Stats questions

[Kasey Murphy]

As I am reviewing stats 1 more time

Could you explain or give an example for content validity vs construct validity

Can you also explain p level / p value?

Thank you!

[Jessica Lewis]

Hi Kasey!

In terms of construct validity, construct is referring to something that can’t be directly observed but measured by observing other things that are associated with it. In order to measure the construct validity of a test, we are determining if the test as a whole measuring the concept that it is intended to measure. An example would be a test measuring quality of life. Quality of life is more of an abstract construct and not a specific, black and white, construct. Construct validity makes sure that the test is actually measuring quality of life as a whole and not just the child’s pain, social economic status, level of mobility, etc. Most research looks at construct validity because you can do a comparison with a “gold standard” test and if there is an association then we can typically say that the tests are measuring the same construct.

Content validity is measuring whether or not the test is representative of all aspects of the construct. We determine if there are any aspects missing from the test or if there are irrelevant aspects included in the test.

As for p-level vs p-value, I believe these are the same thing. The term p-value is more commonly used and it determines the level of significance. Are you thinking about alpha level vs p-value? The alpha level is the number we choose to evaluate the p-value against. Typically, the alpha level is set at 0.05. You compare the p-value with the alpha level to determine whether or not the data are statistically significant.

Let us know if you have any other questions about this!

Jessica

[Kasey Murphy]

Thank you! I found your explanation to be quite helpful.

Could you also explain or give an example of:
“R squared (or Coefficient of Determination): Measures the percentage of variation in the
values of the dependent variable that can be explained by the variation in the
independent variable”

[Helen Carey]

r squared, or the coefficient of determination, reflects the magnitude of an association between variables and is most appropriate when comparing the magnitude of different correlations. A higher value of Pearson’s r indicates a stronger relationship between the variables. To determine the coefficient of determination, this r value is squared.
Example: if you have independent variables of age and GMFCS level and a dependent variable of TUG test time, you could have an r squared value for age vs TUG test time association (hypothetically .017) and a different r squared value for GMFCS level vs TUG test time association (hypothetically .068). The r squared value for the GMFCS level vs TUG test time association (.068) is a much greater magnitude than the r squared value for age vs TUG test time association (.017), therefore, GMFCS level could be said to explain 4 times as much of the variance than age. To interpret this clinically, we could say that a child’s GMFCS level has a greater magnitude of association with TUG test times compared to age.

(r should be in italics, however, the forum post box won’t let me change the font).

[Author]

Posts