===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARDI'll weigh in briefly. I teach a grad course in research methods and stats in which I've used SPSS as a mandate of the program. Recently, the program, recognizing that the likelihood of any of the students actually doing research in the future was close to nil, has moved to the use of Excel. That's good to the extent to which student's have more familiarity with Excel and zero with SPSS; so they've not just been forced to learn statistics and research methods, but a completely foreign software.
With that said, in preparation to move from using SPSS for assignments, I'm exporting the datasets from SPSS to Excel and finding that the structure of data in SPSS is not conducive to analysis in Excel. Furthermore, Excel's structure seems clumsy. For example, to do an independent samples t-test in Excel, entering the data ranges requires that either two columns/variables be created - one for each group; or if there is a group variable (e.g., treatment and control), the ranges require that users copy first those values for one group into a range, and then for the other group into the second range. The likelihood for error with all the copying and pasting is increased quite a bit.
Then there's the problem of completeness of analysis. The methods used in either the basic correlational analysis or the Analysis Tool Pack gives only the correlation, totally devoid of significance, intercept, standard error, or confidence interval limits.
I'll use Excel because it represents a move the program has now mandated, but with all the difficulty inherent in the lack of familiarity with SPSS, it probably is more intuitive tool for analysis in the long run.
Brian Dates
From: SPSSX(r) Discussion <[hidden email]> on behalf of Reka Solymosi <[hidden email]>
Sent: Tuesday, June 12, 2018 5:59:39 AM
To: [hidden email]
Subject: Re: SPSS vs R===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARDDear John,
Thank you for all the detailed comments. While I never taught in SPSS (so I cannot comment on that experience like Juanjo) I did receive all my training in SPSS. It works in a university setting great, but once I left academia, it was not very useful for me. Licences for SPSS are expensive, and are *per PC* from what I remember. I worked at a local council as a transport planning analyst, and then later as a crime analyst in London. Neither were in well-funded places. I had access to Excel, and MapInfo. I self-taught R and it was great not only because of it being free and because of all the support and community around the open source ethos of it, but because of the flexibility. There are an ever growing number of packages available for R for free, which means that it can be used yes for graphics, but also for statistical modelling, for network analysis, for text analysis, for data mining, as a GIS, to build interactive dashboards, to build presentations, and mostly for reproducible data manipulation and analysis. I have yet to find any other tool (other than maybe writing SQL and more recently Python) that allows such a range in data querying, manipulation and cleaning.
That said, bringing it back to teaching in academia, I agree completely with your suggestion that students need an introduction before throwing them into R. I teach the first semester course (the one that students take before moving on to the R based one) in Excel. I chose Excel over SPSS simply due to the fact that no matter where they will go, there will be Excel, even in the poorest local council. It gives enough of an introduction to data analysis that they can then move on to R. So in that sense we are following the suggestion you make, that they start with an easier route to contingency tables via excel, and then move on to R. Excel can actually be a very powerful tool for data analysis if used right, and also an accessible route in to interpreting, understanding, and examining data.
To be honest I think any way we can get students interested in data analysis is good. I don’t hugely care if it’s SPSS, Excel, R, STATA, etc, I think the most important is the core concept and getting students interested in data analysis. I think we manage to achieve that here, we even have a few of our graduating students this year applying for a masters in data science, which coming from a criminology programme I think is somewhat unusual! But I also appreciate that we are standing on the shoulders of giants in a way. I for one have made so much use of the work and support from those like yourself with numerous years of experience in teaching and researching the best ways for teaching quantitative methods. I think there is a lot of very valuable material there, and I think that it can be applied to any platform. I used many materials and resources that showed examples in SPSS, and translated those to Excel or R.
Ideally some resource for sharing platform agnostic resources could actually be compiled and perhaps shared around? My material on using Excel is all available here: https://maczokni.github.io/
MSCD/ I know that last summer I sent around quite a few requests for help to the quantitative methods teaching list, and I would be happy if I could pay back somehow. Maybe some central open-source repository of training material that can be applied by any one to any platform they choose to use, but that is based on all the work everyone is doing, to bring us all together?
Let me know any thoughts!
Many thanks,
Reka
From: John F Hall [mailto:[hidden email]]
Sent: 12 June 2018 10:24
To: [hidden email]
Cc: Juan Medina-Ariza; Reka Solymosi
Subject: SPSS vs R
As part of the ESRC-Nuffield Q-step initiative (http://www.
nuffieldfoundation.org/q-step ) to improve quantitative methods teaching in undergraduate social science degrees in the UK, a new one-semester quantitative criminology course is being taught to undergraduates at Manchester using R, mainly because of its graphic capabilities.See http://jjmedinaariza.github.
io/R-for-Criminologists/ for full course notes.
In the accompanying pedagogical rationale Professor Juanjo Medina explains why (although he admits that R has problems with crosstabs at which SPSS is excellent.)
It is simple. I was sick to the bone of teaching with SPSS. Why should I bother to be a publicist for IBM? . . .But I never quite fell in love with SPSS, its ugly graphic system, its patched up inconsistent menu design, etc. Its whole architecture, easy in the eye for casual users, seem designed to encourage bad habits among future analysts. In the meantime I continue using a variety of tools for my own research (STATA, MPlus, etc) until I met R and fell in love with it.
See https://rawgit.com/
jjmedinaariza/LAWS70821/ master/rcommander.html# motivation
As someone who had to teach and assess Data Management and Analysis (at both undergraduate and postgraduate level) within a tight 13-week semester I still feel that SPSS is an easier, and better, route to Quantitative Methods (via contingency tables rather than multivariate statistics) perhaps leading to R at a later stage.
John F Hall MA (Cantab) Dip Ed (Dunelm)
[Retired academic survey researcher]
Email: [hidden email]
Website: Journeys in Survey Research
Course: Survey Analysis Workshop (SPSS)
Research: Subjective Social Indicators (Quality of Life)
| Free forum by Nabble | Edit this page |