I currently running SPSS and PLS to test my study hypothesis. the case that there are two variables that are associated it (theoretically) and they are independent variable. Thus, i need to combine that two independent variables to become one independent variable. Furthermore, that one independent variable would like to test the relationship with one dependent variable.
Please enlighten me
You cannot combine two independent variables to create an independent variable to those two. However you create it will be correlated to the original ones. For studying the correlation of this new variable with a third variable you can just plot them one against the other, then se which relation they have (if they have it) and fit a regression line from which can calculate the correlation coefficient (r).
As Jokin stated, you cannot combine two independent variables to create one independent variable. If it is theoretically related, you can create a latent variable of these two independents variables and use that latent score as one variable. Ideally, you need three independent variables for creating one latent variable, but two would be sufficient for testing your hypothesis.
To my mind the problem (question) is not clear: How many variables are there and what is the relationship between these variables that you want to investigate (or that you expect theoretically)?
If you talk about a combination of two variables:
(a) You can combine them in different ways, for example creating a total score by simply summing them up (what often is done with items of a scale) or by using them in a latent variable model (a measurement model as D. Eastern Kan Sim seems to suggest -- however, a measurement model with only two indicators needs some restriction to be identified, for example equality of error variances or loadings) to predict a third variable by using this total score or latent variable. Of course, you can combine two variables in that way.
(b) But it may also be that you want to use to two correlated variables as predictors of a third "dependent" variable as in a multiple regression model (here you could calculate a multiple correlation R). But then I don't understand where your problem is.
In the subject title of you post you write "which theoretically stated that two variables are associated with it". Do you mean to say that you expect that two variables are associated with a combined variable?
But in explaining you question you also state that "there are two variables that are associated ..." (I don't understand the "it" next to it) "... and that they are independent variable" (perhaps you wanted to write "variables"?). If you mean to say that they are associated and independent from each other this is a contradiction -- if they are associated (= correlated) they are not independent (= uncorrelated).
As I wrote in the beginning, your question is not clear. Perhaps you can draw a conceptual model showing the variables and the relationship between these variables that you want to investigate (or that you are expecting theoretically).
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA