Hypothesis Testing - 2

Cohen’s d tells you

how big the differences is between 2 means in terms of SD (rather than whether the differences is statistically significant)

Percentage of variance accounted for (r²) tells you

how strong the relationship or effect is

Confidence Intervals is the range of values that are

most likely to contain the true population parameter if you repeated the study many times

basic logic of hypothesis is to always assume

treatment doesn’t work (the null hypothesis) [pessimistic]

Z-critical value increase as

α level decreases

When the α level is low, it becomes harder to

Reject the null hypothesis bc the rejection regions (outside the boundary, the two tails) move farther out

The smaller the α level, the

more trust you have for the null hypothesis

When T or Z score equals 0,

no difference → fail to reject (retain) the null hypothesis

Within the T or Z region like within ±1.96 equals

fail to reject (retain) the null hypothesis

Outside the T or Z region like anything that’s more or less than 1.96 equals

reject the null hypothesis

For big Z score, you want: …. population SD (σ)

small

For big Z score, you want: …. sample size (n)

big

For big Z score, you want: …. between sample mean (M) and pop mean (μ)

big differences

We want sample mean (M) to be

bigger than pop mean (μ) to get a more positive number

If p < 0.05 →

there is a significant differences meaning we reject the null hypothesis

is p > 0.05 →

there isn’t significant differences meaning we fail to reject (retain) the null hypothesis

T-critical value decreases as

df value increases (opposite of z)