So, the key difference between an independent samples t-test and a paired-samples is…?
…the way standard error is calculated.
dependency [between data points] will always distort our analysis from reality. In what way?
it exaggerates in favour of our hypothesis, or against it