See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/335345700

Quantitative Data Collection Methods

Chapter · August 2019

CITATIONS READS
0 9,874

1 author:

Murtada Busair Ahmad

Nile University of Nigeria
18 PUBLICATIONS   3 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dr. Murtada Ahmad & Associates on Folk-media View project

RESEACH PUBLISHED IN CONFERENCE PROCEEDINGS View project

All content following this page was uploaded by Murtada Busair Ahmad on 23 August 2019.

The user has requested enhancement of the downloaded file.

 Research Methods In Communication & Media Studies

CHAPTER 15
Quantitative Data Collection Methods
Murtada Busair Ahmad

Introduction
Data collection is a significant factor in social scientific research and, by extension, in empirical
communication research. Arising from the background of specific theoretical assumptions or
from a set of existing research findings, investigating a social phenomenon mostly depends on
adequate and representative sample of observations in order to generate evidence from different
points of view on the phenomenon (Babbie, 2001, pp.20-22). The evidence to look for and how
to look for it are contingent on the methodology chosen and its specific assumptions. The method
adopted for an inquiry would be determined by the existing literature or body of knowledge that
deals with the phenomenon to be investigated, the conceptualisation of the phenomenon, the
problem formulated, the research questions or hypotheses and the research design. The focus of
the aforementioned elements of social scientific research is narrowed to a particular research
topic that is worthy to be studied. Such a topic may emanate from a theoretical proposition that
requires testing, a practical problem that begs for solution, or an experience that seeks
explanations (Frey, Botan, & Kreps, 2000, p. 14-15).

The foregoing suggests that data gathering does not exist in isolation; it is rather anchored in
definite and significant nuts and bolts whose substances reflect in form of questions to be
answered and items to agree or disagree with in the data gathering instruments especially self-
administered questionnaire and interview for survey and field research respectively, and coding
sheet for content and textual analyses. Scholars have suggested how questionnaire ought to be
planned, designed, and worded in order to give good and sufficient data to address research
questions and test hypotheses. To that extent, they suggest that consideration should be given to
the essential information needed, the pattern which the questions should take, and the statistical
techniques to use for analysis (Khan & Khan, 2006; Pallant, 2007, p.7). Likewise, they imply
the necessity and combination of closed-ended and open-ended questions in the case of
interview to enable free and unimpeded responses from respondents (Pallant, 2007, pp.8-9) and
combination of open-ended questions along with observation and participation in qualitative or
field research (Arensberg, 2006).

Coming from that background, communication research draws its significance from the use of
diverse social scientific methods for collection of data that can reveal varied but definite roles
played by people, as sources and receivers of messages, channels, messages, and contexts in
specific social phenomena. The significance of communication research does not only lies in the
fact that it has become the means to understand how people, channels, and contexts interrelate in
order to influence messages in the society, but also in the centrality of the contemporary mass

188
 Research Methods In Communication & Media Studies

media to “organising every aspect of contemporary life, from the broad patterning of social
institution and cultural systems to intimate everyday encounters and people‟s personal
understanding of the world and their sense of themselves” (Deacon, Pickering, Golding &
Murdock , 1999, p.1).
This chapter discusses an array of data collection methods applicable to communication and
media research against the backdrop of quantitative content analysis and cross-sectional survey
research designs.
Quantitative Content Analysis
Traditional content analysis has to do with examining, and counting, of manifest content of
media messages both in written and verbal forms. Studies dealing with content analysis
pragmatically respond to the questions: who says what, to whom, why, how, and with what
effects? To conduct content analysis, the following must be determined: categorization of the
content to be studied, the unit(s) of analysis, and the attitudinal direction of massage(s). These
three fundamentals of content analysis apply to the study of all media, specifically mass
communication media whether print or electronic. Hence, decision should be made on the choice
of mass communication media to investigate.

Category Construction or Content Categorization

Category construction refers to identifying specific genres in the medium of our choice.
Newspapers and magazines‟ contents cut across genres like news, editorials, opinions, features,
sports, letter-to-the-editor, cartoons, adverts, and advertorials. Radio‟s messages or programmes
are classified into genres such as news, commentaries, current affairs, public affairs,
documentaries, features, dramas, music, adverts, sermons, etc. Likewise television‟s messages
are grouped into genres like news, current affairs, documentaries, sports, soap opera, movies,
adverts, etc. Content categorization can also surface in other ways. This will depends on what
makes up the universe or population of study. A study may involve only one news genre. Such
genre could therefore be divided into two or more categories relative to the subject matter
involved or the issue of our research interest. Most importantly, the categories in which the
message in focus is coded must be mutually exclusive, equivalent, exhaustive, and parsimonious
(Frey, Botan, &Kreps, 2000, p241).

Therefore, the specific genre(s) to be studies should be included in the research design right from
the proposal stage. Identifying genres to be studied will partially specify boundaries of the
content involved, which are normally considered through operational definition of concepts.
Most often contents to be analysed are determined by the problem statement, research questions,
and hypothesis explicitly stated in the study. A good problem statement always puts the
researcher on the right track in deciding the exact genres of investigation (Khan & Khan, 2006).

189
9
 Research Methods In Communication & Media Studies

Units of Analysis
Units of analysis are the individual parts of contents to be specified and isolated for study.
Selection of specific units of media contents for analysis depends largely on the nature of the
study relative to variables highlighted in the research questions and hypotheses. Selection of
units of analysis may also depend on the universe or population of our research. For instance, if
newspaper‟s editorials form the population of our study, then editorials of selected newspapers‟
editions would serve as units to be analysed. Then, there would be the need to develop content
categories of editorials of our concern in nominal terms. Categorizing the editorials (or similar
news genres) is an exceptionally creative process that relies on the researcher‟s sense of
diligence and resilience as there are indefinite numbers of categories into which editorials could
be grouped. Hypothetically, newspaper‟s editorials on Boko Haram‟s insurgency could be
contextually categorized into three: (a) editorials before Chibo Girls’ abduction, (b) editorials
during Chibo Girls’ captivity (c) editorials after Chibo Girls’ rescue.

It can therefore be said that whatever parts of the contents specified as units of analysis would
serve as cases to be studied. Technically, words, sentences, paragraphs, columns, pictures, a
whole story, a whole editorial, a whole cartoon can be made units of analysis in a study that
involves newspapers or magazines.

In the like manner, words, statements, speeches, pictures, body gestures, costumes, scenes, can
be specified as units of analysis in television or film oriented studies. If the study involved deals
with patterns of coverage and camera shots of scenes in a play, drama, soap opera, or
documentary, shot types can be identified as units of analysis. Therefore, close-ups, long shorts,
zoom-in, zoom-out could feature as units of analysis in the study.

In a situation whereby a genre such as editorial is adopted as unit of analysis and coded into two
or more categories, decision should be made on what should be units of observation among the
units of writing. Words, phrases, clauses, sentences, and paragraphs can always serve as units of
observation. Likewise, column inches could be counted regarding coverage of issues and events.
Krippendorf identifies five types of units that might be observed for analysis: physical units,
syntactic units, referential units (or character units), propositional units, and thematic units.
(Krippendorf Cited in Frey, Botan , &Kreps, 2000, p.240).

Physical units include the amount of space or time devoted to news stories, editorials, and similar
genres on specific issues and events. Newspaper‟s column inches devoted to such messages also
belong to this type of unit. Whenever we focus the average story length or proportion of items on
front page, inside page, and back page, we are dealing with physical unit of analysis. Physical
unit of analysis characteristically use standardized measurements. The remaining four unit types
are called meaning units because they deals with symbolic meanings of contents. They are
therefore less standardized compared to physical units.

190
 Research Methods In Communication & Media Studies

Syntactical units involve discrete units of language or units of writing composition like words,
sentences, and paragraphs. For example, a researcher‟s interests may have to do with
investigating the extent of balanced coverage of a controversial issue or event by selected
newspapers. The researcher could go for counting the number of words or sentences that are
allotted to both sides of the event or issue.

Referential units involve units of allusions made to persons, objects, events, in selected
newspapers‟ reports, editorials, features, etc. Counting of allusive and insinuative words,
phrases, clauses, and sentences may interest a researcher in relation to special issues and events
reported in selected radio or television channels.

Thematic units are themes or topics or subject matters reported about an event or issue of interest
in selected newspapers, magazines, radio stations, or television channels. A researcher may
decide to content-analyzed themes or topics mostly emphasized in newspapers‟ coverage of
specific events or issues. For example, the researcher may unitize stories on Boko Haram‟s
insurgency against the backdrop of the themes under which they are reported. Themes such as
peace brokering, kidnapping, bombing, arrests, rape, etc. may serve as units. In fact, it is possible
that a single story entails two or more themes. For instance, while bombing and killing may
occupy the beginning of a newspaper story, peace-initiatives and cease-fire may be mentioned
toward the story‟s end. Therefore, the number of occurrence of such themes in different stories
published by a given number of newspapers or broadcast by a given number of radio stations or
TV channels which are selected for study are recorded.

Very significant in the realm of unit of analysis is the need to beware of error of ecological
fallacy. Error of ecological fallacy arises whenever we mistake attributes of a group for attributes
of individuals belonging to the group. This brings about confusion in generalization on units of
analysis in such a way that we might erroneously “draw conclusion about individuals based
solely on the observation of groups” (Babbie, 2001, p.100). Within the context of content
analysis design, an investigator ought to be painstakingly mindful of the specified units of
analysis in his/her research design. If a sentence is specified as the unit of analysis, it should not
be confused with a word. Likewise, a sentence should not be made to replace a whole paragraph
as the unit of analysis if the latter has been specified as the unit of analysis in the research design.
Even though “the pattern observed between variables at the level of groups may be genuine, the
danger lies in reasoning from the observed attributes of groups to the attributes of the individuals
who made up of those groups when we have not actually observed individuals (Babbie, 2001,
p.100).”

Attitudinal directions
An attitudinal direction of media contents specified for analysis has to do with the nature of the
massages which might be positive, negative, or neutral. In other words, there is the need to
classify the units of analysis in terms of the directions in which they flow. Media messages on a

191
 Research Methods In Communication & Media Studies

controversial issue would normally take a side to such issue. It is therefore normal that a
researcher categorize reports on the issue or event relative to the associated meanings. Meanings
and inferences made out of a given text would guide the content analyst on how to appropriately
locate identified units of analysis within the relevant categories of attitudinal directions (Frey,
Botan, and Kreps, 2000, p.237). Categorising media contents against the backdrop of their
attitudinal directions as well as labeling the directions depends largely on the analyst‟s sense of
creativity, preference, and the nature of issues involved. Contents‟ attitudinal direction could be
classified as favourable, unfavourable, or neutral. It could also be categorized as soft, harsh and
neutral.

Content analysis may also be conducted from the perspective of examining texts. Qualitative
textual analysis deals less with numbers. It depends largely on versatility of the researcher in the
knowledge of the English language, semiotics, and rhetoric all of which enable deep probing of
textual denotations and connotations.

Procedures of Collecting Data for Quantitative Content Analysis

Gathering of data is one of the five components of content analysis as a research design. The
components are (1) formulation of research problem or hypothesis, (2) sample selection, (3) data
collection, (4) data analysis, and (5) interpretations of results. The first three components are
involved in the data gathering procedures, which is the concern of this chapter.

Problem statement assumes the initial position in the scheme of data gathering because the
statement provides the clue to what to study. The statement identifies the media to focus for
inquiry and suggests the actual content category of the media to be examined. In other words,
our problem statement tells whether we should isolate radio or television or newspapers for our
inquiry. If TV or radio is the identified medium in the statement, the genres to be content-
analyzed would also be identified. The statement would tell us whether we are to content-
analyze documentaries or dramatized entertainment. Likewise, if newspaper is identified in the
statement, the news items to be studied would be mentioned. The statement would tell us
whether we are to content-analyze editorials or straight news or combination of both. This means
that the problem statement sets the limit on relevant information to collect.

Characteristically, a functional research problem specifies and justifies the subject-matter in

focus, the symbolic content in this case, relative to existing theoretical explanations in the
literature. Such problem statement is practically rearticulated in form of a set of research
questions or in form of a set of well-defined hypotheses or in form of combination of both. When
combination of both becomes necessary, research questions stream out of the problem and
hypotheses are formulated based on theoretical perspectives and practical problems. Well-
defined research questions or hypotheses guide the construction of relevant content categories,
which provide relevant data for analysis (Khan & Khan, 2006).

192
 Research Methods In Communication & Media Studies

Distinctively, hypotheses are normally derived from conceptualisation of the main concepts
identified in the specific theoretical literature that is reviewed for our study. Conceptualisation is
the process of explicating a set of concepts from their higher level of abstraction to their lower
level of concrete meanings. It is this level of concrete meanings that ease formation of
“observable lower variable at operational level. An operational hypothesis is actually a
declarative statement relating two or more operational definitions” (Khan & Khan, 2006).

Sample Selection
Sample selection is the second component and the second overall step to data gathering. Sample
selection has to do with identifying the study‟s population, sample frame, and sample size. The
sample size is normally drawn from the units that make up the study‟s population, which could
be editorials or similar news genres in newspapers, and which could also be soaps, dramas,
movies, documentaries, etc. on specific TV channels. The content categories that are identified
as units of analysis making up the study‟s population are expected to have been considered
through operational definition of the key variables in the study to be conducted. Therefore
sampling will be based on such units of analysis.

If newspapers are to be studied, decision should be made on the following: (a) specific
newspapers to use, (b) specific editions of the newspaper selected, (c) the exact messages to
focus in the editions specified, and (d) the precise units to observe for analysis in the messages.
If TV is chosen for the study, the following should be decided: (a) TV channels to watch, (b)
genres of TV message to watch, (c) duration of televising the messages, (d) actual time of the
day for watching, and (e) period of the study –numbers of days.

Representative sample size of the text identified is selected through probability sampling
procedure. It should be ensured that a relatively large sample size is derived from the study
population. To arrive at a meaningful size of sample that can be generalized to the population, a
combination of probability sample techniques can be made. We can combine simple random,
systematic, stratified, cluster, and multi-stage probability sampling procedures.

Hypothetically, to conduct analysis on newspaper coverage on Boko Haram‟s insurgency over a

period of six months, we may decide to classified newspapers to be studied into clusters on the
basis of when they are published, in this case daily newspapers and weekly newspapers. We may
thereafter group each of the two clusters on the basis of circulation, namely state and national
newspapers.

We may now opt for two daily newspapers and two weekly newspapers out of which we could
select a proportionate number of editions. For instance, we may select two editions from each of
the identified daily newspapers on weekly basis, and selection of one edition of the weekly
newspapers also on weekly basis.

193
 Research Methods In Communication & Media Studies

We may decide to systematically identify Monday and Friday editions as the clusters for the
daily newspapers, and only Saturday or the specific day in which the weekly newspapers are
published as their cluster. Definitely, we would be having 48 issues of the weekly newspapers
and 24 issues of the weekly newspapers in the specified six month period. Systematic random
selection of Monday and Friday can form the sample frame for our hypothetical study if we
isolate editorials as units of analysis. The reason is that one issue cannot have more than one
editorial.

However, if other genres, say news stories and interviews for example, would be parts of units to
be studied, we would have to move further in the scheme of randomization. Therefore, we may
decide that stories on the front page, interviews regardless of page, and editorials regardless of
page, should form parts of the sample frame. We may decide to apply systematic random
sampling procedure to creating the sample frame. We would therefore obtain a list of all the
published news stories in the front pages, all the published interviews and editorials in the
selected newspapers in the days specified and within the six months stipulated for the study.
Apart from editorials, there may be two different stories on the insurgency on the front page of
an edition of the selected newspapers. Likewise, there may be two or more interviews pertaining
to the issue in an edition of the newspapers. Thus we may decide to select one news story with
big headline published on the front page of each of all the editions specified for the study. We
may also decide to select only one interview and we may specify the interview to be selected as
the one which is the first in the sequential arrangement of pages in every edition of the
newspapers identified for the study. Therefore the sample size for the study would emanate from
the sample frame on the basis of the aforementioned probability sampling procedures.

Data Collection
Data collection involves four fundamental things: determining unit of observation (unitizing);
creating code categories; training coders; and coding units into categories (recoding).

Designing the Coding Sheet / Training the Coders

Coding sheet is the tool for data gathering. Coding sheet is used for recording each unit of
analysis relative to number of occurrences and categorization of genres, themes, and attitudinal
direction or tone. It should be noted that the key items in the coding sheet emanate from the
conceptualization (Babbie, 2001 pp.310-311).

Coding sheet construction for a study would depend on the study‟s focus. Key items that are
identified for the study must be adequately taken care of in the coding sheet. Categorization of
the items or themes should be numerically indicated. The process of coding involves conversion
of raw data into a standardized form (See Figure 1.1).

Most significant along with coding sheet construction is the training of coders. The coders are
research assistant who would help enter codes into appropriate sections of coding sheet. The

194
 Research Methods In Communication & Media Studies

coders should be made to identify the themes and other core elements of the coding sheets for
appropriate entry.

Strengths and of Weaknesses of Content analysis

The scope and limitations of content analysis application to investigating researchable
phenomena can easily be understood through our knowledge of its strengths and weaknesses.
The underlying assumption is that content analysis as a method of inquiry cannot satisfy all
research needs. This speaks the reason data from manifest content of communication should be
complemented with data from survey or field observation.

Strengths
1. Content analysis is an obtrusive method. It requires no interaction with people from
whom the messages for research emanate. Media messages in the form of written or
spoken words are readily constant to be used for analysis. This provides open access to
contents to be investigated.
2. The stand-still nature of communication content gives researchers the opportunity to
examine past and current events that have been documented in printed forms as well as
those that have been recorded in form of static, motion pictures, and sounds. Therefore
content analysis renders objective account of events, issues, and themes to inquiry.
3. Content analysis can give reliable and replicable quantitative data. This can be enabled by
availability of a standardized framework (e-g. in grid format) which can be applied to
analysis of different types of media.
4. It is inexpensive and relatively easy to gather data for content analysis.

Weaknesses
1. The fact that an investigator select, categorize, record, and code data on the basis of his /
her interpretation and judgment makes objectivity of content analysis somewhat
questionable. For example, categorizing specific patterns of behaviour in a soap opera or
movie as „aggressive‟ will depend on how aggression is conceptualised by the
investigator. Therefore data collection could be compromised if there is problem with the
initial conceptualization of the term.
2. Coders‟ bias can affect data interpretation and coding. Two different coders might
differently interpret same sets of content. This is likely to affect the validity and
reliability of both the data and coding sheet. Such situation demands that systematic
3. comparison should be made of the two pools of data so that uniform interpretation and
coding could be effected. This could be time-consuming and financially expensive.
4. In content analysis, it is difficult to be certain that sample studied is representative of the
research population.
5. It is hard to make correct inferences on the basis of content analysis.

195
 Research Methods In Communication & Media Studies

Survey
Survey research is the most commonly used scientific method of investigating social phenomena
and how they interrelate to influence human society. It is used to systematically collect
information about a defined social problem of an identified community or a group of people
living in a specific area or sharing similar life experience, with a view to understand such
problem and develop mechanisms to solve it (Farooqui,2006). Survey research is normally
focused on large and small populations, which are also referred to as universe. But rather than
collecting information from the whole population which might be characteristically diverse and
relatively complex, survey research demands selection of smaller representative samples to be
studied from the whole population.

There are three fundamental types of survey that are common with social investigators: factual
survey, attitude survey, and explanatory survey. The factual survey is used to collect descriptive
information about a population of people. National population census of any country falls within
this category of survey. The factual survey can be used to investigate sociological problems of
specific population of people.

The attitude survey is designed to investigate subjective stance of individuals on events, issues
and trends. A typical example of attitude survey is opinion polling on political issues and events
such as elections, voting patterns, and government‟s policies and activities. Market research that
is conducted to know consumers‟ attitudes towards products also belong to this category of
survey. Likewise, some perception studies can be categorized as attitude surveys.
The explanatory survey is the most sophisticated and most comprehensive in that it takes an
inquiry beyond description. It is a survey that involves theories development through hypotheses
testing. It is also the survey type that is used to test validity of existing theories and to generate
new hypotheses from such theories in order to examine patterns of statistical correlation among
variables. A typical example of the explanatory survey can be seen in a study conducted
purposively to know why a section of a community prefers to vote for a particular candidate in
an election. In such a research situation, inquiry may go beyond seeking to detect whether there
is correlation between two variables to inclusion of a third variable or even a set of third
variables as a method of elaboration.

Designing the Survey

The purpose of information collection is to describe, compare, and explain opinions, attitudes,
and behaviours of people that are sampled. To collect accurate and relevant information, survey
research demands that the research should be designed against the backdrop of a well-defined
problem and objective or a set of objectives which should be further articulated in form of a
research question or a set of research questions. The survey design right from the beginning
should also include conceptualisation of specific concepts theoretically dealing with the
identified problem and operationalisation of such concepts, leading to a reliable, well-
constructed, and valid data collection instrument. The basic assumption here is that the

196
 Research Methods In Communication & Media Studies

instrument, specifically a self-administered questionnaire will normally emerged from

operationalisation of concepts tapping constructs that are processed through composite measures
of a set of items or indices or indicators.

Meanwhile, when designing a survey research, consideration should be given to the theoretical
foundation on which the survey will be based. Essentially, two basic reasons seem to call for new
research: first, the “view that existing scientific knowledge must always be provisional”; second,
the need “to reduce the ambiguity of much research evidence” (de Vaus.2001, p.11). The
significance of factoring in the theoretical source of the survey in focus emanates from the need
to be skeptical of the existing evidence: instead of investigating towards evidence that is
consistent with the theory, it is more appropriate to come up with findings that provide a
compelling test of the theory. The two related strategies for putting a theory into an acid test are
elimination of rival explanations of the existing evidence and search for new evidence that can
disprove the theory. The quest for plausible alternative hypotheses is well-articulated by de
Vause (2000) when he argues that:
As well as evaluating and eliminating alternative explanations, we should
rigorously evaluate our own theories. Rather than asking „What evidence would
constitute support for the theory?, ask „What evidence would convince me that
the theory is wrong? It is not difficult to find evidence consistent with a theory. It
is much tougher for a theory to survive the test of people trying to disprove of it
(p, 14).

He furthermore argues:
The more alternative explanations that have been eliminated and the more we
have tried to disprove our theory, the more confidence we will have in it, but we
should avoid thinking that it is proven (p,15).

Therefore, a good survey design should go beyond the simplistic propositions that tend to
explain how two variables correlate. Such design should include right from the outset other
necessary factors or variables that can spell out how the predictor and outcome variable might be
related. This emphasizes the need for a clear-cut assumption of a more complex causal model
that could fully reveal new facts about why a phenomenon predicts a specific outcome.
Providing a focus and a framework for the survey, the fuller model will help guide formulation
of problem, facilitate suitable composition of research questions, and enrich contents of in-depth
interview or contents of a self-administered questionnaire or similar tools for data collection.
Putting the aforementioned in context, de Vaus suggests that it would be helpful for a researcher
to ask four key questions:
1. What am I trying to explain (i.e. what is the dependent variable?)
2. What are the possible causes (what are the independent variables?

197
 Research Methods In Communication & Media Studies

3. What causes will I explore?

4. What possible mechanisms connect the presumed causes to the
presumed effects (what are the intervening variables?) (pp, 20-21).

Formulation of Research Problem

At the heart of survey research design is the formulation of a research problem because it tells us
what to study and thus guides us on the nature of information to gather. Problem statement tells
the reason for the research. The investigator should be able to state categorically what is it that is
optical to our understanding regarding the issue identified for inquiry and why such phenomenon
deserves to be studied. To do this, the investigator is expected to fully understand the nature of
the problem and the background from which it emanates. The statement, which supposes to be
made in a sentence, should therefore be able to capture the scope of the relevant data to collect.
In order to generate answers, the problem also should be articulated into research question or a
set of research questions.

Concepts and Conceptualization

Conceptualization plays a vital role in the scheme of data gathering. The process of
conceptualizing starts with identifying a concept or a set of concepts to be studied in our
research. A concept is defined as the abstraction from reality (Kerlinger 1973), indicating that
concepts emanate from real happenings in society. A concept is commonly stated in abstractive
terms and at a higher level of abstraction, not specifying a particular referent among others that
share its general meaning. For example furniture is a concept which embraces related but
different objects for human convenience such as chairs, tables, cupboards. Therefore, a concept
has to be defined from a specific dimension in order to be clearly meaningful. To derive a
concrete meaning from a concept, the concept has to be delineated from its higher level of
abstraction to the lower level that gives a clearer picture. It is at the lower level of abstraction
that concepts become measurable.

In mass communication and related social scientific research, scholars deal with concepts such
as media credibility, bias, exposure, perception, and social reality. Each of these concepts can be
defined or explicated from different dimensions of their meaning. Definition and explication of a
concept will depend on theoretical framework of the study involved because “the meaning and
significance of some concepts are interpreted in the context of the theory” that guides the study
(Farooqi, 2006).

Conceptualization commences with delineation of concepts from their higher level of abstraction
to the lower level abstraction. Social scientists translate the abstract concepts identified in
theoretical review into something concrete, linking the concepts to empirical realities and
making them to be measurable in research. When they do this, social scientists bridge the world

198
 Research Methods In Communication & Media Studies

in which they operate: the world of ideas and theories and the world of observations and
experiences (Nachmias& Nachmias, 1990 p47. Conceptualisation is the method social scientists
used to tell their colleagues and readers of their research works what they mean when they use a
concept (Idid, 2006).

Before defining a concept, the researcher must have read about it and how it has been discussed
and debated in the relevant theoretical literature. With that, the researcher will have two options:
he may accept the existing conceptualization if it is line with his or her objective of study; he or
she may come up with a totally different conceptualization or better still define the concept in
focus from a different dimension or from different anchor point.

Most significantly is the fact the dimension from which a concept is defined will affect any
study‟s results and generalization of the results. For instance social reality, as a concept, is
explored at two levels of abstraction, one leading to the other. At the individual level of
abstraction, it refers to the worldview held cognitively by individuals. At the societal level, it
implies a perception of the world commonly shared in the society (McLeod & Chafee, cited in
Eveland & Glynn, 2008). Therefore, if social reality is conceptualized in a given research from
dimension of a perception of the world commonly shared in the society, the research findings
should go in line with this conceptualization; not in the dimension of the worldview held
cognitively by individuals. Idid (2006) points out how “aggression” has received differing
conceptualizations by communication researchers. He says:

One definition is “physical abuse,“ another would term aggression as “physical

abuse aimed to lower the self-esteem of the other.” Another definition takes
aggression as both the physical and verbal abuses aimed to lower the self-esteem
of the other. There are different meanings in the definitions or conceptualisation
of the concept. At least two differences are evident here one includes only
physical behaviour while the other includes both physical and verbal behaviour
that lower the self-esteem of the other...

Another example can be seen in how media exposure is conceptualized. The difference should be
noticed in a study that conceptualizes television exposure as time a viewer spends watching
television and another study that conceptualizes television exposure as a viewer’s ability to
comprehend television messages. The difference in the two conceptualizations is very glaring:
the first conceptualization emphasizes time that a viewer spends watching the medium; the
second one emphasizes ability to understand the images on the screen. Editorial readership can
also be another interesting concept. Editorial readership can be conceptualized from different
dimensions; depending on the inquirer‟s research interest and the angle from which he/ she
would like it to be measured. Thus, conceptualization determines what information or data we
seek from our sampled population.

199
 Research Methods In Communication & Media Studies

Operationalization / Construct Measurement

Operationalization has to do with how the concepts or constructs are measured.
Operationalisation is a step beyond conceptualisation in that the latter help to produce constructs
as an entities on the theoretical level, whereas the former defines the constructs in terms of
indices or items or indicators on operational level. To an extent, concept operationalization plays
a significant role in construct validity, hence every construct has to be indexed on multiple,
diverse indicators given on scales (Judd& Kenny, 2006). That is to say that use of one-item
scales is often unreliable in construct or variable measurement (West, 2006).

An investigator measures a construct by asking respondents to answer a given set of questions in

form of providing a score on the index scale that contain indicators which measure the construct.
Specifically, the investigator may ask the respondents to tick an option (on a Likert-type scale)
ranging from strongly agreed to strongly disagree with a set of statements otherwise called items
or indicators. The items or statements should be appropriately constructed, tapping scope and
intensity dimensions of the construct on the scale.

The composite measures so produced by scales and items are statistically examined through
reliability test. The essence of reliability test is to examine internal consistency of the items that
make up the scale. In other words, reliability test reveals the extent to which the items hang
together to produce a specific variable. Reliability test on SPSS replaces the old fashion test-
retest reliability, which is assessed by administering a questionnaire to the same community of
people on two separate times as well as calculating the correlation between the two scores
obtained (Pallant, 2007, p7). With reliability scale on SPSS, internal consistency of items is
measured by Cronbach‟s alpha whose values range from 0 to 1. While some statisticians do
agree on .5 as a minimum value for items reliability, Nunally (cited in Pallant, 2007, p.7)
recommends a minimum of .7 Cronbach‟s alpha. However, the alpha values depend on the
number of items in the scale. Thus where the number of items is fewer than ten, Cronbach‟s
alpha can be somewhat small. Reliable as the reliability scale might be, it cannot tell whether the
items examined do really procure one dimension. This situation provokes the investigator‟s
curiosity to explore factor analysis, a second alternative to reliability of measures. Factor
analysis normally reveals two, three, or four dimensions or components into which items in a
scale could be sub-divided, leading to the birth of new constructs or variables.

Data Collection in Action: the Elements

All the aforementioned core procedures in survey research design, ranging from problem
formulation through operationalization of constructs are in preparation for data gathering. Data
gathering is the crucial phase of the research process and, thus, requires serious precautions in
considering the procedures namely, research population, sample frame or selection, sampling
procedure, sample size, unit of analysis, and data collection instrument(s).

200
 Research Methods In Communication & Media Studies

Research Population
Research Population is the universe or population of people from which the sample size to
be investigated is drawn. The study population must be identified relative to the problem
statement.

Sample Frame
If, for instance, the research theme has to do with readership of newspaper editorials on Boko
Haram insurgency among undergraduates in Kwara State University, it would be unnecessary to
identify the whole undergraduates in the University. Hence, the needs for a sample frame. A
sample frame can be obtained through creation of a list of students from specific departments in
specific faculties or colleges in the University. Selection of sample can easily be understood
through the use of sampling procedure.

Sampling Procedure
Probability sampling techniques are mostly applied to sample selection. In the case of selecting
specific colleges in the higher institution, simple random sampling method could be applied to
selection of three out of the five colleges in the university. Simple random could also be used for
selection of two departments from each of the three colleges. Thereafter, it could be arbitrarily
decided that an exhaustive list of graduating students should be obtained from each of the six
departments simple-randomly selected.

Once the lists have been acquired, systematic random sampling method could be applied to
selection of students from each of the lists. Arbitrarily, it could be decided that every fourth
person on each of the lists should be picked. If every fourth (4th) person would be systematic-
randomly picked from a list of fifty (50) graduating students from the Mass Communication
department (a department from ICT College), it logically speaks that only twelve (12) graduating
students from the department would be selected. Similarly, every fourth (4th) person would be
systematic-randomly picked from a list of sixty five (65) graduating students from the Computer
Science department (another department from ICT College), leading to selection of fifteen (15)
graduating students. Therefore, the ICT College would contribute only twenty-seven (27)
students to the total number of students that would make up the sample size.

Sample Size
Sample size is the total number of people drawn from the research population. Therefore, the
total numbers of students who are systematic-randomly selected from the identified sample
frame would form the sample size for the hypothetical research on newspaper editorials
readership among undergraduates in Kwara State University. Data would be collected only from
the actual students included in the sample size, who will serve as units of analysis.

Formation of Data Collection Instrument

Self-administered questionnaire is the most systematic instrument for data gathering in that it is

201
 Research Methods In Communication & Media Studies

amenable to computer-based entry and analysis. Besides, most of its elements and contents are
conceptually and operationally structured. Specifically, indicators or indices that are used to
measure the constructs or variables are mostly generated by panel method. Generation of items
by panel method takes effect whenever a group of people who are conversant with a
phenomenon are asked to individually suggest a set of diverse statements that interpret a
construct. This is normal in line with the dictum that says: “two heads are better than one”.

Instrument Pre-testing
At the initial formation stage of a questionnaire, it is likely that it has the problem of ambiguous
wording, and related errors that might affect data reliability of data. In order to detect and solve
such problems, a pilot study or two separate studies ought to be carried out for the questionnaire
pre-test before full-blown data collection. Pilot studies are normally conducted with some 30
respondents from the research population. Conventionally, respondents used for questionnaire
pre-test should not be part of the sample size.

Most importantly at the pre-test stage, the investigator should look at the frequency distribution
of responses, perhaps some items or indicators might appear constant in terms of responses or
lack heterogeneity of responses. Such problem is normally indicated in the frequency distribution
of items that is displayed in the SPSS data output. At this instance, such items should be dropped
or reworded appropriately.

Thereafter, a reliability of the items should be examined with the use of scale of reliability
analysis on the SPSS. Reliability of items is defined by Cronbach‟s alpha that has been discussed
under conceptualisation (please, revise conceptualization above). In addition, each items strength
or weakness can be seen in the inter-item correlation table also in data output displayed in
reliability analysis scale. Strength of an item is indicated by correlations coefficient of .30 or
above. It is advisable to drop item with lesser coefficient. This will enable the coefficient of the
remaining items to rise.

For the purpose of reliable data collection, delivery persons have to be properly trained because
they serve as research assistants. In ethnic and religious sensitive research, relevant persons who
are properly trained for the research should distribute and retrieve copies of the questionnaire
from respondents.

Strengths and of Weaknesses of Survey Design

That survey research design has much relevance in the cause of investigating social and
behavioural phenomena is an unquestionable fact. Its universal appeal to the search for answers
to social scientific-related problems is an attestation to its validity as a way of inquiry and
discovery. All the same, like other research designs, survey has its merits and demerits relative to
research goals and nature of phenomena to be investigated.

202
 Research Methods In Communication & Media Studies

Strengths
1. Through its ability to describe characteristics of a relatively sizeable group people, survey
research provides in-depth knowledge of a larger population to which such group
belongs. With information rendered by the investigated group, answers are analytically
provided to the problem formulated for research. Specifically, problems such as socio-
economic conditions, good or bad, are highlighted and identified for necessary policy
actions to be taken.

2. Government, agencies, firms, and organizations that have interests in bringing about
change in a community of people can unravel what the people need and how their
problem could be solved through the application of survey design. Impacts of
development programmes executed for the people can also be monitored and evaluated
with the use of survey. Longitudinal survey is practically the most relevant method to
study the trends of change in most rural communities.

3. Data generated through survey design are often relatively valid and reliable not only
because of the standardized ways in which self-administered questionnaire is constructed,
but also owing to the fact that investigators have direct contact with people and their
explicit realities. Besides, space created for unstructured response and comment provide
investigators with knowledge of the actual situation of things regarding the subject-matter
in focus.

Weaknesses
1. Selection of sample size for a survey research can be tricky. Defects in sampling might
arise out of inclusion of persons not really conversant with subject-matter of the study.
Therefore some responses to survey interview or questionnaire could be artificial as they
may tend to satisfy the investigators. This is to say that sample selection is prone to
sampling errors, which may likely affect findings. Kerlinger convincingly observes
that: “there is always the one chance in twenty or a hundred that an error more serious
than might be caused by minor fluctuations of chance may occur” (Kerlinger (1973,
p407).

2. The problem so identified with sampling selection is likely to affect the validity and
reliability of data- gathering instrument, specifically self-administered questionnaire. The
raises an eyebrow about the standardized nature of the questionnaire. Standardization of
the instrument is significantly observed in order to ensure its suitability to all
respondents. In the cause of so doing, there is the probability of omitting information
appropriate to many respondents, rendering the whole process to be superficial (Babbie,
2001, p.268). In some cases, respondents may skip certain questions in the self-
administered questionnaire, leading to incomplete response. An exception to this might
be the use of face-to-face interview which allows for open-ended questioning. In this

203
 Research Methods In Communication & Media Studies

case, interviewees are given the chance to take investigators to the depth of the
issue in focus. But face-to-face survey is good only for data-gathering from
smaller sample size; its use could be tedious, cumbersome, and expensive for a
large sample size. Its weakness lies in its probable inability to source responses to
highly controversial issues, for instance some interviewees might be afraid and
reluctant to express their takes on sensitive issues and events. Conversely, the
same persons may be ready to anonymously respond to such questions in a self-
administered questionnaire. The weakness so identified justifies a triangulation of
both instruments in a survey research design.

3. Survey research design has been confirmed weak in dealing with social life of
respondents. Both face-to-face interview and self-administered questionnaire may
be inadequate to capture “the total life situation in which respondents are thinking
and acting” (Babbie, 2001, p.268). Survey research, whether cross-sectional or
longitudinal, is designed to take care of special events and issues; it does not take
place in a vacuum. Thus respondents perceive research occasions as
circumstantial and are likely to supply suitable information that is far removed
from reality study. In such research situation reliability and validity of research
findings become questionable (Farooqui, 2006). This justifies the need to
combine survey research design with field observation research design wherein
direct observation can be modified as field conditions warrant (Babbie, 2001,
p.268). There and then flexibility of field observation complements the
unchanged nature of survey. The fixed nature of survey is glaring in the fact
variables not included in the initial stage of the study design cannot be included
when later discovered that it is necessary. It should be recalled that variables are
constructs derived from conceptualisation and measurement.

Conclusion
The discussion of both quantitative and qualitative data gathering methods in this chapter
reveals that each method can only provide a cluster of information for researching a
multisided social phenomenon. If the objective is to explore a large sample size for the
purpose of examining correlations between specific variables, collection of qualitative
data is appropriate.
On the whole, social research demands that attention should be paid to the topic of the
study and the focus of the problem formulated, research objectives, and research
questions. The problem should emphasize the rationale for the methods. In the research
proposal, the phases in which the two separate data would be gathered should be clearly
stated because their explanation forms a part of the findings discussion in the final
reports.

204
 Research Methods In Communication & Media Studies

References

Arensberg, C. M. (2006). The community study method. In de Vaus, D. (ed.), Research design,
Vol. IV (pp. 57-81). London, SAGE Publications.
Babbie, E.(2001). The practice of social research. 9th edition. Australia, Wardsworth.
de Vaus, D.(2001). Research Design in Social Research. London, SAGE Publications.
Deacon, D., Pickering., M., Golding, Murdock , G. (1999). Researching communication.
London. Arnold.
Eveland, W. P., Jr., & Glynn, C. J. (2008). Theories on the perception of social reality. In W.
Donsbach & M. Traugott (Eds.), The SAGE handbook of public opinion research (pp.
155-163). London: Sage publications.
Farooqui, J. (2006). The survey method in sociological research. In Hussain, M. Y. (ed.).
Towards developing an integrated research method in human sciences (pp. 47-73). Kualar
Lumpur, International Islamic University Research Methods in Social Sciences.
Frey, L. R., Botan, C.H., & Kreps, G.L. (2000). Investigating communication: An introduction to
research method. Boston, Allyn and Bacon
Idid, S. A. (2006). The role of theory in research. In Hussain, M. Y. (ed.). Towards developing
an integrated research method in human sciences (pp. 17-31). Kualar Lumpur,
International Islamic University Research Methods in Social Sciences.
Kerlinger, F. (1973). Foundations of Behavioural Research. 2 edition. New York , Holt.
Khan, F. R & , M. A.(2006). Quantitative content analysis as a method of research in Mass
Communication. In Hussain, M. Y. (ed.). Towards developing an integrated research
method in human sciences (pp.141-167). Kualar Lumpur, International Islamic University
Research Methods in Social Sciences.
Nachmias, C., & Nachmias, D. (1990). Research methods in social sciences. London, Edward
Arnorld
Pallant, J.(2007). SPSS survival manual: a step by step guide to data analysis using SPSS for
window. 3rd edition. England, Open University Press. McGraw-Hill.
West, S.G. (2006). Validity. In de Vause, D. (ed.), Reseach design, Vol. 1 (pp121-140). London,
SAGE publications.

205

View publication stats