Paired data refers to a type of data where each observation in one dataset is directly related or
connected to a corresponding observation in another dataset. On the other hand, unpaired data
refers to two independent datasets where there is no direct connection or relationship between
the observations in each dataset.
To provide an example of paired data, let's consider a study in the field of medicine where the
blood pressure of individuals is measured before and after a specific treatment. In this case,
the blood pressure measurement before the treatment is paired with the measurement after the
treatment for each individual.
For unpaired data, let's consider a study comparing the effectiveness of two different
medications in treating a specific condition. In this study, one group of individuals receives
medication A, while another group receives medication B. The data collected for each group
is independent and not directly connected to each other.
Now, let's describe these datasets in more detail and discuss the appropriate use of the one-
sample t-procedure or the two-sample t-procedure for each.
In the paired data example of blood pressure measurements, the dataset consists of two
columns: one for the blood pressure measurement before the treatment and another for the
measurement after the treatment. Each row represents an individual's data. The key distinction
in paired data is that the observations are related or connected within each pair. In this case,
the one-sample t-procedure is appropriate because we are comparing the mean difference
between the paired observations to a hypothesized value (e.g., zero) to determine if there is a
significant change in blood pressure after the treatment.
In the unpaired data example of comparing two medications, we have two separate datasets:
one for the group receiving medication A and another for the group receiving medication B.
Each dataset consists of a single column representing the outcome variable (e.g., effectiveness
of the medication) and multiple rows representing the individual observations. In this case, the
two-sample t-procedure is appropriate because we are comparing the means of two
independent groups to determine if there is a significant difference in effectiveness between
the two medications.
In summary, the choice between the one-sample t-procedure and the two-sample t-procedure
depends on the nature of the data. Paired data, where observations are directly related within
each pair, requires the one-sample t-procedure. Unpaired data, where observations are
independent between groups, requires the two-sample t-procedure. It is important to correctly
identify the type of data and choose the appropriate procedure to ensure accurate statistical
analysis.
Reference:
American Psychological Association. (2020). Publication manual of the American
Psychological Association (7th ed.). https://doi.org/10.1037/0000165-000
Yakir, B. (2011). Introduction to Statistical Thinking (With R, Without Calculus). Jerusalem,
IL: The Hebrew University of Jerusalem, Department of Statistics. Retrieved from
https://my.uopeople.edu/pluginfile.php/1408108/mod_page/content/22/Textbook%20Statistica
%20Inference.pdf
