Software

This page is mostly based on a meeting held on Sept 26, 2008

Warning!

Of course, anyone interested in making software decisions should contact the unit's IT person for discussion of feasibility!

General comments

  • General consensus seems to be that user friendliness is the most important quality we look for in a piece of software.
  • Affordable price is another big concern.
  • It may or may not be a good idea to teach Statistics and statistical software in the same class depending on the amount of lab time available. It might be a good idea in some cases to just leave it up to the students to play with Excel following the textbook.
  • People tend to like the piece of software they were trained on. Bruce likes Minitab, Carie likes SPSS.

Specific software. JMP Excel SPSS R more

JMP (JMP 7)

Users: Mike W (BIO300), Eugenia (STAT200)(past)

JMP is developed and sold by SAS institute.

OSes: Mac, Windows, Linux (limited; can't find student licences for linux).

Good Bad
  • Very easy to use. According to Mike W it is easier to learn JMP from zero than to learn how to do Stats in Excel even if you're familiar with Excel in general.
  • Forces users to graph data. It won't do any analysis without showing you the relevant plot.
  • Graphs are interactive. It's possible to change scales, identify points etc using your mouse in real time.
  • Menus for analysis are context dependent. It's good for answering questions like: "What can I do to analyze relationship between this two variables?".
  • Expensive. A student 1-year license for a full version can be had for $50. (Mike W mentioned that 1-year licence costs $100. Not sure what it means. MD)
  • Licensed (Eugenia).
  • Output is not flexible. Hard to change the way graphs look (Eugenia)
  • Hard to create computed variables (Eugenia). Apparently SPSS is good at it.

Notes

Eugenia was under impression that it's impossible tocopy-paste graph directly from JMP to word processing software butMike W has demonstrated that it's possible (at least in the latestversion on Macs).

Pricing

JMP is not offered by UBC IT as of now.

MS Excel

Users: Jonathan (Business students), Eugenia (STAT200)(past), maybe Carie (future)

OSes: Windows, Mac

Good Bad
  • widely available; not free but most PCs have it
  • students are familiar with it so we can focus on teaching Statistics
  • students perceive Excel skills as useful
  • it's good enough for most basic stuff (Jonathan)
  • can be improved with add-ons (Johathan):
    • DDXL from Data Desk (Win or Mac), 15 days trial, US$75.
    • Data Analysis Plus comes with Keller's book
  • performs incorrect analysis or crashes sometimes
  • can only do a very limited number of analyses
  • terminology inconsistent even within the program; e.g. sometimes "significance level", sometimes "p-values"
  • it generally seems like Microsoft does not care about Statistics in Excel at all

Notes:

  • you need to enable Data Analysis add-in before you can do anything
  • according to Carie, there is a "Student version" of Excel that is even dumber than the "standard" version
  • Upd (Dec 1, 2008). Bruce has a contact in Microsfot.
  • OpenOffice.Org Calc, which is open source, can perhaps be used instead of MS Excel.

SPSS (SPSS 17)

Users: Carie (Psychology)

OSes: Windows, Mac

Good Bad
  • Carie thinks it's nice, and so do most psychologists
  • great for recoding variables (Jonathan)
  • Carie secured a 1-year license for her students for C$20
  • Bruce thinks it's counterintuitive
  • less intuitive than JMP (Jonathan)
  • expensive

Pricing

Explanation of different educational packs (PDF): from SPSS website, local copy (same file).

UBC IT will be (from April 2009?) offering SPSS licences for $200/year/computer.

R project (R 2.7.2)

Users: Eugenia (STAT)(past)

Freely available from R-project website.

OSes: Win, Mac, Linux, Unix, anything because it's open source

Good Bad
  • free: as in speech, and as in beer
  • flexible: parameters can be changed, the way graphs look can be changed relatively easily
  • teaches algorithmic skills
  • can do anything with the right package
  • used a lot in academic world across disciplines
  • steep learning curve
  • no standard GUI; there are some projects (e.g. Tcl/Tk-based Rcmdr) but nothing is of industrial quality yet.
  • learning new programming language will distract student from learning Statistics.

Other software mentioned

NCSS (NCSS 2007)

  • Windows only
  • cheap, has Power and Sample Size calculator (PASS), easy learning curve (Jonathan)
  • Pricing:
    • for Education:US$25 per 2-year student licence, US$100 per lab comuter/year. No research!
    • for Academics: US$500 for NCSS, or US$1000 for NCSS+PASS
  • 7 day Free Trial

Minitab (Minitab 15)

  • Windows only
  • Bruce's favourite. Jonathan has use it as well.
  • There's a comparison of Minitab and Excel in Keller's book (Jonathan).
  • Pricing for students or faculty: US$30 for 6 month, US$50 for a year, US$100 to buy.

STATA (STATA 10)

  • Windows, Mac, Linux, Unices
  • Suggested by Jonathan
  • Educational pricing. Students can buy it for US$95 (1 year), US$155 (perpetual) for STATA/IC. Labs: US$300 + (N * 85). Dumbed down version is available for cheaper.

Pricing summary

The following prices are for educational use only. If used for research prices may or may not vary. All prices are in USD.

  Student
6 m
Student
12 m
Student
24 m
Student
perpetual
Faculty Lab Free trial
JMP 30 50     see students   30 days
SPSS 80 (46) 140 (80)   200 200 (via UBC IT) 200/c/y 30 days
NCSS     25   500 100/c/y 7 days
Minitab 30 50   100 see students   30 days
STAT/IC   95 155   155 via GradPlan 300+85/c/y  

a place of mind, The University of British Columbia

Department of Statistics

E-mail: slate@stat.ubc.ca

Emergency Procedures | Accessibility | Contact UBC | © Copyright The University of British Columbia