Inferential Statistics: Significance, Pitfalls, and P-Value Use

This assignment delves into inferential statistics, focusing on the significance of p < .05 in determining statistical significance and the associated pitfalls of hypothesis testing. It explores the reasons behind the widespread use of p < .05, tracing its origins to Fisher's convenient threshold for judging deviations. The discussion highlights the limitations of statistical significance, including its potential disconnect from practical significance, the risk of inappropriate hypothesis testing methods, and the impact of small sample sizes. Furthermore, it addresses publication bias and misinterpretations arising from uncorrected multiple testing. The assignment also examines the skepticism within medical fields regarding the strict application of p < .05 due to the high stakes and the need to minimize errors in statistical inference, referencing Cohen's critique of null hypothesis testing. Ultimately, the work underscores the importance of understanding the nuances and limitations of statistical significance and hypothesis testing in research.

Running head: INFERENTIAL STATISTICS 1
Inferential Statistics
Students Name
Institution Affiliation

INFERENTIAL STATISTICS 2
Inferential Statistics
Why do you think p < .05 is used to determine statistical significance?
Quantitative studies are research that involve the use of data which are analyzed so that
conclusions can be drawn about the phenomenon under study. Statistical inference theory is
based on the idea that the results from a sample can be generalized to the population. Many
quantitative studies use p<.05 to determine statistical significance. Dahiru (2008) argues that, 5%
level of significance is associated with greater feasibility. The researchers argued that with
p<0.05, it is feasible to set up studies with fair possibility of selecting effects that are significant.
The use of p<0.05 as a quantification of evidence against the null hypothesis was first proposed
by Fisher long time ago. It has continuously been adopted by researchers, consultants, and
teachers without significant opposition. Fisher argued that p<.05 was convenient point that could
be used as a limit when judging if a deviation ought to be regarded as significant or not. The
p<.05 rule of the thumb takes all deviations that exceed two standard deviations as significant.
What are the pitfalls of statistical significance or hypothesis testing?-166
Statistical significance or hypothesis testing is a highly applied concept in inferential
statistics. Despite of the fact that it has been exclusively applied in statistics, there are significant
pitfalls associated with it. Confidence intervals and hypothesis testing are main methodologies
used in statistical inference. Because of their weaknesses, it is advisable for researchers to use
both methods simultaneously for more precision (Dancey & Reidy, 2017). Statistical
significance doesn’t necessarily imply practical significance. Hypothesis testing especially when
inappropriate methods of hypothesis testing are applied. A hypothesis test might indicate
whether the parameter value hypothesized is plausible. But fails to identify the exact parameter

INFERENTIAL STATISTICS 3
values that are plausible. The confidence interval as used in statistical inference only provides
sets of plausible parameters. Also, a small sample size might result in a test design that needs the
system to execute at points much beyond specified levels so that the power of the test
successfully approaches reasonable values (Cohen, 1994). Across studies, hypothesis testing
might result to exaggerated focus which might result to a disregard of prior knowledge.
Publication bias is also a common pitfall since most researchers might be tempted to distort
results in favor of positive results. Within studies, statistical significance can lead to uncorrected
multiple testing and semantically induced misinterpretations.
In medical fields, p < .05 is not acceptable to reject the null hypothesis. Why do you think
this is?
Cohen (1994) discussed the concept of statistical power in psychology inference testing.
The researcher was critical about the use of null hypothesis testing around p<0.5 statistical
significance. Many medical researchers are critical of the application of p-values and hypothesis
testing in the medical field. The use of p < .05 as a determinant of whether a null hypothesis is
accepted or rejected might threaten scientific research and their findings by yielding misleading
conclusions. Despite of some medical researchers accepting that the concept of p-values and
hypothesis testing are important tools of critical research, they are still unconvinced and skeptical
about the absolute rule of 0.05. Health is an important aspect of life and since life is dignified
and precious, medical researchers are interested in using methods that reduce the probability of
errors in statistical inference.

INFERENTIAL STATISTICS 4
References
Cohen, J. (1994). The earth is round (p < 0.05). American Psychologist, 49, 997-1003.
Dahiru, T. (2008). P-value, a true test of statistical significance? A cautionary note. Annals of
Ibadan postgraduate medicine, 6(1), 21-26.
Dancey, C., & Reidy, J. (2017). Statistics without maths for psychology (7th ed.). Harlow, UK:
Pearson.