## Question creating a statistical Analysis for treatment validation with placebo cond.

Hi there, promising board of infinite knowledge,

1. Randomizing the Conditions

This is what I want to do. One individual, two conditions.
A is treatment.
B is fake-treatment. 10% of the (continuous) session, to not withdraw the treatment!

My Idea:

1. Randomize Fake-Treatment Onset and Length to get rid of sequential effects. Problem: Occurence of Condition B occurs to a much lower probability than Condition A, but the length of a full Session is not set a priori.

*Question 1: How to randomize the Onsets of the Fake-Treatment Trials, so it is in some way equally distributed over a session with indefinite length?

*Question 2: How to randomize the Length of Fake-Treatment Trials, so that the the length is not too short (<1s), nor too long (>1min), so it is equally distributed over a session with indefinite length?

2. Analysis

My Idea:
Add up all the seperated trials of each condition to create a continuous plot and do a Regression Analysis.

*Question 3: Samples in Condition B is way less than Condition A (~10%). Is it even possible to compare these two with such different number of samples in a Regression-analysis? (Assumptions? Robustness?) Should I calculate a rel(t)?

*Question 4: (might be a stupid question), but is it automatically a multivariate Regression, if I compare the 2 Conditions with each other, instead of the gradient of each condition separately with 0?

*Question 5: Or is there a better way to analyze this case? ANOVA? I think CI, Effectsize and Power is important.

You would really help me a lot with this If you need clarification on some things, I'd like to help out!