08-11-2014 10:51 AM
Just wanted to have your views on how you do Significance tests, do you prefer Excel or SAS? I have recently learnt doing t-tests and ANOVA in SAS Base but at work, we use Z-Scores in Excel to compare different groups for our Campaign analysis. I find it hard to understand the excel Z Scores and find t-tests in SAS much easier. Just wanted to know what's the best practice to do group comparisons.
08-11-2014 12:43 PM
Granted this is now 15 years old, but you should look at http://users.df.uba.ar/cobelli/materias/laboratorios/excel97.pdf. The last line in the abstract is:
Persons desiring to conduct statistical analyses of data are advised not to use Excel.
Now there have been improvements, but a search on BD McCullough reveals the following http://www.forecastingprinciples.com/paperpdf/McCullough.pdf from 2006.
There is a reason SAS, R, SPSS, S Plus, Matlab, Mathematica, Python, Ruby and Java apps all exist. A large part of them have to do with the statistical inadequacy of Excel.
08-17-2014 05:22 PM
08-18-2014 07:24 AM
Just because it is common doesn't always mean that the results are the best you could get for the money. Note Bruce McCullough's point about tabulating the normal distribution and calculating p values.
I use Excel a lot, but I don't trust it to give me statistical answers.
Need further help from the community? Please ask a new question.