There are a couple of mistakes. First, it confuses 'not rejecting the null' with the 'null is true'. A false null might not be rejected due to a small sample. Second, the effect size in this case should be computed with the SD from the pre- or baseline or alternatively with the pooled SD from pre and post. The SD of the difference depends on the intra-subject correlation, so it's used for the test but not for the effect size.
Your lectures are amazing and very easy to understand. Thanks.
There are a couple of mistakes. First, it confuses 'not rejecting the null' with the 'null is true'. A false null might not be rejected due to a small sample. Second, the effect size in this case should be computed with the SD from the pre- or baseline or alternatively with the pooled SD from pre and post. The SD of the difference depends on the intra-subject correlation, so it's used for the test but not for the effect size.