You are on page 1of 4

Statistics in Educational Research

Ganesh Chandra Naik

GVM’s Dr. Dada Vaidya College of Education

Ponda, Goa

Abstract:

It is difficult to conceive and do educational research without the knowledge and use of statistics.
Knowledge of statistics also helps in reading and appreciating research literature. But
unfortunately this subject is given not so importance by many teachers and students. This has
resulted in poor quality of research. This paper discusses some of the reasons of this problem.
Also it presents some common problems found in the use of statistical tools and its related issues
in research reports.

---------------------------------------------------------------------------------------------------------------------

Introduction: Educational research particularly empirical / quantitative is statistics-dependent


research. Without a thorough knowledge of statistics, it becomes very difficult to conceive and
execute a research work. To a great extent quantitative research methods in education is 90%
statistical methods. Statistical methods are indispensable for collecting, organizing, analyzing
and interpreting data collected in numerical form[ CITATION Kou84 \l 1033 ]. From writing the
statement of the problem to writing the conclusion of research a research worker is guided by the
statistical methods. If that is so, the question arises on the importance given to statistics in
research methodology courses at Master’s level courses and the way it is treated in the
curriculum transaction. This question would be explored under theory and practice debate.
Secondly, some common problems and mistake the researchers face and commit in doing
educational research would be discussed.

1. Theory – Practice debate in teaching statistics in research methodology: This problem


looks very trivial but I have encountered it both as student and teacher at various levels.
Teachers’ attitude towards this question determines the way statistics is treated in the course.
Some teachers and departments hold the view that because of availability of statistical software,
teaching of calculation and rigorous practice of solving statistical problems is not required. So, in
the teaching of statistics students are not given practice on statistical problems. Only theory part
of statistics is discussed. Therefore, the students get minimal and that too only theoretical
knowledge of some statistical tools (Naik, 2015). This has been described as ‘week syllabus’ in
terms of statistical techniques taught at Master’s level (Sansanwal 2015). This problem is also
due to the lack of qualified teachers (teachers competent to teach statistical methods) at Master’s
level programmes. Even the students at Master’s level are scared of statistics and prefer less
importance be given to practice part of statistics.

In B. Ed also we have educational statistics under educational evaluation. There too we were not
asking calculation type questions in the university examination. Because there were no questions
on calculation, we used to pay little attention to calculation in our teaching. Recently we started
asking simple questions involving calculation in this paper. But to my surprise not more 2%
students attempt this kind of questions. And out of these, many would get the answer wrong.
This exemplifies the status of statistics with students. It looks as if there are no takers of statistics
either at B. Ed and M.Ed level. Because of poor knowledge of research guides and research
students in statistics, several mistakes could be found in the research reports and research articles
in journals. The following sections describe this problem.

2. Some common problems and mistakes involving statistics found in research reports.

a. Research students go on collecting all possible data believing that these would be useful and in
that process collect huge amount of data. This happens when the data is qualitative and personal
and also do not belong to interval scale. These data are so varied and diffused that the researcher
faces a tough problem in deciding appropriate statistical tools for data analysis. Data becomes
unmanageable and finally redundant.

b. Sometimes the required data may not have been collected due to poor planning. Or the data
that has been collected is inadequate to arrive at some definite conclusion and not in the desired
format. It is not possible to go back to field for data collection. Because of no planning or poor
planning, the collected data is not relevant to the purpose of research.

c. After collection of data students face the problem of selecting appropriate statistical tools for
data analysis. This is very common problem involving research students. Even if they have
access to Statistical software like Excel and SPSS, they do not know which statistical techniques
is appropriate keeping in mind their requirement.

d. Indiscriminate use of parametric statistics is found where the non-parametric statistics could
have served the purpose. Parametric statistics are very powerful statistical procedures with strict
assumptions to be met. The researchers ritually go on using parametric statistics believing that
their result would have more acceptability.

e. Sampling techniques determines statistical tools. And very often it is statistical tools to be used
( though it should not happen) that determine sampling. Whatever it may be the case, knowledge
of statistical techniques and their appropriate use is very much necessary here in selecting
appropriate sampling for research. Without considering all these, the researcher goes for some
sampling mostly incidental or any kind of non probability sampling. But at the time of statistical
analysis they go for powerful statistics. Even the size of sample matters a lot in deciding
statistical tools even within parametric group of statistics. In Fourth Survey of research in
education, sampling size and rational for particular sampling technique in different Ph. D works
has been thoroughly criticized by Buch (1991).

f. In using complicated experimental design and even in co-relational research, if appropriate


statistical tool is not used, misleading finding would come. There is statistical method of
matching experimental and control groups. If not used, the finding would be different from what
it would have been otherwise and it would affect internal validity of research. In co-relational
research point bi-serial and partial correlation are used selectively find out specific relation
between two or more variables. Blind use of simple correlation techniques would not be
appropriate in many situations.

g. Problem in interpretation: Data analysis is there in the textbooks. It is also at times taught. But
when it comes to interpretation, it is neglected. Statistical analysis is taught but not its
interpretation[ CITATION San12 \l 1033 ]. Researchers commit mistakes in hypothesis testing and
in interpreting the result. They find an unusual problem of reconciling when to their surprise
their null hypothesis is accepted against their hypothesis indicating differences in two groups.
The investigator tries all excuses to say why the result is otherwise and not expected. It shows
the desperation on the part of the researcher to accept the research findings. Sometimes the
researcher even goes a step ahead to deny the result and support the observed data rather than the
statistical finding (Naik 2015).

h. The researcher does not write the justification of statistical techniques: This is a part of report
writing authentically. The research report must be authentic, reporting everything that make the
readers convinced that the research and the work behind is authentic. The researcher rarely
reports why s/he decided a particular statistical tool for data analysis. Most action of the
researcher in research design should be justified. This enables the readers to appreciate the
intention and thinking of the researcher. For example when the investigator uses mean in a
situation, s/he should write why mean not other measure of central tendencies there to be used.
This kind of reporting enhances the quality of research reporting and the readers benefit from
reading this kind of research reports.

Conclusion: The importance of statistical methods in research particularly quantitative research


need to be realized by all; research guides, research students and faculty members of Post
Graduate departments. Knowledge of statistical methods would improve the quality of
educational research. Several mistakes and lacuna found in research reports due ignorance of
statistical methods could be avoided if all stake holder of educational research are comfortable
with statistics.

Bibliography
Buch, M. (1991). New Direction for Educational Research In India. In M. Buch (Ed.), Fourth Survey of
Research in Education (pp. 1-45).

Koul, L. (1984). Methodology of Educational Research. New Delhi: Vani Educational Books.

Naik, G. (2015). Methodological Issues in Educational Researtch. University News , 53 (17), 20-23.

Sansanwal, D. N. (2015). Methodological Issues in Experimental Research. University News , 53 (10), 12-
16.

You might also like