# difference in differences - which statisics (preferrably "Excel-able")

#### seraphodiabolus

##### New Member
<apology: please move this question to "statistics" if it is not appropriate for the group of "regression analysis". I, the OP, read incorrectly the group's name as "statistics analysis" and thought it's the right group to post my question>

I have no idea what statistics to use for the following (difference in differences):
- 4 groups of data with mean and standard error for each group.
- to calculate the difference of (the difference between group 1 and 2) and (that between group 3 and 4)
- hypothetical example:
4 17 y/o males, 5 17 y/o females, 6 18 y/o males, 7 18 y/o females.
I'd like to know if the difference in height between males and females differ between 17 y/o and 18 y/o.
What statistics should I use to calculate the p-value ?

#### j58

##### Active Member
Well, I would actually formualate your problem as a regression problem, so, as far as I'm concerned, you're in the right forum.

You have two factors, sex and age; so a natural analysis would be a linear model with terms for sex, age, and their interaction. You'd be looking to see whether the coefficient for the interaction term were significant. I have no idea whether Excel has a canned function to perform such an analysis. SPSS almost certainly does. Alternatively, why not consult with an actual statistician?

#### j58

##### Active Member
I would actually formualate your problem as a regression problem, so, as far as I'm concerned, you're in the right forum.

You have two factors, sex and age; so a natural analysis would be a linear model with terms for sex, age, and their interaction. You'd be looking to see whether the coefficient for the interaction term were significant. I have no idea whether Excel has a canned function to perform such an analysis. SPSS almost certainly does. Alternatively, why not consult with an actual statistician?