SmartPLS to create Latent Variables for logistic regression
Posted: Sun Dec 14, 2008 9:42 pm
Dear all,
my research models have binary output variables, therefore I cannot use PLS to estimate the structural relations between independent and dependent variables. As far as I understood I have to use logistic regression instead.
I however still need to create LV from formative and reflective measurement models. Can I use PLS to estimate the measurement models (confirmatory factor analysis)? Is it best to estimate the measurement models of LV one by one, together at once, or can I even create a "full structural dummy model" assuming linearity of dependent and independent variables?
Thanks alot for your help
Florian
my research models have binary output variables, therefore I cannot use PLS to estimate the structural relations between independent and dependent variables. As far as I understood I have to use logistic regression instead.
I however still need to create LV from formative and reflective measurement models. Can I use PLS to estimate the measurement models (confirmatory factor analysis)? Is it best to estimate the measurement models of LV one by one, together at once, or can I even create a "full structural dummy model" assuming linearity of dependent and independent variables?
Thanks alot for your help
Florian