Can you give me the steps/calculations for figuring Cronbach's alpha on a pilot survey sent out by email via Survey Monkey that had 18 5-point Likert scale questions and 72 respondents? The pilot survey was sent to individuals as similar as possible to the ones who would be in my main study (but not actually in the main study). Also, I can break my questions in the survey into three similar groups. Based on what I have read, it looks like I could use Chronbach's alpha on all 18 questions and 72 respondents first and then I could do 72 respondents with each of the three similar groups and include all in my dissertation. Would this be beneficial? Based on the score between 0 and 1 will determine if each of my pilot survey questions are reliable and valid. If the percentage is low, should I just take out that question from my survey before sending it out for the main study or revise it? If it is an acceptable percentage between 0 and 1 I know I just keep it