Page 448 - Contributed Paper Session (CPS) - Volume 2
P. 448
CPS1922 Imran K. et al.
the privacy of the names of authors and the journals from which the Type-IV
error has been found have been deliberately hidden.
3. Result
“Error of the Fourth kind or Type IV error”
Before the availability of free and friendly statistical software, the
researchers would collaborate with the statistical consultant for collection,
analysis and interpretation of data as it was not possible for them to perform
analysis of their own. In the last 10years, use of statistical software has grown
tremendously among the researchers, academicians of government and non-
government organizations. With the mushrooming of many user-friendly and
free statistical software, the statistical analysis is done badly by the researcher
who doesn’t have knowledge of statistics and hence given birth, by a click, to
a new type of error. Statistical software is nowadays used recklessly by the
researchers and hence commit an error, by clicking the mouse, we call “Error
of the fourth kind or Type IV error”, thus give the wrong answer to the right
question. Such type of errors has been observed by the authors while dealing
with the research problems of other disciplines. Almost every researcher is a
statistician in today’s world. Using statistics without collaborating with the
statistical expert is just like taking a medicine without consulting a doctor. With
the availability of free online statistical software and videos, most of the
researchers analyze and interpret their research data/findings of their own
without collaborating with the statistical expert.
Type IV error is mostly committed due to improper use of user-friendly
statistical software like entering data in a wrong way, filling data cells wrongly
(like filling blank cells by 0), no/poor knowledge about outliers and extreme
observations or we can say improper handling of problem data, inappropriate
use of statistical tools, incorrect interpretation of statistical results. Statistical
software like data are innocents and they dance the way researcher makes
them dance rather the way they should. Statistical software will definitely
analyze the data and give the results, it’s the statistical expert/collaborator
who knows the correct way of extracting useful information from precious data
of the research. The reasons for not collaborating with statistical experts are
three-fold(i) researcher assumes he has enough information regarding the
statistics, (ii) Unavailability of Stat Labs in the university or the research
institute, (iii) researcher dislikes to incorporate another author to his
publication. Research is a team work and if conducted in collaboration will lead
to a correct and high-quality product. It’s very difficult for the researcher to
discover where he has gone wrong. Though Type IV error is difficult to
calculate but from our experience its probability is more than 0.80 (found in
publications where the statistical expert has not been consulted). The
statistical errors that can spoil the research work while using Statistical
437 | I S I W S C 2 0 1 9