Neeta Jain , I am not David Morse ;) but maybe I can help. In Stata, the code to calculate statistical power for a multiple regression using a F-test is just:

power rsquared 0.1 0.2, n(10(10)200) ntested(3) ncontrol(3) graph alpha(0.01,0.05,0.1)

Please see the figure attached. Here I am assuming:

• Different sample sizes from n =10 to n = 200
• Conventional significance values of 1%, 5% and 10%: alpha(0.01,0.05,0.1)
• 3 control covariates: 1)ethnicity, 2)education level, 3) type of work
• 3 IV of interest: loneliness, creativity and resilience

Here is the trick: "rsquared 0.1 0.2" means that you are assuming that a model with only control will explain only 10% (0.1) of variance of DV (working from home), but a full model with controls and the 3 IVs of interest explains 20% (0.2) of the DV. You can use previous studies to approximate values of R2, Stata will calculate the associated effect size (delta). For example, for a sample size of n = 142 and a significance level of 5%, the Stata command is:

power rsquared 0.1 0.2, n(142) ntested(3) ncontrol(3)

And the results are:

Estimated power for multiple linear regression

F test for R2 testing subset of coefficients

H0: R2_F = R2_R versus Ha: R2_F != R2_R

Study parameters:

alpha = 0.0500

N = 142

delta = 0.1250

R2_R = 0.1000

R2_F = 0.2000

R2_diff = 0.1000

ncontrol = 3

ntested = 3

Estimated power:

power = 0.9509

Hi ! Following link can be useful:

https://forum.smartpls.com/viewtopic.php?t=4009

Rolando Gonzales Martinez my DVs are loneliness, creativity and resilience these are not IVs as you have stated above. Hence why multivariate regression is being used and not multiple regression and I didn't know what to choose in G*Power