POSITIVISM IN CROSS SECTIONAL SURVEYS: CONSEQUENCES AND CURES

What is Cross sectional Survey?

Survey is one of the research strategies which the opinions of the

sample or population is sought by the researcher, usually with a
more objective research instrument, say a structured
questionnaire. It is usually associated with a deductive approach
and is conducted usually in business and management research
to prove or disprove certain assumed propositions and
hypothesis. It allows for the collection of large amounts of data
from a large population economically. It is most frequently
conducted to answer research questions relating to ‘who, what,
how much and how many’ involved in a problem of study.

In terms of time horizon, research design can be longitudinal or

cross-sectional. A cross-sectional study examines a particular
phenomenon at a specific period of time (Saunders et al., 2007).
According to Malhotra and Birks (2007) one sample of a
population can be taken and studied at a particular time as in a
single cross-sectional study or two or more samples of a target
population could be studied once as in multiple cross-sectional
study.
Conversely, in terms of time horizon, a study may be
longitudinal where a particular phenomenon is studied at different
periods of time. Longitudinal study can take the form of a single
longitudinal study where only one sample is studied at different
time periods or multi-longitudinal where two or more samples are
studied at different periods of time (Malhotra & Birks, 2007). This
study is typically a cross-sectional study in that data will be
collected from a cross section of customers of organized retail
outlets and private banks.

A cross-sectional survey is defined as an observational research

type or study that analyses data of variables collected at one
given point of time across a sample population or a pre-defined
subset. This method is often used to make inferences about possible
relationships or to collect preliminary data to support further research and experimentation .
This type of survey is also known as transverse study or
prevalence study.
The data collected in a cross-sectional study is from people who
are similar in all variables except the one variable which is under
study. This variable remains constant throughout the cross-
sectional study. This is unlike a longitudinal study, where
variables in the study can change over the course of research.

Cross-sectional studies are conducted across multiple industries

like:

 Retail – In retail, a cross-sectional study can be conducted

on males and females between the ages 24-35 to check for
spending or purchase trend differences between the sexes.
 Business – In business, a cross-sectional study can be
conducted to understand how people of different socio-
economic status from one geographical segment respond to one
change in an offering.
 Healthcare – A cross-sectional study in healthcare is used to
understand how prone kids between the age of 2-12 across
different boroughs in the US are prone to a low calcium
deficiency.
 Education – A cross-sectional study in education is
particularly helpful to understand how either males or females
from a similar age bracket but different ethnicities react to their
grasp of a certain object.
Cross-sectional research helps garner a lot of quick actionable
data that helps in decision making and offering products or
services.

Types of cross-sectional studies

 Descriptive
A cross-sectional study may be entirely descriptive and it used to
assess the frequency and distribution of the study topic in a
certain demographic. For example, how a random sample of
Universities across a state is assessed to check for obesity
amongst women and women.

 Analytical
This type of cross-sectional study is used to investigate the
association between two related or unrelated parameters. This
methodology isn’t entirely full proof though because the presence
of risk factors and outcomes are simultaneous, and their studies
are simultaneous too. For example, to validate if coal workers in
the mine could develop bronchitis only looks at the factors in the
mine. What it doesn’t account for is that bronchitis could be
inherent or may have existed from before.
 In a real-life cross-sectional study, attributes from both types are
used.

Defining variables of Cross-Sectional Studies

Some of the key variables of a cross-sectional study are:

 The cross-sectional study is conducted with the same set of

variables over a certain period of time. The study is conducted
in a single instance, unlike longitudinal studies, where variables
can change over the period of extensive research.
 Cross-sectional studies give the flexibility to the researcher
to look at multiple variables together as a constant, with only
one variable being the focus of the cross-sectional study.
The easiest way to encapsulate a cross-sectional study is that it is
a snapshot of a group of people at that point in time. They are
used to determine what is happening in real time, at the moment.
Hence, this research type is used to draw the pulse of the
population data at any given point in time.

You can also use it to map prevailing variables that exist at a

particular point. For example, cross-sectional data on past
drinking habits and a current diagnosis of liver failure.

Cross-Sectional Study Examples

Below are examples to better illustrate this method of research

1. Phone companies rely on advanced and innovative features

to drive sales. Research amongst the target demographic
market by a phone manufacturer validates the expected
adoption rate and potential sales of the phone. In a cross-
sectional study, males and females across geographies and
ages are enrolled for this research. If the study results show that
Asian women would not buy the phone because it is bulky, the
mobile phone company can tweak the design before its launch
or develop and market a smaller phone for women.
2. Another example of a cross-sectional study would be a
medical study looking at the prevalence of cancer amongst a
certain population. The researcher can evaluate people from
different ages, ethnicities, geographical location and social
backgrounds. If a significant amount of men from a particular
age are found to be more prone to have the disease, the
 researcher can conduct further studies to understand the
reason behind it – like a longitudinal study.  

Analysis of cross-sectional studies

In a cross-sectional study, to calculate prevalence, multiple

parameters are measured simultaneously – questions,
observations, and answers.

Prevalence = No of cases at a given time / No of people at the

same given time

For continuous variables, they fall along a continuum within a

given range. To calculate prevalence, the values have to be below
or above a predetermined level or else median levels may be
calculated.

Cross-Sectional Studies Advantages and Disadvantages

Here are some of the key advantages and disadvantages of

conducting online research using cross-sectional study.

Advantages of cross-sectional studies

 Relatively quick to conduct

 All variables are collected at one go
 Multiple outcomes can be researched at once
 Prevalence for all factors can be measured
 Good for descriptive analysis
 Can be used as a springboard for further research

Disadvantages of cross-sectional studies

 Cannot be used to get timeline based research

 Tough to find people that fall under the exact same variables
 Associations are tough to interpret
 When strong feelings are involved, there could be the case
of a bias
 Does not help to determine cause

Cross-Sectional vs. Longitudinal Studies

Even though they are both quantitative research methods, there
are a few differences between cross-sectional studies vs
longitudinal studies.

In cross-sectional studies, the variables are collected at a certain

given point of time but longitudinal studies span across multiple
sessions and the variables could change.

Cross-sectional studies are preferred to find common points

between variables but longitudinal studies, due to their nature are
then used to dissect the research from the cross-sectional study
even further.

The other factor is that due to the longevity of the research, there
could be attrition in the longitudinal study which skews the results
finding and the corresponding research whereas, in cross-
sectional research, there’s no chance of that happening because
the study is done with the same variables and sometimes at the
same time of collection of variables.
Positivism in Cross sectional Survey
Positivism has to do with the situation where knowledge or the
world is thought to exist independent of people’s perceptions of it
and that science uses objective techniques to discover what exist
in the world” (Sullivan, 2001 p.47). Positivism uses logical,
quantitative, more objective scientific methods to test
hypothetically deductive generalizations. On the other hand,
phenomenological or interpretive philosophy holds that “reality of
the world is thought to arise out of the creation and exchange of
social meaning during the process of social interactions” (Sullivan,
2001 p. 48). Phenomenological perspective uses qualitative, more
subjective, naturalistic approaches in inductively and holistically
to understand human experiences in context-specific settings
(Amaratunga, Baldry, Sarshar & Newton, 2002).

As a philosophy, positivism adheres to the view that only “factual”

knowledge gained through observation (the senses), including
measurement, is trustworthy. In positivism studies the role of the
researcher is limited to data collection and interpretation in an
objective way. The quantitative paradigm is based on positivism.

Human beings are seen objectively, and as a result, social

scientists look to different avenues to study human society (De
Vos et al., 2011b:5). Babbie (2011:35) states that the roots of
positivism can be traced to Auguste Comte, who saw human
beings as a phenomenon to be studied scientifically. Thus,
positivism may be seen as an approach to social research that
seeks to apply the natural science model of research as the point
of departure for investigations of social phenomena and
explanations of the social world (Denscombe, 2008:14;
2010b:120). The natural sciences are suitable for the study of the
social world; and hence, many researchers assume that the
positivist approach is scientific. Naturally, one would ask the
question: What is positivism supposed to comprise?

In response, Glicken (2003:20); Denscombe (2010a:324) and

Lincoln et al. (2011:107-108, 122) answer as follows: Positivism
firstly entails a belief based on the assumption that patterns
(trends), generalisations, methods, procedures, cause-and-effect
issues are also applicable to the social sciences. This view of
positivism maintains that the objects of the social sciences,
namely people, are suitable for the implementation of scientific
methods.
Welman et al. (2009:6) also link positivism directly to the
scientific model. This model or approach strives to formulate laws
applicable to populations. These said laws explain the causes of
observable and measurable behaviour. The positivist researcher
prefers working with an observable social reality; and such
research would produce generalisations similar to those produced
by the natural scientists. Positivists also believe that an objective
reality exists outside personal experiences with its own cause-
and-effect relationships (Remenyi et al., 1998:32; Saunders et al.,
2000:85; Riege, 2003:77; Neuman, 2006:82; Babbie & Mouton,
2008:23; Saunders et al., 2009:113; Muijs, 2011:4).

The positivist researcher maintains that it is possible to adopt a

distant, detached, neutral and noninteractive position (Morris,
2006:3). A position such as this would enable the researcher to
assume the role of an objective analyst, making detached
interpretations about those data that have been collected in an
apparently value-free manner. For the same reason, positivists
prefer an analytical interpretation of quantifiable data (Druckman,
2005:5). The abstract ideas of the social relationship should,
consequently, be linked to the precise measurements of the social
world.

Secondly, positivism entails a belief that valid knowledge can only

be produced on the basis of direct observation by the senses; and
this would include the ability to measure and record what would
be seen as knowledge. Observation in this sense means
accepting only empirical evidence as valid evidence. Valid
evidence is thus produced through the senses of sight, smell,
touch, taste and hearing. It would clearly mean that there is no
place for phenomena which cannot be observed either directly,
through experience and observation, or indirectly, with the aid of
instruments.

Moreover, it should be quite obvious that things that cannot be

seen (observed), for instance people‟s thoughts and attitudes,
cannot be accepted as valid evidence and knowledge. Thirdly,
many accounts of positivism suggest that scientific knowledge is
arrived at through the accumulation of verified facts. These facts
feed into the theoretical edifice pertaining to a particular domain
of knowledge. Thus, theory expresses and reflects the empirical
research. Such findings are often referred to as laws pertaining to
a particular field, namely empirically established regularities
(Bryman, 2005:15).
Fourthly, as De Vos et al. (2011b:6) remind us, scientific theories
are seen by positivists as providing hypotheses, which are then
submitted to empirical testing. This implies that science is
deductive, as it seeks to extract specific propositions from general
accounts of reality. Logically, this would entail the construction of
a specific theory to explain the laws in a particular field. A
hypothesis is thereby derived to enable the researcher to submit
the hypothesis to rigorous empirical examination before rejecting,
revising or accepting the hypothesis.

Finally, positivism is also taken to entail a particular stance in

relation to values. Consequently, the researcher would need to
be purged of values, since these could impair the objectivity and
so undermine the validity of knowledge. Positivism‟s position on
values is to draw a sharp distinction between issues, statements
and norms (Flick, 2007:12). While positivists recognize that they
can investigate the implications of a particular normative position,
they are unable to verify or disprove the position itself.

In conclusion, regarding scientific theories, it may then be stated

that the positivist is concerned first and foremost with the
creation of laws applicable to all people at all times (Welman et
al., 2009:192). Collis and Hussey (2009:58, 61-62) proceed from
the previous argument when they clearly state that the purpose
of positivism is to seek generalisations (theories). The said
generalisations are, however, based on and grounded in the
natural science laws, which are not necessarily applicable to
social structures. In sum, positivism “equates legitimacy with
science and scientific methods” (Scott & Usher, 2011:13); and as
such, it involves a number of assumptions, (cf. paragraph 5.2.1.1,
p. 301). In this thesis, generalisations are sought in Chapter Four,
where the core contents of different sport management and
educator training programmes are analysed to determine their
common content.

The positivist tradition however, has not met with approval and
support by all scientists, since it has produced some serious
problems as well as some questionable assumptions. Henning et
al. (2004:17), Babbie (2010:41), Rubin and Babbie (2010:15) as
well as Denzin and Lincoln (2011a:8) point out that early positivist
social scientists assumed that social reality can be explained in
rational terms, because people always act rationally. Babbie
(2010:41) in particular states that people do not always act
rationally. Nonetheless, even non-rational behaviour could be
rationally understood and predicted. Babbie (2010:42) further
alleges that everybody acts, thinks and interprets subjectively to
a certain extent. This subjectivity is unique to any individual; and
the endeavour for objectivity could best be obtained through the
discovery of intersubjective interests between individuals.
Inevitably, the positivist view would not agree with this approach.

Following Babbie (2010:44; 2011:44), it was rather difficult to

choose the best suitable paradigm among those presented in this
chapter. Nonetheless, because this study does not focus on the
natural sciences, it cannot be aligned entirely with the positivist
paradigm (approach), as will become evident in the ensuing
paragraphs of this chapter. It is nevertheless important to pay
careful attention to Babbie (2011:44) when he implicitly warns
against the total negation of the positivist paradigm, because
each paradigm compensates for the other by suggesting
complementary perspectives. The different paradigms should be
seen as different arrows that could be used as the situation
demands or requires.

There are two more important principles of positivism: to isolate,

analyse and understand the causes of human behaviour. For
Livesey (2011b:3) the basic thought is that behaviour is caused
(initiated) by something, which if understood, could be applied to
explain and predict human behaviour. The second principle is
concerned with objectivity. Gratton and Jones (2010:25) firmly
believe that for the positivist, there would be an emphasis on
methodology to facilitate replication and quantifiable observations
for statistical analysis. Here, the researcher is independent of
and neither affects nor is affected by the subject of the research.
Welman et al. (2009:6) clearly state that the positivist approach
underlies the natural scientific method in human behavioural
research and holds that research should be limited to what can be
observed and measured objectively. In relation to the current
study, this implies that the generation of data should be
independent of human opinions and judgment.

The researcher distributed a questionnaire to school sport

managers at selected schools in a diversity of South African
schools, and he assumed the role as observer of phenomena in a
natural setting.
Qualitative methods of data generation would not fit easily within
the positivist approach to research (cf. par. 5.8, p. 328). For this
very reason, Curtner-Smith (2002:38) states that positivist sport
pedagogy researchers specifically tend to use quantitative
methods to gather measurable numerical data. In this way, their
work is aimed at the provision of numerical illustrations of
teaching or coaching (descriptive studies), discovering
relationships between components of teaching or coaching
(correlation studies), or attempting to change some aspect of
teaching or coaching (experimental studies).

Wisker (2008:65); Welman et al. (2009:9), as well Gratton and

Jones (2010:25), also propose the use of quantitative methods,
where control groups could be used, for the positivist approach.
Muijs (2011:3), as well as Thomas et al. (2011:19,21) also
mention the use of experimental and survey methods to collect
data; but they indicate that the use of these methods would not
necessarily produce understanding, explanations and
interpretations.

Concurring with Gratton and Jones (2010:25), one could say that
the positivist approach undeniably has strengths, notably in terms
of precision, control and objectivity. Such sport-related theories
naturally would be the result of statistical analysis, which
removes the need for more individualistic or intuitive
interpretation. Positivist research is also generally more
straightforward in terms of planning, simply because the data are
collected in one go, and the analysis of all the data takes place at
the same time. Finally, it emerged during the literature review
that in the early years of sport-related research, such research
was dominated by the positivist approach. For reasons, which will
be outlined below, alternative approaches are now becoming
more widespread. One of these approaches is known as post-
positivism.

Post-positivism
Dissatisfaction with positivism became increasingly widespread,
thereby increasing the appeal of post-positivism (Teddlie &
Johnson, 2009:68). Because of the increasing appeal of post-
positivism, post-positivistic works gained credibility throughout
the social science community. Teddlie and Johnson (2009:68) cite
the works of post-positivists: Popper (1934; 1959), Hanson (1958),
Toulmin (1960), Campbell and Stanley (1963), Hempel (1965),
Kuhn (1962; 1970; 1996), Phillips (1987; 1990), as well as Phillips
and Burbules (2000).

Post-positivism will not be considered a distinct philosophical

tradition in its own right. Creswell (2009:6) sees post-positivism
as an extension of positivism, since it represents the thinking
after positivism, challenging the traditional notion of the absolute
and objective truth of knowledge in the social sciences. Gratton
and Jones (2010:26-27) hold the view of post-positivism that in
reality, it is not possible to gain understanding merely through
measurement. Post-positivist approaches show a much greater
openness to different methodological approaches, and often
include qualitative, as well as quantitative methods. This allows
for the development of alternative research strategies to find
information in unlikely and creative ways (Glicken, 2003:28).
Additionally, researchers in this paradigm normally believe in
multiple perspectives from participants rather than a single reality
(Creswell, 2007:20; 2009:7).

Positivism contends that there is an objective reality out there to

be studied, captured and understood, whereas post-positivists
argue that reality can never be fully apprehended, only
approximated (De Vos et al., 2011b:7). According to Denzin and
Lincoln (2011a:8), post- positivism relies on multiple methods for
capturing as much of reality as possible. At the same time,
emphasis is placed on the discovery and verification of theories.
Traditional evaluation criteria, such as internal validity, are
stressed, as is the use of qualitative procedures that lend
themselves to structured (sometimes statistical) analysis.
Computer-assisted methods of analysis that permit frequency
counts, tabulations and low-level statistical analyses may also be
employed.

The post-positivist researcher focuses on the understanding of the

study as it evolves during the investigation. The study begins
with an area of study. A question and a hypothesis are
conjectured before starting the study (Morris, 2006:77). Post-
positivists believe that positivist research is often difficult and
impractical for many forms of social research (Glicken, 2003:27).
Emanating from any research, there are tendencies towards a
specific notion which can – by repetition – bring valuable data to
light.
Post-positivists accept that the natural sciences do not provide
the only model for social research. However, they do believe in
an objective reality. Rather than focusing on certainty and
absolute truth, the post-positivist will focus on confidence: How
much can the researcher rely on his/her findings? How well can
one predict certain outcomes?

The proponents of post-positivistic research argue that research,

even scientific research, is frequently a product of historically
located practices. Post-positivism reflects a distrust of absolutes
and foundational truths; following the correct method can no
longer guarantee true results. Instead of only one truth, there are
many. Truth is fundamentally dependent on language; and it is a
socially constructed phenomenon. This distorts the reality on
which positivism is built.

Post-positivism provides the researcher with more subjective

measures for gathering information. The degree of honesty of the
researcher could be a problem in this kind of research. Could the
subjectivity in a post-positivistic study influence the data
negatively? Glicken (2003:29) is convinced that post-positivist
research offers the social scientist the ability to do research on a
small scale by using very creative methodologies. Thus, a mixed
method of research was used (cf. par. 5.7, p. 321). The
researcher employed in-depth interviews and questionnaires to
collect the data for the investigation.

The limitations of positivism have led to the development of an

alternative perspective – a collection of related perspectives –
those of interpretivism.