Oculometer/
Pupilometer
Measures the impact of various
stimuli on the sense of sight.
Eye-tracking equipment such as
oculometers record movements of
the eye to determine how a
respondent reacts to ads, packaging
options, shelf or store displays.
Pupilometer measures the interest
of the person in the stimulus by
measuring changes in the diameter
of the respondent’s pupils. The
hypothesis being that more the
increase in the size of the pupil,
more positive is the attitude of the
individual towards the stimulus.
Voice Pitch Meter
Measure emotional reactions
of the individual by reporting
on any change in the
respondent’s voice by
measuring any change in the
voice pitch of the person.
CONTENT ANALYSIS
• Studying a previously recorded or reported communication and systematically
and objectively breaking it up into more manageable units that are related to the
topic under study.
• It is peculiar that it is classified as a primary data collection technique and yet
makes use of previously produced data. However, since the analysis is original,
first hand and problem specific, it is categorized under primary methods.
• Some researchers classify it under observation methods, the reason being that in
this, one is also analysing the communication in order to measure or infer about
variables. The only difference being that one analyses communication that is
ex-post facto rather than live.
• One can content-analyse letters, diaries, minutes of meetings, articles, audio and
video recordings, etc.
Method of conducting content analysis
• Structured and systematic and thus of considerable credibility.
• For example, we want to know what makes the young Indian tick,
we could make use of blogs written by youngsters, essays and
reality shows featuring the age group. We decide that we want to
assess value systems, attitudes towards others/ elders, clarity of life
goal and peer influences.
• This step is extremely critical as this indicates the assumptions or
hypotheses the researcher might have formulated.
U: Universe of Content
Space and
Word Theme Characters Item
Time
• Word. Can be easily subjected to computer analysis. For eg., the values that we want to evaluate are
individualistic or collectivistic, aggressive or compliant. Thus, we can sift the communication and
place words such as ‘I’ or ‘we’ under the respective heads. Words like ‘hate’ ‘dislike’ go under
aggression and ‘alright’ ‘fine’ for complacency. Then counts and frequencies are calculated to arrive
at certain conclusions.
• Theme. Difficult to quantify as it involves reporting the propositions and sentences or events as
representing a theme. For eg., disrespect towards elders is the theme and one picks out the
following as a representative: a young teen’s blog which says my old man (father) has gone senile
and needs to be sent to the looney bin for expecting me to become a space scientist, just because
he could not become one... This categorization becomes more complex as the element of observer’s
bias comes into play. This kind of analysis could be useful when carried out by an expert. However,
the reliability and validity of the findings would be questionable in the hands of an untrained analyst.
• Characters. The person producing the communication, for eg. the young teenager writing the blog.
• Space and time measures. Related to the physical format, i.e., the number of pages used, the length
and duration of communication.
• Item. Refers to categorizing the entire communication as say ‘responsible and respectful’ or
‘aggressive and amoral’. Observer’s bias is likely to be high. To ensure reliability of findings, can ask
another coder to evaluate the same data.
• Content analysis of large volumes becomes tedious and prone to error if
handled by humans. Thus, there are various computer program available
that can assist in the process.
• TEXTPACK can tag defined words for word frequency by sorting them
alphabetically or by frequencies.
• Open-ended questions can be sorted by Verbastat or Statpac, which has
an automatic coding module.
• Content analysis is a useful technique when there is large quantity of
textual data and it needs to be structured in order to arrive at some
definite conclusions about the variables under study. Computer assistance
has greatly aided in the active usage of the technique.
• However, it can appear too simplistic, when one reduces the whole data to
counts or frequencies.
FOCUS GROUPS
• More cost effective and less time consuming
• Generates large amount of information in a short time span.
• Group dynamics play a positive role in generating data that the individual
would be hesitant about sharing when he was spoken to individually.
• The process generally involves a moderator who maneuvers the discussion
on the topic under study and a group of carefully-selected respondents who
are specifically invited and gathered at a neutral setting.
• The moderator initiates the discussion and then the group carries it forward
by holding a focused and an interactive discussion.
Key Elements of a Focus Group
There are certain typical requirements for a conducive discussion. These need
to be ensured in order to get meaningful and usable outputs from the
technique.
1. Size: Not too small, not too large
2. Nature: Similar background
3. Acquaintance: Strangers
4. Setting: Neutral
5. Time period: Single sitting, should not exceed 90 minutes
6. The recording: Camera over human recording
7. The moderator: Skillful, attentive, good communicator, knowledgeable,
unbiased, flexible, ability to make respondents feel comfortable, sensitive,
etc.
Steps in Planning and Conducting Focus
Groups
1. Clearly define and list the research objectives.
2. Split objectives into information needs to be answered by the group. These may be
bulleted as topics of interest or as broad questions to be answered by the group.
3. Prepare list of characteristics, which would be used to select the respondent
group. Based on this screening, a questionnaire is prepared to measure the
demographic, psychographics, topic-related familiarity and knowledge.
4. Chart out outline for conducting the whole process. It is critical to involve the
decision maker, the business researcher as well as the moderator. Done to ensure
complete clarity for the moderator in terms of the intention and potential applicability
of the discussion output.
5. Carry out the discussion.
6. FGDs reveal rich and varied data, thus the analysis cannot be quantitative or in
frequencies. Thus, summarize the findings by clubbing under different heads as
indicated in the focus group objectives and reported in a narrative form. Include
expressions like ‘majority of the participants were of the view’ or ‘there was
considerable disagreement on this issue’.
Types of Focus Groups
1. Two-way focus group: One group sits and listens to another group first.
2. Dual moderator group: For managing group and group mind
3. Fencing moderator group: Take opposite sides on the topic
4. Friendship group: When comfort level of members needs to be high
5. Mini group: 4-6, expert committees
6. Creativity group: Longer than 90 minutes (may be 1-2 days long,
workshop mode), extract alternative thinking
7. Brand-obsessive group: Involve respondents passionate about a brand
8. Online focus group
Advantages of Online Disadvantages of Online
FGDs FGDs
• Geographic locations not a • Used only with computer savvy
constraint people
• People who are busy and • As anonymity is assured,
otherwise are not able to authentication of the respondent
participate, can also be tapped being a part of the population
• Interaction is faceless so under study might be difficult to
anonymity is assured and is thus establish.
less inhibited. • Other sensory cues of tone,
• Cost and time advantage as body language and facial
compared to a traditional focus expressions are not available.
group.
Advantages of FGDs Disadvantages of FGDs
Personal Telephone
Interview Interview