Regression coefficient has measurement units like ( Kg. product per gm of treatment ), now if the independent variable ( x ) represent by ordinal , what formula can be use to calculate regression coefficient .
If I understood your question correctly, you can indeed use any statistical software package (R, SPSS, STATA) to calculate a linear (o non-linear) regression with those variable types. Ordinal variables such as 5-point Likert scale (1 to 5) are considered as continuous in many fields of the social sciences. In case you do not feel comfortable with this, you can normalize your independent variable (that is obtain its Z score) and enter that as predictor in your regression equation. In case your ordinal variables are not scalar, because they represent completely different groups, you may need to count an AN(C)OVA, or create "dummy variables", in which each dummy variable represents a group (e.g., 0 = Female / 1 = Male). I hope this helps. Good luck in your research
Lucas
Dear Mr Lucas
Thank you for this information , I know the solution for this case by spss , but I need a formula to calculate regression coefficient with out software
Hey,
the formul stays the same, if your dependent variable is continuous: beta_dach= (X'X)^-1 X'y
But instead of using the ordinal independent variable directly you dichotomize it. Don't forget to drop a reference category if you have an intercept included to avoid exact multicollinearity.
Best regards,
Florian
If you want to know how you can estimate the regression coefficient, you should see the formula. The formula is B=(X'X)^-1*(X'Y)
which X is your independent variabel
Y is your dependent variable
B is regression coefficient
X' is transpose of X
(X'X)^-1 is the inverse of (X'X)
I'm sorry but what you want is not ordinary least squares regression but rather ordinal logistic regression. For full details see the attached notes. Best regards, David Booth
I apologise for misunderstanding your question. The notes I sent should still be helpful though. Best, David
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression