I think that I your N is less than 50, you can use T-student for ndependent samples or W Wilcoxon (it's depends on your sample normality). But if your N is more than 50, you should use ANOVA (the test used depends on your sample normality, too).

Hello Carla, Thanks for your response. I have N= 49. Can i go with ANOVA. In my case, The respondents themselves compare between two environment and i am just projecting the output with a statistical method.

Mr. sateesh , Tell me , r u using same participent for diffrent envirinment or diffrent participant for diffrent environment?.

Same participant in different environment. I have decided to go with ANOVA two factor without replication. Any more suggestions are most welcomed. Many thanks!!

If u r using same participant on diffrent environment. your sample is dependent, in this case you use Paired sample T test. Anova is appropriate when u have more than two environment.

If you have N participants, each responding to the same two envionments, you have 2N observations.

Two-way ANOVA without replications is possible if you use the interaction as error. You have 2N-1 df: N-1 for respondents, 1 for environments and N-1 for interaction/error. Testing that environments have no effect is done using F(1,N-1).

Application of the "paired t-test" to the differences is done using a t-test with N-1 df.

Don't worry. It does not matter which of both methods you apply, since what you are doing is essentially the same.

Although it looks like different, it is not. Note that the squared value of t(N-1) equals F(1,N-1) !