Dear Mplus-Team, I have data from the same individuals measured at three time points each one year apart. Sample size of this panel study is N= 2700. Currently I have at level 1 the measurement and at level 2 the individual. For my dependent continous variable "attitudes towards the unemployed", I wonder if it is possible to include a variable which specifies the time as such. That is I would like to test for the effects of "newspaper coverage" on "attitudes towards the unemployed" within the described panel design. "Newspaper coverage" is measured objectively and clearly varies between time 1, time 2 and time 3. However, as it is measured objectively the values for "Newspaper coverage" are the same for all individuals at each time point. I now wonder if and how it would be possible to take potential effects of "Newspaper coverage" for the dependent variable "attitudes towards the unemployed" into account, e.g. within a latent growth curve framework. Any comments would be highly appreciated, best regards Elmar
bmuthen posted on Monday, March 21, 2005 - 1:21 pm
It sounds like "newspaper coverage" could be used as a time variable and therefore growth modeling would be relevant. Growth modeling does not necessitate using real time (or age) as the time variable but anything that is relevant to cause change over time in the dependent variable. Just remember that the centering issues remain the same - for instance, you may want to transform the coverage variable so that it is zero at the first time point.