This is a response to someone else’s post, need one reference, no min word count.
Partial correlation is the relationship between two variables while controlling for a third variable. It can also be viewed as a three way overlap (a-b-c) and then remove c in the picture in order to determine the correlation between any two of the variables (a &b) especially if c is constant. If c is not constant, then there is certain amount of correlation found any of the variables is tied in with the correlation that each of those two (a & b) has with c.
My example: “a” is the amount of effort spent on making the discussion in an on-line education by the graduate student in nursing, “b” is the graduate student’s score on the discussion in online education, “c” is a measure of time where the graduate student work as a nurse in any institution while taking an online class.
Independent variable is seen as having an effect on the dependent variable (Kellar & Kelvin, 2013). In the sample, “a” (which is the amount of effort spent on making the discussion in an online education by the graduate student in nursing) is the independent variable while “b”(score of the graduate student in an online education) is the dependent variable which is referred to as outcome variable or the response variable. “c” (measure of time when the graduate student work as nurse in any institution while taking an online class) is the confounding variable, which is defined as other than the independent variable that may affect the dependent variable (Mc Donald, 2014).
Partial correlation is the most appropriate test on the sample given because there is a 3rd variable that might mislead or distract the result of comparison of first and second variable. Partial correlation is a technique that allows to describe the relationship between the two variable after statistically controlling the 3rd variable (Kellar & Kelvin, 2013). Thus in my sample study, to be able to obtain the right relationship between “a” and “b” is to make “c” constant. I would control my “c” as selecting my subjects as only graduate students in nursing, who are taking an online education while working as a full-time nurse in any institution. Therefore, if the 3rd variable is controlled statistically, then the outcome will attain a true result.
References:
Kellar, S. & Kelvin, E. (2013). Munro’s statistical methods for health care research (6th ed.). Philadelphia; Wolters Kluwer/ Lippincot.
Mc Donald, M. (2014). Handbook of biological statistics (3rd ed.). Maryland: Sparky House Publishing.
