Network of Scientific Journals of Latin America, the Caribbean, Spain and Portugal

Scientific Information System

Carmen Lafuente Ibáñez, Ainhoa Marín Egoscozábal

Research methodologies in social sciences: Phases, sources and selection of techniquesSchool
of Business Administration Journal, No. 64, September-December, 2008, pp. 5-18,EAN
University
Colombia

Available at: http://www.redalyc.org/articulo.oa?id=20612981002

Business Administration School Journal, ISSN (Print

version): 0120-8160 [email protected]
EAN University
Colombia

More information about the article Magazine page

How to cite? Full fascicle

www.redalyc.org
Non-profit academic project, developed under the open access initiative
 EAN
Magazine

METHODOLOGIES OF THE

RESEARCH IN THE
SOCIAL SCIENCES:
PHASES, SOURCES AND SELECTION OF TECHNIQUES
Carmen Lafuente Ibanez*
Ainhoa Marin Egoscozabal**
SUMMARY

This article offers a general review

1. INTRODUCTION

Scientific research is a task that sooner or

of the methodology and different phases later clashes with the development of an
of the scientific research academic career and sometimes also with
process, as well as the selection criteria
and description
the development of professional work. This
of the different article reviews the different stages that
qualitative and quantitative techniques make up the scientific research process,
that can be used in scientific explains the importance of setting
research applied objectives, defining variables, establishing
to the social sciences. the hypothesis and, above all, selecting
appropriate techniques to achieve the objectives
For this reason, it is of special interest that led to the start of the research.
to those in the process of designing
and carrying out academic research
, and to
those interested in the selection and usefulness
of the different existing techniques,
the contrast between hypotheses and the treatment
of information for research.
Scientific
researchMethodologyTechniques * PhD in Economic and Business Sciences from the National
University of Distance Education. Professor of Quantitative
Methods at Nebrija University. Member of the Nebrija –
Tecnológico de Monterrey Research Center for Family
KEYWORDS Business and Entrepreneurs. He directs research projects that
lead to doctoral thesis, focusing on family business and
entrepreneurship. Carries out market research and consulting
projects for public bodies and private companies.
** PhD in Economics and Business Administration from
UNED. Professor of Applied Economics at Nebrija
University.
This article was submitted on August 4, 2008 and its publication was approved by the Editorial Committee Member
on August of the African Studies Group of
16, 2008.
the Autonomous University of Madrid. An expert in
economic and trade relations of the European Union,
she has been invited to the European Parliament and
has been a visiting researcher at the European Centre
for Development Policy Management in Brussels.
EAN

The scientific method can be defined as the set of This article is about a general review of the methodology and
different phases of the scientific research process; Moreover, it
tactics used to construct knowledge. These are the
shows the selection criteria and description of several
steps and instruments that lead us to explain qualitative and quantitative techniques that could be applied to
phenomena, or to establish relationships between a scientific process in Social Sciences.
facts. The tactics used are diverse, although it is Therefore, this study is of
common to distinguish between two types of methods: special interest for those processes which
the deductive method and the inductive or empirical are involved in the design and realization of
an academic research process
method. and for those professionals interested in the
selection and application of different existing
The deductive method is part of the so-called rational techniques to contrast
the hypotheses and the information management
logic and consists of: starting from some general in research processes.
premises, reaching inferring particular statements. If it
happens that these initial general conceptions are not
demonstrable (axioms), the method will then be
axiomatic-deductive.
The inductive or empirical method consists of creating general statements from experience, starting with the
observation of a phenomenon, and repeatedly reviewing comparable phenomena, to establish by inference laws
of a universal character. In this sense, it is possible to affirm that both types of methods follow inverse processes,
where the tactics used go from the general to the particular (deductive method), or from the particular to the general (inductive
or empirical method).

KEY WORDS
The scientific method provides the means to achieve an objective, but it does not
provide the objective itself, which must be posed through the process discussed below.

Scientific Research

Methodology
Techniques
 EAN
Magazine

In practice, one cannot speak of a purely hypothesis that will be validated or refuted. After

inductive or deductive research strategy. choosing the topic that frames the problem, the

Synthetic methods are usually used, a mixture of process continues with a review of the
both methods. publications and references on the subject.
Note that in these early stages, a research
2.2 The research process: problem,
objectives and hypothesis process in the strict sense has not yet
begun, however a fundamental step is taken, the
The first task that must be faced in the development definition of the research objective, as outlined in
of an investigation is the statement of the problem, the following table:
which would be the area that would give rise to the

TABLE 1
Phases in the research process

Two types of objectives can be The origin of the hypothesis is in the problem
distinguished: the general objective, which posed, and there is no method to reach it. It
must reflect the essence of the problem to may appear as a result of inspiration, genius,
be investigated and which is somehow reading, for example, but it always has its
expressed in the title of the doctoral thesis. origin in a problem. On the other hand, the
On the other hand, there are specific objectives (both general and specific) must
objectives, which can be defined as the be clear and precise, because in this way
steps taken to achieve the general objective. the risk of making mistakes in the selection
Consequently, the specific objectives should of techniques to achieve them is lower.
not only not hinder the main objective, but
should complement it and allow the research
to be directed towards its culmination.
The research process formally begins
with the formulation of the hypothesis.
UNIVERSIT
Y
EAN
»ee
EAN
Magazine

2.3 Variables in the research process Variables are of great importance in the
scientific method because we classify and
Variables are observable qualities, group them, we relate and study them, and
properties or features of the elements of the we interpret them. Variables are involved in all
scientific study. Variables are susceptible to change phases of the process, so we must define them
and to being linked to each other, according to thoroughly and specifically.
various types of relationships.

TABLE 2
Data and techniques in the research process

Data processing and exploitation

---------------:—•--------------------------
Descriptive

Qualitative Quantitative Techniques

Techniques
♦Focus groups ♦ Descriptive measures
♦In-depth interview ♦ Statistical inference techniques
♦Delphi method ♦ Multivariate dependency techniques
♦ Multivariate interdependence techniques
 EAN
Magazine

2.4 Research techniques analysis of the data derived from the

judgment of a group of experts in the field we
From a scientific point of view, research is a are studying.
process that allows us to apply the
appropriate scientific method and techniques Qualitative techniques by themselves
to find an answer to the problem initially constitute a valid instrument for making
posed. The use of some techniques or decisions based on the forecasts of the
others defines the type of research we are group of experts, but in many studies they
carrying out. are complemented by the development of
quantitative research.
If we start from the basis that the essence of
research is the obtaining and processing of • According to the function that the
data, we can propose different types of research fulfills
research according to the following
classification criteria: To achieve the general objective of the research, we
need to achieve several specific objectives,
♦ According to the nature of the which makes it difficult to define a single line
information
of research. However, and although in most cases
they complement each other, we can choose one of
The nature of the variables is a factor to take
the following types of study depending on the
into account when choosing the type of
function they perform: exploratory, descriptive and
research. The nature and quantity of
information determines our choice. explanatory-descriptive research.
According to this criterion we can choose
between qualitative research and On many occasions, in order to develop the
quantitative research. topic of our main research, we find ourselves
in the need to carry out exploratory research,
A type of quantitative research will be very that is, a first understanding of the problems
useful when we can measure the variables under or conditions of the phenomenon under
study in an objective way and with a high degree of study. It is, therefore, a preliminary
precision. When we carry out this type of investigation prior to the definitive
research we must be sure that the investigation.
observation of the variables will provide us We carry out descriptive research when we
with specific numerical data that can be subject want to show the characteristics of a group,
to subsequent statistical treatment. In the
a phenomenon or a sector, through the
development of quantitative research, the use of
observation and measurement of its
Statistics is common and necessary practice.
elements. The information provided by a
descriptive analysis, in addition to being an
If the information we can access is poor in
end in itself, can be used as
data, or is dispersed, but rich in descriptions
of the variables, we will carry out qualitative
research. The characteristics of this type of -
research involve, in most cases, an intuitive
UNIVERSITY

(EA EAN
N)
Magazine

a starting point for the development of more describe and interrelate the research
specific research. variables for subsequent decision-making.
We will choose explanatory-predictive
research when our study is not limited to just When we carry out statistical analysis, we
measuring variables, but we also aim to must go through the following phases:
understand behavior and determine and
quantify the relationships between variables. • Data collection.
With this type of research we try to • Ordering and presentation of data.
understand the structure and factors that • Data processing and exploitation.
influence a variable, in most cases, with the
ultimate goal of establishing predictions Data collection, being an elementary task, is
about it. not without its difficulties. Our effort in
obtaining data should not be less than that
3.1 Statistical research
dedicated to carrying out other tasks or
activities. For our research to be useful, we
must ensure, to the extent possible, that the
data collected does not contain errors or

3. QUANTITATIVE TECHNIQUES
biases that could invalidate the study. For
the data to be true, we must ensure that the
variables have been thoroughly defined,
since the data are the direct result of
Throughout this article we have referred to observing said variables.
Statistics as the science on which the
research process to be carried out to The errors that we may encounter do not
achieve our objectives must be based. At this always have a theoretical meaning, but may
point, it is time to ask: What does statistical analysis be due to incorrect annotations, tabulations
consist of? And what techniques and instruments and/or classifications, or to the existence of
does it offer us for solving problems? values that are too large or too small,
considered as atypical data. In these cases,
Statistical analysis indicates the procedures the techniques of statistical analysis allow us
that allow us to present, summarize, to detect the existence of anomalous data; if
describe and compare a set of numerical in a frequency table or in the graph of a set
data. of data we observe that a value is repeated a
lot or a little, or we find maximums or
To understand the social and economic minimums, respectively, it may be due to the
reality we need a set of data, already existence of incorrect data. Once these data
existing or created by ourselves, over time or have been found, we must correct the error
at a given moment, of the variables specific and we can even eliminate those that are
to this discipline. atypical and that make the measures used
lose representativeness, justifying why they
In statistical analysis we find valid -
were eliminated and how this can influence
instruments to summarize the information in
the theory that we want to empirically
a few measures that allow us to know,
 EAN
Magazine

contrast. The survey is based on the preparation of a

questionnaire that we will formulate to the
In the organization and presentation phase, people, institutions, companies, etc. from
what we do is organize, tabulate and which we want to obtain information.
graphically represent the data in order to
summarize it so that it is more manageable. If we want to know the forms of
internationalization of companies in a certain
In the last phase, which we call data sector of the economy, the innovation
processing and exploitation, we determine processes, the degree of performance, we
the measures or parameters that summarize must use the survey as an appropriate
the amount of information, which gives us an technique to acquire data because we will
initial view of the most relevant hardly find information that fits our research.
characteristics of the set of data we are
processing. It is in this phase where we There is abundant bibliography on how to
formulate and contrast statistical use the sampling survey technique, such as
hypotheses, where we quantify the Azorín (1986), Santos and others (1999) or
relationship between the different defined Ortega (1992). Without going into detail
variables and where we make forecasts to about each of the steps to follow, it is
achieve the proposed objectives. necessary to be clear about some concepts
and the problems that we may encounter
when using the survey as a technique to
3.2 Data collection. understand the social and economic reality.
Sample survey
We must carefully and clearly define what
We have already talked about how to obtain the survey unit will be and its characteristics
data and the importance of keeping it error- (establishment, company, consumers) and
free. Now we are going to focus on learning delimit the scope of the survey, that is, the target
about the types of data and the different population, the territory and the time in which we are
techniques to obtain them and what we can going to carry it out.
work with.

To find the data we can turn to two types of

sources: secondary sources, when the data
comes from governmental, industrial or
individual institutions and primary sources
when we directly collect the data for the
development of our research.

If we need direct information or primary data to

achieve our research objectives, one of the most
commonly used data selection techniques is the
survey.
UNIVERSIT
Y
EAN
that
EAN
Magazine

The target population is made up of units The questionnaire is the instrument we use to obtain
that share one or more common - information through the sampling survey.
characteristics that we want to study. It is Although this technique is widely used, it is
usually quite difficult to define the population increasingly losing credibility due to the bias
and determine its size due to the differences in in the responses of those surveyed. If we
information depending on the source consulted. In create a questionnaire that is objective,
other cases, this problem arises because we do clear, precise and correct, we can reduce
not have data since the population we want biased responses. In this sense, it can be
to analyze is unknown. very helpful to investigate the existence of
questionnaires carried out on topics similar
In these cases, to delimit the population and or identical to those of our study. If we
determine its size, we must use the cannot find them, we must create our own.
information from the source that best fits our
research, and in some cases it will be To develop a good questionnaire, we must
necessary to make all kinds of assumptions. begin by determining the type of questions
we are going to use and establishing the
Once the population has been defined, we must
appropriate measurement scale (nominal,
consider whether it is possible to study it in its
ordinal, interval or ratio) for each of the
entirety. Its size may be very small, or we must think
population characteristics we want to know.
about whether we select a set of elements from it
since it is materially impossible to study the entire
When formulating questions, we must take
population or group. into account two aspects: on the one hand, the
computer software that we are going to use to
In most cases we will have to extract a process the information (it is easier to work with
closed questions than open ones, and to process
sample from the population and consider
single-answer questions than multiple-answer
whether the study will be limited to describing and
questions); on the other hand, it is important to
understanding the elements that make up the sample
remember that the data obtained is the basis on
without any further pretensions, or whether we want
which our research is based and that not even the
to apply the sample results to the entire population.
most sophisticated methodology is capable of
In this last case we will need statistical inference
providing good results if the initial information is false.
techniques so that the generalization of the results is For this reason, it is very useful to include questions
valid, starting by using those techniques that in the questionnaire that allow the validity of the
guarantee that the sample is representative of the answers to be detected.
population from a statistical point of view.

Whether the survey is directed at the entire

population or at a representative group of the
population, the use of this technique is
expensive in terms of money and time.
Choosing a research topic of interest to one
of the agents involved in the activity
analysed can be a way of obtaining funding.
 EAN
Magazine

Once the questionnaire is completed, we will procedures, but in this case Statistics does
choose the type of sampling that we will not provide us with techniques that validate
carry out to obtain the units to which it is the inference of sample results to the
directed. To achieve maximum accuracy in population. The second condition to
the selection of the sample, we must take guarantee the goodness of the choice is that
into account two fundamental conditions. we must take into account that the chosen
The first condition is that the generalization sample must be adequate to achieve our
of the results obtained in the sample to the objectives.
population must be valid. This requirement is
met, from a statistical point of view, when the We determine the sample size based on the
choice is made using random or probabilistic type of sampling chosen, the desired
sampling. We can also use non-random precision, the financial resources, the time
procedures, that is, directed or opinionated available, and the ease of selection.

TABLE 3
Operation planning and sampling plan design

Questionnaire design

• Nature of the
TOScope
information (quantity, complexity and

Territorial accessibility)

• Media (mail, telephone, personal, Internet, etc.)

• Types of questions Data
(open, closed, etc.)
Scope

Temporary
Sampling type

Scope

Population
Fountain. Own elaboration
 UNIVERSIT
Y
EAN
»ee
EAN
Magazine

3.3 Data processing and exploitation Usually, our interest will not be limited to knowing and
describing the characteristics of a sample, but we will
At this stage of statistical analysis, we consider which want to describe and know the characteristics of an
techniques and instruments we are going to use to entire population based on knowledge of the sample.
achieve the specific objectives proposed that will lead This will be the time to formulate all kinds of statistical
us to achieve the general objective. hypotheses and to choose the appropriate contrasts to
validate or refute them.
The proposed objectives and the nature of the
information determine the type of research we are The use of most of these statistical instruments is only
going to carry out according to the function it fulfills. possible when we have quantitative or numerical data.
It is common practice to apply these instruments to
When our study focuses on determining and describing
qualitative data, once they have been coded and
the key characteristics of social and economic reality,
assigned a number. This practice is not statistically
the frequency with which they occur and the degree of
correct, so we must be cautious in its application and
association between the different variables of interest,
subsequent interpretation of the results. It is important,
we can use statistical measures capable of
summarizing and synthesizing the information already as we already mentioned in the “information gathering”
stage, to assign each characteristic an appropriate measurement
organized and tabulated.
scale or numerical assessment, since in this way, we limit the
usefulness of the assigned numbers and the mathematical
In this synthesis process we look for a few numerical
operations that we can perform with them, and therefore choose
values that represent the sample data set. Position
the appropriate technique and instrument.
measures (means, mode, median, and percentiles) are
values that summarize all the available information.
In some cases, a descriptive study may be sufficient to
These measurements obtained from the sample
achieve our objectives; in others, we carry it out to
usually have a clear and concrete meaning and are
complement the development of explanatory-predictive
easy to calculate.
research.

With a position measure, and with a single value, we

We consider explanatory-predictive research
describe the global phenomenon under study, but we must
appropriate when we want to quantify the influence that
decide whether or not said value represents the data set. That is,
certain factors have on a characteristic that we call the
we need to measure the variability or separation of the
target variable.
values in the distribution. The calculation of dispersion
measures (range, standard deviation, Pearson
In this type of research, we may also be interested in making
coefficient of variation, etc.) allows us to calculate
predictions about the future behavior of the target
whether or not the position measures actually represent the -
variable.
distribution, and therefore, whether they describe the
characteristic of the phenomenon being studied.
Again, the nature of the data, quantitative or qualitative,
temporal or cross-sectional, forces us to choose
To complete the descriptive study, if the research
between the different techniques available.
requires it, we can establish what type of relationship
exists between the variables and to what degree they
Statistical regression techniques are a good choice
are associated, using correlation and association
when we have numerical data, regardless of whether
techniques.
 EAN
they are temporal, cross-sectional or panel. Using Magazine
developing a regression model.
regression techniques, we can identify, from the available
information, the dependency structure that best explains the
behavior of the target variable across the entire set of
variables or factors with which we assume it is related;
based on the theoretical knowledge that there is a
causal relationship between these and the former.
Regression models allow, on the one hand, to predict
the future value of a variable based on the knowledge
of another set of factors and variants, and on the other
hand, to find those factors or variables that best explain
the behavior of the target variable.

Once the causal relationship between the variables has

been established, we are faced with the problem of
choosing the mathematical equation that best relates
them. Knowledge of economic theory and the search
for empirical studies already carried out will help us in
choosing the ideal model to achieve our objectives. In
any case, we must know that a model of relationship
between variables is not better because it is more
complex.

The reliability of the results will depend on the validity

of the chosen model. There are techniques and
instruments that allow us to reject, modify or accept the
selected model. In addition to the dependency
techniques mentioned, statistics offers other
techniques applicable to those case studies where
there is no dependent variable, investigations where
there is no cause-effect relationship between the
variables, that is, cases where all the variables depend
on the rest. We refer to multivariate interdependence
techniques such as Factor Analysis, Cluster Analysis
or Correspondence Analysis. They are descriptive
techniques that simplify the complexity of a
phenomenon through the reduction of data and/or
variables.

The application of these techniques allows us to identify

underlying factors or dimensions that explain the correlations
between a set of variables, as well as detecting small groups of
variables that encompass the original set of variables
and/or data. These techniques are also very useful for
obtaining indicators and as a preliminary step for
UNIVERSITY

(EA EAN
N)
Magazine

4.
With the experience and knowledge of
QUALITATIVE TECHNIQUES specialists, we can identify problems, obtain
unknown information about some aspect, detect
We have described the quantitative disruptive factors in the change process and predict
techniques and instruments to develop the evolution of future trends considering the different
descriptive and/or explanatory-predictive possible scenarios.
research.
The most commonly used qualitative
Most of the techniques discussed require a techniques are the Delphi method and
set of cross-sectional and/or historical opinion methods. If we choose one or the
numerical data on the economic variable of other, the basic process we must follow is
interest, so we will not be able to apply them the following (Sancho et al., 2001):
if we do not have them. Furthermore, when
we want to make predictions, there may be - • Select the experts.
significant changes in the starting conditions,
or the pattern observed in the past may not • Conduct the interview.
be maintained in the future; this makes the
• Collect the answers.
use of historical data to predict future values
of the series questionable.
• Get results.
We will therefore use qualitative procedures
The selection of experts is, according to
in those cases where we cannot quantify the
Sancho et al. (2001), one of the crucial
information and when the data is not
stages, since the opinions given by
applicable or is not available.
professionals in the field constitute the raw
material of our work.
Most qualitative methods offer us techniques
for collecting qualitative information and
The interview can be done with an entire
predictions based on the opinions of a group
group simultaneously or individually. We can
of experts. With the instruments provided by
carry it out in a way directed by the
these techniques we obtain internal interviewer, based on a questionnaire, or
information about individuals using small develop it through free response from the
samples, therefore their use is advisable to obtain interviewee, or through a combination of -
prior information on an unknown topic to be
both. Once we have obtained the answers
developed later in depth. and depending on the type of data we
obtain, we can apply one of the quantitative
If we want to know the impact of a possible techniques for its treatment.
financial crisis on the internationalization
process of an economic sector, it would be
appropriate to obtain information on the
possible effects by interviewing a group of
professionals.
 EAN
Magazine

Although the results obtained through are forced to use qualitative methods. The
quantitative techniques do not necessarily advantage of one over the other is that they are
have to be more reliable than those derived relatively inexpensive and quick in obtaining
information.
from qualitative research, they do allow us to
measure the confidence with which they can In any case, research may be purely
be accepted. In some cases, it may happen qualitative or
that the data is scattered or scarce and we

5. CONCLUSIONS

In this article we have presented the different

techniques and instruments that disciplines
such as Statistics and Econometrics put at
our disposal, to gather and/or create
information on the research topic, presenting
the information in a manageable way,
exploiting the information and thus obtaining
results that allow us to make decisions
considering different scenarios.
To this end, we have named and briefly
described different specific methods for each
research topic. We have conveyed the need
to combine different techniques to achieve
the same objective, inviting the creation of
research designs that encompass different
methodologies, both in the use of different
sources to obtain information and in the
combination of different techniques for data
processing.
 UNIVERSIT
Y
EAN
that
EAN
Magazine

LITERATURE

Anderson, D., Sweeney, D., and Williams, T. (1999): Statistics for Administration and
Economics.
Mexico, International Thomson Publishers.

Azorin, Fy Sánchez-C., J., (1986): Sampling methods and applications. Madrid, Alianza
Editorial.

Berenson, M.L. And Levine, DM, (1996): Basic Statistics in Administration. Mexico, Prentice
Hall Hispanoamericana.

Bernal, C., (2000): Research methodology for administration and economics. Colombia.

Casas Sánchez, JM, (1996): Statistical inference for economics and business administration.
Madrid, Ramón Areces Study Center Publishing House.

Sanchez Houses, J.M., Santos Peñas, J. (1999): Business Statistics, Madrid, Ramón Areces
Study Center Publishing House.

Clairin, R. And Brion, P.(1997): Sampling manual. Salamanca, The Wall.

Hildebrand, D.K. and Lyman O. R., (1997): Statistics applied to administration and
economics.
Mexico, Addison-Wesley Iberoamericana.

Ortega Martinez, E., (1992). Business Research Handbook. Madrid, Pyramid Editions.

Perez, G. (1994): Qualitative research. Challenges and questions. Madrid, 21st Century.

Pulido, A, (1989): Econometric models. Madrid, Pyramid.

Ryan, C. (1995): Researching tourist satisfaction. New York.

Sancho et al. (2001): Notes on Tourism Research Methodology. Madrid, UNWTO.

Santos Peñas, J. And others, (1999): Design and statistical treatment of surveys for market
studies. Madrid, Ramon Areces.

Uriel, E and Muñiz M., (1993): Economic and business statistics. Madrid, AC.
UNIVERSIT
Y

Uriel, E., (1995): Data analysis, time seriesEAN

and multivariate analysis. Madrid, AC.
Magazine