Response to a post
Respond to the post and explain the benefits and consequences of the “relaxed” level of significance.
Statistical significance is achieved when a researcher is able to prove that a relationship exists between
two variables such that a hypothesis can be proven or a null hypothesis disproven. One way to achieve
statistical significance is through the application of a p-test. A p-test enables the proving of a hypothesis
that is predicated upon observing the characteristics of a parameter obtained via random sampling of a
larger population. If the mean, along with a significant number of the values of the parameter associated
with the sample are shown to be higher or lower than the mean and a significant number of the underlying
population’s parameter, then significance can be proven. The term “significant” is predicated upon the p-
value which indicates the portion of the sample parameter data is higher or lower than the baseline data
associated with the null hypothesis (Warner, 2012). A p-value of .01, for example, would indicate that 99%
of a specific parameter’s sampled values are higher or lower than the comparison population. According
to Frankfort-Nachmias (2020), a general rule of thumb used by social psychologists is a p-value of .05 or
5% is sufficient to prove a significance level to justify a given hypothesis. Anything higher than an alpha
value of .05 (ie: p>.05), cannot be deemed to be statistically significant.
If hypothetical researchers were to indicate that “due to their research being exploratory in nature,
a significance level of .10 was used,” I would determine that their findings were suspect and not
statistically significant. Although it is possible that their findings may have some merrit, I would be
unwilling to accept their hypothesis based upon a 90% confidence rate in the data. Based upon such a
comment, it leaves the reader wondering if they conducted their research and then post facto relaxed
their significance test to “fit the data”. If the researchers could not be bothered to readjust their
hypothesis or re-conduct their experiment in order to achieve a significance level of .05 or lower, then I
would not be willing to ascribe credibility or statistical significance to their findings.
References
Frankfort-Nachmias, C., & Leon-Guerrero, A. (2020). Social statistics for a diverse society (9th 3d.). Sage
Publications.
Warner, R. M. (2012). Applied statistics from bivariate through multivariate techniques (2nd ed.).
Thousand Oakds, CA: Sage Publications.