Hello Ali,

Both types of coefficients express the same thing, albeit using different scales. Each represents the amount of change predicted/estimated in the DV, per unit change in the IV (independent of the other IVs). For unstandardized coefficients, you have the values given in the original scale of the DV (per unit change in the original scale of that IV); for standardized coefficients, the changes are all in SD units (as if all variables were z-scores).

Only the standardized coefficients may be compared for magnitude across IVs to answer the question, "Which IV is given more emphasis in the model?" (as suggested by Faizan Mazhar 's answer).

However, which type is better suited to an application depends on: (a) the specific research question(s) being addressed; and (b) the use of the coefficients. People quite frequently report both, by the way.

Both types are almost universally supplied by statistical software (as noted by David Eugene Booth ), so what this boils down to is:

It sounds to me as if you'll either furnish the unstandardized coefficients (as requested) or have to come up with a good explanation for why they are unnecessary!

Good luck with your work.

Look up what your output means in your software manual. Generally you get both estimates. Best David Booth

Hello Ali,

Both types of coefficients express the same thing, albeit using different scales. Each represents the amount of change predicted/estimated in the DV, per unit change in the IV (independent of the other IVs). For unstandardized coefficients, you have the values given in the original scale of the DV (per unit change in the original scale of that IV); for standardized coefficients, the changes are all in SD units (as if all variables were z-scores).

Only the standardized coefficients may be compared for magnitude across IVs to answer the question, "Which IV is given more emphasis in the model?" (as suggested by Faizan Mazhar 's answer).

However, which type is better suited to an application depends on: (a) the specific research question(s) being addressed; and (b) the use of the coefficients. People quite frequently report both, by the way.

Both types are almost universally supplied by statistical software (as noted by David Eugene Booth ), so what this boils down to is:

It sounds to me as if you'll either furnish the unstandardized coefficients (as requested) or have to come up with a good explanation for why they are unnecessary!

Good luck with your work.

I agree with the answers above and would like to point out that interpretation in the original units (unstandardized coefficients) is most often more informative to understand how the predictor(s) are related to the outcome variable. I can hardly come up with an example where standardized coeffeicients are really useful for interpretation with regards to content. As has been pointed out, they are often used to assess the relative impact of the predictors on the outcome, but even this iterpretation can be problematic. A better and (in my opinion) easier way to comapre the impact/importance (if had this in mind) is to show the semi-partial correlations. But have a look at the book by Darington & Hayes (2017) - Regression Analysis and Linear Models, for a discussion about "importance of regressors" (chapter 8).

I agree with Faizan. The unstandardised coefficients are mostly reported to reflect the change in dependent variable per unit change in the predictor. The standardized coefficient are used in comparison of impact of any independent variable on the dependent variable.