Unbounded variables and regression

[Jean-Philippe NAU]

I'm sorry I forgot to edit the title of my email.

2009/8/19 Jean-Philippe NAU <[hidden email]>

Dear all,

I read in Field "Discovering Statistic Using Spss" that the outcome variable of a regression should be unbounded. Despite that I've seen regression ran on five or seven points scales...Could you help me understand "how wrong" it is to run regression on bounded scales?

Thank you very much,

Best regards,

Jean-Philippe