# Compare differences between exp. and contr. of two groups

#### rothmann1510

##### New Member
Hi!
I already used lots of search engines without finding an answer to my problem.

I am testing two methods for manipulating religious belief. For each of the methods there exists an experimental and a control group. So there are in total four groups (between-subject design).
My hypothesis is the following:
Exp1 - Contr1 > Exp2 - Contr2

How do I test if one of the differences is bigger than the other? The interaction between the factors "Exp. vs. Contr." and "method of manipulation" is significant.

Cheers,
Konstantin

#### Silicona

##### New Member
how do I test if one of the differences is bigger than the other?
The interaction between the factors "Exp. vs. Contr." and "method of manipulation" is significant.
Whether difference 1 > difference 2 or vice versa must be clear from the descriptive statistics.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
How is religious belief formatted (categorical [binary], continuous, ordinal)?

#### rothmann1510

##### New Member
Religious belief is measured on a 7 point Likert-scale.

So the significance of the interaction is enough for me to simply consider the descriptive statistics? This would make things really easy.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Well, ordinal data can be tricky. So you ran linear regression modeling? If so, did you test the model assumptions to make sure it was appropriate. Also, what is your sample size. Lasly, was your model: y = B1(treatment status) + B2(control status) + B3(treatment * control).

Provide many more details, I am not completely following the study design!!

#### Silicona

##### New Member
Religious belief is measured on a 7 point Likert-scale.
This seems to confuse two concepts. Was ist a Likert scale (interval scaled), composed of several Likert-type items? Or was was it just 1 Likert-type item (ordinal scaled)?