modeling a regression

Hi all,

I am working on a research project where we are trying to measure how events
experienced by one set of individuals impacts the a set of similar
individuals that did not experience the event (e.g., a contagion effect). So
basically, for each individual-specific event we have approximate 200
non-event individuals. We are now trying to assess the importance of factors
we expect to affect the reaction by non-event individuals to the event.

We thought of the following regression structure:

reaction non-event individuals = intercept + B * VECTOR of characteristics
of the event individual + Error

In terms of data this will be something similar to:

non_A_1 = int + b1*factor_1_event_A + ... + bn*factor_n_event_A +

Error

.... ...

non_A_200 = int + b1*factor_1_event_A + ... + bn*factor_n_event_A + Error

non_B_1 = int + b1*factor_1_event_B + ... + bn*factor_n_event_B +

Error

.... ...

non_B_200 = int + b1*factor_1_event_B + ... + bn*factor_n_event_B + Error

non_C_1 = int + b1*factor_1_event_C + ... + bn*factor_n_event_C +

Error

.... ...

non_C_200 = int + b1*factor_1_event_C + ... + bn*factor_n_event_C + Error

non_D_1 = int + b1*factor_1_event_D + ... + bn*factor_n_event_D +

Error

.... ...

non_D_200 = int + b1*factor_1_event_D + ... + bn*factor_n_event_D + Error
....

....

....



Now the problem: although on the left hand side all variables change for
each observation, on the right hand side, variables come in blocks, so each
event (A, ..., D, ...) results in (on average) reactions in 200 individuals
and values on the right hand side are constant for all the 200 observations.
Is this a problem? Any solution or any other model that not OLS?

Thanks in advance


.

Loading