Wizard good enough for publications.

81 views
Skip to first unread message

niclas...@gmail.com

unread,
Dec 17, 2020, 4:43:13 PM12/17/20
to Wizard User Group
Hi Evan and members. 
I rally love Wizard and hope to love the upcoming alternative. 
I have SPSS that I use to run the analysis I want to use in a paper, I do this cause SPSS is like legit to refer to as using in scientific journals. I think that Wizard is so much more useful unless you are a pure statistician with Im not and don't intend to be. Still I believe that the sometimes conservative approach to statistics in journal might be less accepting for the use of any alternative software, OR? 
have anyone used Wizard as the statistical tool reported in a paper?
The results are the same and its jut tedious to run all again in SPSS. 

best regards 

Niclas 

Allan Gulledge

unread,
Dec 17, 2020, 7:52:14 PM12/17/20
to wizard...@googlegroups.com

 

I don’t think journals care much about the software package, more about what tests, etc. were used (and obviously the quality of the data and appropriateness of the tests).

 

I have cited Wizard in two peer-reviewed publications (e.g., Baker et al., J. Physiology 596.9, 2018).

--
You received this message because you are subscribed to the Google Groups "Wizard User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wizard-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/wizard-users/d57095c4-ce5a-4b5f-9e69-2c5031757fc1n%40googlegroups.com.

sste...@gx.camden.k12.ga.us

unread,
Dec 18, 2020, 9:25:53 AM12/18/20
to Wizard User Group
Short Answer: Use it.

Longer Answer:  The key is using the right measures based on normalcy. 
SPSS: You need to know whether to use parametric or not parametric measures based on normalcy.
Excel: Same as above, plus you must be sure to add the data analysis packages with the right tools. 
Wizard: Takes that into consideration by default, and you can see which method they used. You can also choose your own. 

Even longer: When I was working on my dissertation in 2013, I used Excel, SPSS, and Wizard to compare results. They were consistent across programs. I compared four sets of data. One was normal, one was kurtotic, one skewed to the left, and one skewed to the right. After I taking all of the proper measures in SPSS, per my proposal, I decided to see what the different programs would produce. I noticed that I didn't see the option for choosing a method (in the default) pivot tools in Wizard. Wizard automatically CHOSE the correct one. You can see the Kruskal-Wallis, Pearson,  Spearman etc. in the results.  I kept all of my results but reported the SPSS since that was what the school required. However, I knew then that Wizard would be something I'd keep. 
Reply all
Reply to author
Forward
0 new messages