Determining ideal sample size, testing normality in SPSS, bonferroni adjustment

R

rsuri

Guest
#1
Hello all,
I am new to the forum. I had a bit of training in statistics and research, particularly statistics in psychology, but I have some basic questions that are holding me back. I would appreciate the help.

1. How can I determine the ideal sample size for a study (upper and lower limits)?
2. How do I test for normality in SPSS? Does it vary for different hypothesis/analysis using the same dataset or does it remain the same (is normality dependent on DVs and IVs or not)?
3. When exactly do we need to use Bonferroni adjustment? I know it is debatable to use (or not use) Bonferroni adjustment when the hypotheses are independent of each other, but I was wondering if someone could elaborate a bit more for me (explain in layman terms)?

Thank you again. I have more questions, but I guess this would be a good starting point
Raghav
 

Karabiner

TS Contributor
#2
1. How can I determine the ideal sample size for a study (upper and lower limits)?
You can perform an a priori power analysis.

2. How do I test for normality in SPSS?
SPSS offers the Kolmogorov-Smirnov test, which
preferably is performed with Liliefors-correction
("explore" menue). Maybe SPSS also offers Shapiro-
Wilks somewhere. Usually, using plots such as Q-Q
plot ist preferable over a significance test.

Mind that in almost all cases testing variables for
normality is not necessary.There are some instances
in which procedures (such as linear regression etc.)
require normally distributed prediction errors
(residuals)
, at least if sample size is small.

With kind regards

K.