What Is Statistical Significance A Measure Of?
What Is Statistical Significance A Measure Of?
What Is Statistical Significance A Measure Of?
BEYOND DESCRIPTIVES
Making Inferences Based on Your Sample
Measures we can use to interpret the meaning and importance of our findings:
1. Statistical significance
When the results of a study fall in the extreme 5% (or 1% if you use a more
stringent criterion) of the sampling distribution, suggesting that the obtained findings
are not due to chance alone and do not belong to the sampling distribution defined by
the H .
When our findings fall in the region of rejection and we reject the null
hypothesis, we state that we have found statistical significance (or a statistically
significant difference) between the sample and the sampling distribution that we are
comparing. If we find significance, we can conclude that our sample must have come
from a different distribution than the one defined by our null hypothesis.
Statistics is all about taking a piece of the population and making a guess about
what that population’s behavior might be like. If you were working with parameters
(parameter vs. statistic explanation), there would be no need for guesswork; You’d have
all the data. In real life getting all the data can be costly, time-consuming, or impossible.
For example, Gallup Polls uses statistics to estimate who will win the next
election. Drug manufacturers use statistics to estimate how many people might have
side effect from their drugs. And businesses use statistics to forecast sales figures for the
future.
What is Statistical Significance a Measure of?
Statistical significance is a measure of whether your research findings are
meaningful. More specifically, it’s whether your statistics closely matches what value
you would expect to find in an entire population. As a simple example, let’s say you
worked for a polling company and asked 120 people how they were going to vote in the
next election. You would want your report to reflect everyone in the country, right? In
other words, you want your report to have significant findings. How is “significance”
measured? With a few calculations.
To test for statistical significance, perform these steps:
1. Decide on an alpha level. An alpha level is the error rate you are willing to work with
(usually 5% or less).
2. Conduct your research. For example, conduct a poll or collect data from an
experiment.
3. Calculate your statistic. A statistic is just a piece of information about your sample,
like a mean, mode or median.
4. Compare the statistic you calculated in Step 3 with a statistic from a statistical table.
2. Effect Size
The effect size tells you the magnitude or strength of the effect of a variable.
One of the easiest types of effect size to understand is the percentage of variability in
one variable (the dependent variable), which is accounted for by the independent
variable (in the case of experiments) or which is accounted for by the relationship with
another variable (in the case of correlations).
This effect size, expressed as a percentage, can range from .00 to
1.00. For example, if the effect size equals .10, then 10% of the variability in the
dependent variable scores would be accounted for by the independent variable. That
would mean that 90% of the variability in the dependent variable is not associated with
the independent variable.
The effect size in psychological research is more likely to be smaller. In
interpreting the effect size, between 1% and 4% is considered a small but reasonable
effect, 9–25% is considered a moderate effect, and 25–64% is considered a large effect.
These ranges were never intended to be strict cutoffs but rather to serve as guidelines
to enable us to evaluate the strength of relationships between variables.
Interpretation of Effect Size:
Effect size tells you how meaningful the relationship between variables or the
difference between groups is. It indicates the practical significance of a research
outcome. A large effect size means that a research finding has practical significance,
while a small effect size indicates limited practical applications.
3. Practical Significance
Practical significance refers to the usefulness of our results or findings from our
study. In other words: How do the results affect or apply to daily life? Even if we find
statistical significance, the difference we find may not be noticeable or noticeably affect
people’s lives. When we conduct studies, we should consider the practical use or
implications of our findings. Findings that make a noticeable difference in the world
outside the laboratory in addition to being statistically significant are memorable and
define areas that are likely to generate additional studies, so practical significance is
another aspect of research that we should consider.
Statistical significance, effect sizes, and practical significance vary independently
so that you can obtain any combination of the three factors in a study. In one study you
may have statistical significance but find a very small effect size and no practical
significance.
A few specific examples may help you to understand the different combinations
of outcomes from a study. Suppose you find that psychology majors
(M = 42.50, SD = 1.24) send significantly more text messages than college students in
general (M = 39.10, SD =2.82; p = .03) and that your effect size is 2%, so that the major
accounted for only 2% of the variability in texting frequency. The difference between
texting 42 times a day vs. 39 times a day is minimal in terms of impact on time or
attention to one’s phone. In this case, you found statistical significance, a small effect
size, and very little practical significance in this study.
Consider if instead you found that there was no difference (p = .15) in the mean
texting frequency for psychology majors (M = 60.5, SD = 4.67) and college students (M =
39.1, SD = 5.82) but the effect size was moderate and the major accounted for 20% of
the variability in testing. In this case, the psychology majors sent or received
approximately 20 more text messages than their college peers, a difference that might
be noticeable and thus have significance in terms of different amounts of time spent on
texting, which could influence interactions with others or attention in class. Hopefully,
these examples demonstrate that attending to all three factors (statistical significance,
effect size, practical significance) will help you to better understand the meaning of your
study’s results.