POWER ANALYSIS Explained: How To Calculate Sample Size For Your Research

how analysis is a critical step in the design of an experiment or study and it's used to determine the minimum sample size required to detect an effect

of a given size with a certain degree of confidence here's how to do it first step determine the research

hypothesis Define it so first determine the effect size the effect size is a measure of the magnitude of the phenomenon you are studying common

metric tricks for determining the effect size include coins D for T tests odd ratio for logistic regression and

correlation coefficient for correlation studies next step is to specify the significance level when you are defining the research

hypothesis specifying the significance level this is determining the probability of rejecting the null hypothesis when it is is true so your

type one error a common Choice as we all know is 0.05 the next step is to choose the desired power and the desired power is

from one to Beta so the power is looking at the probability of correctly rejecting the N hypothesis when it is false so your type two error commonly

used power levels are8 or .9 meaning

there is an 80% or 90% chance chance of detecting an effect if it is there thirdly you want to determine the statistical test so you select the

appropriate test depending on your research design you select the statistical test you will use such as a tea test an anova regression analysis

Kai Square test Etc the fourth step is to calculate or estimate the effect size use prior

research or p studies if possible use data from previous studies to estimate the expected effect size coins guidelines can also provide a benchmark

EG small medium large effects the fifth step is to use a power analysis tool so the software tools that

allow you to do power analysis include statistical software such as G power r or online calculators to input the the parameter

such as the effect the size the alpha the power and test type and calculate the required sample size lastly you want to interpret the

results so you review the input the software will just provide you the minimum sample size needed for your study but ensure that this sample size is feasible giving your resources an

adjust if necessary so looking at this example suppose you are conducting a two sample Le T Test you want to compare the means

of two independent groups so the first thing is to decide what your effect size is and you can use coin D and say let's

say that the studies suggest an effect size of 0.5 which is a medium effect then we have to set the significance level Alpha and we can set that at the

conventional 05 then we need to determine the desired power which is 1 minus beta and so so this is set

at8 then the statistical test is a two sample T test because we're comparing means of two independent groups if it were small groups we would maybe look at

an an NOA so then using G power you open G power G power can be downloaded from G power. hhu dode you can download it for

power. hhu dode you can download it for a Mac or Windows and then once downloaded and it's installed you can start to do the

analysis so you open G power and you select T Test under test and means difference between two independent means

two groups then you input the effect size as .5 the alpha as 005 and the

size as .5 the alpha as 005 and the power as8 then click calculate to get the required sample size per group

[Music]