Effect of SD under the null hypothesis

Categories: general
Outset of the problem I recently discussed generating data for a power/design analysis and the following problem came up. Does the standard deviation (SD) has an influence on the test statistics (t-value, p-value, BF) under the null hypothesis? The noise in the or the SD of the data should have no effect. To understand this, look at Cohen’s D, which is the mean divided by the SD. Under the null hypothesis will always be zero because the mean is zero regardless how big the SD is.

Read More →

Design analysis for noveltyVR

Categories: experiments
Aim of this document The aim of this document is to conduct a design analysis for a fixed N design using one-sided Bayesian t-test. This analysis will inform our choice of N for a registered report. The planned experiment is a 2 x 2 x 2 design with one between- and two within-subject factors. Without going into too much detail, there will be a novelty and a control group (Factor N) and we will examine recollection/familiarity (Factor M) for weakly/strongly learned words (Factor E).

Read More →

Prior deliberation on analysis

Categories: experiments general
Aim of this document The aim of this document is to accompany the design analysis for noveltyVR (see here) and to shed light on the deliberations concerning the planned analysis. The planned design is a 2 x 2 x 2 design with one between and two within subject factors. There will be a novelty and a control group (Factor N) and we will examine recollection/familiarity (Factor M) for weakly/strongly learned words (Factor E).

Read More →