Author Archives: Peter Edelsbrunner

Peter is currently doctoral student at the section for learning and instruction research of ETH Zurich in Switzerland. He graduated from Psychology at the University of Graz in Austria. Peter is interested in conceptual knowledge development and the application of flexible mixture models to developmental research. Since 2011 he has been active in the EFPSA European Summer School and related activities.

Introducing jamovi: Free and Open Statistical Software Combining Ease of Use with the Power of R

For too long, Psychology has had to put up with costly, bulky, and inflexible statistics software. Today, we’d like to introduce you to a breath of fresh air: jamovi, free statistics software available for all platforms that is intuitive and user-friendly, and developed with so much pace that its capabilities will potentially soon outrun SPSS.

Screenshot of jamovi

 

As can be seen above, jamovi has a beautiful user interface with some very handy features: It does real-time computation and presents and updates results immediately with beautiful figures and neat APA tables. These results can then be copy-pasted into your editing software such as Word. Basic analyses (e.g., t-tests, ANOVAs, correlations, contingency tables, proportion tests) are already available and more will be arriving shortly. What’s more, packages from the powerful R software can be easily adapted so that they can be used within jamovi’s beautiful user interface. In this way, jamovi can give you access to the power of the R language, but without having to learn the R syntax. For those wanting to learn R, jamovi can help there too: with just one mouse click jamovi delivers the R syntax underlying each analysis.

Another gadget of jamovi is live data management: You can edit your data directly in the software, and if you change something, results that depend on these changes are immediately updated in the output window. Imagine how this would work in SPSS: Change a data point, click through all the menus again or re-activate the relevant syntax, manually delete the old output, all in order to get ugly figures and tables that need additional time investment to become beautiful or in accordance with APA-format; with jamovi, these strenuous days are over!

One particular and useful type of analysis is also already available in jamovi: The TOSTER module. This analysis allows testing whether data support a null hypothesis (e.g., the absence of a meaningful effect), which is often what we want to know but not possible to test with most statistics packages.

Thus, there are many reasons to install and use jamovi right away, and if you want to help your peers, you can develop your own R-based jamovi modules and make them freely available for everyone in the jamovi store.


Interview with Jonathon Love, jamovi co-founder and developer

jamovi might remind you of another recently established free stastistics software: JASP. Indeed, Jonathon Love, Damian Dropmann, and Ravi Selker were all developers of JASP who now develop jamovi. The two software packages may at first seem similar, but they emphasize different functionality. This means both packages will continue to be developed, and users can enjoy the benefits of both. Let’s see what Jonathon, former lead developer and designer of JASP, and now one of the jamovi core developers, has to say about this and more:

During our last interview, you were lead developer of JASP. What was your motivation to start jamovi and what happened since then?

So developing JASP was really fabulous, and something we all really enjoyed doing. But we did find that our ambitions, hopes and dreams went beyond JASP’s original goals. JASP has always been heavily focused on Bayes, and we wanted the freedom to explore other statistical philosophies.

At the same time, a number of technologies had matured to the point where we could build a more advanced software architecture in a much shorter amount of time. When I began JASP, I had to choose between older, “tried and true” technologies (C++ and QWidgets), and the newer, up and coming HTML5+js technologies. At the time, I concluded the newer technologies just weren’t mature enough for a large project like JASP.

Fast-forward a few years, and everything has changed. HTML5+js have overcome leaps and bounds and have become a capable, mature framework. Similarly, other developments have made things that before were very difficult, much more straight forward. For example, the R6 R package has enabled us to create a much more elegant analysis framework, allowing rich graphical analyses to be developed in much less time, and to support data-editing. Similarly, it has made it feasible to provide one of the most requested features: R syntax for each analysis.

So the decision to begin jamovi was a combination of ambitions beyond the JASP project’s core goals, and seeing the opportunities that newer technologies provided.

You launched jamovi a couple of weeks ago and so far only few analyses are available. When will jamovi offer a scope of analyses comparable to SPSS?

So we actually think SPSS is overwhelming, making the user navigate a huge labyrinth of menus filled with analyses most people will never use. We do want to provide a lot of analyses, but we’ll do it in a different way. Our intention is to provide all the basic analyses used in undergraduate social science courses in the next few months, and we have the ambitious roadmap of being a viable (and compelling!) alternative to SPSS for the majority of social science researchers by August, providing all these analyses, and providing complete data-editing, cleaning, filtering and restructuring.

For additional, or more specialised analyses, we hope to build a community of developers providing analyses as “jamovi modules”. jamovi modules are R packages which have been augmented to run inside jamovi and provide analyses with a user-interface. Importantly, these modules still function  as R packages making the analyses usable from both platforms. People are then able to publish jamovi modules they create through the “jamovi store” (and CRAN), making them available to anyone. We recently worked with Daniel Lakens to produce a jamovi module of his TOSTER package, and that’s come together very nicely. There’s a few more modules in development that we know about, and you can expect further announcements in the coming weeks!

One of the neat things about the jamovi store is that it allows us to keep jamovi itself simple, and allows people to only install the analyses that are important to them. For those familiar with R, this is exactly how it works with CRAN, and we hope to duplicate its success, but for analyses with rich, accessible user interfaces.

jamovi is built on the idea that developers create jamovi modules for their R packages. Why should they do that?

There are two answers here: for science, and for themselves.

For science, because not everyone is, can be, or needs to be an R programmer. People have strengths in different areas. As long as new analyses are only available to people who can work with R, there are a lot of scientists who will be left behind. So I think it is imperative that we make new and advanced analyses available and accessible to everyone – that’s one of the core motivations for jamovi.

But creating jamovi modules can also be significant for the authors of analyses. One of the most significant metrics in science is how widely someone’s work is used, and a jamovi module ensures an analysis is accessible to the greatest number of people possible. So there are good career incentives for people to develop jamovi modules too.

Therefore, we encourage R developers to look into developing jamovi modules. The jamovi developer hub provides tutorials walking you through the process of writing a jamovi module: dev.jamovi.org, and if people would like help or advice, we can pair them up with a “dev mentor”. There are also forums where people can post questions. We’re keen to support the developer community in whatever way we can.

Readers of this interview will inevitably compare jamovi to JASP. What do you see as jamovi’s most distinctive features? Where can you borrow from JASP’s approach?

So our distinctive features are: data-editing, our R syntax mode, and the jmv R package.

Data-editing is one of my favourite features, because it takes something crazy complex, and makes  it seem really easy! You’ll notice that if you run an analysis, say descriptives, and then start changing some values in the data view, the descriptives analysis updates in real-time. This in  itself is cool, but you’ll also notice that only the columns in the descriptives analysis affected by the data changes are updated. Under the hood, jamovi is dynamically figuring out which values in the results need to change in response to the data changing – and only recalculating those. This is pretty neat.

R syntax mode is another favourite. jamovi can be placed in “syntax mode”, where the R code for producing each analysis in R is provided. This is super-cool, because it makes it easy for people to see and learn R code, and it also allows them to copy and paste the R code into an interactive R session. This allows people to make the jump to R, if that’s an area they are wanting to develop skills in.

Our jmv R package is the other half of “syntax mode”; an R package which provides all the  analyses included in jamovi. This is awesome, because it means that a single R package will cover entire undergraduate social sciences programs. In the past, doing something like an ANOVA with all the contrasts, assumption checks, post-hoc corrections, etc. required in the order of 7 packages. So it’s been exciting to bring all of those elements together, and make them simpler for R users as well.

With respect to JASP’s approach, Eric-Jan Wagenmakers and the JASP guys have put a lot of effort, and continue to put a lot of effort into making new Bayesian analyses accessible to a broader audience. Their analyses represent a truly fabulous contribution and we’ll definitely be keeping a keen eye on what they get up to. You should too!

What are the biggest challenges ahead in developing and disseminating jamovi?

The chicken and egg problem. Always the chicken and egg problem!

People are reluctant to adopt a new platform when not all the supporting materials, videos, textbooks, etc. have been created yet. At the same time, the content creators are reluctant to provide supporting materials, because people seem reluctant to adopt it. In this way, markets tend to resist change, and overturning the status-quo often poses a frustrating challenge.

This phenomena isn’t just limited to software; you’ll find that it applies to many areas in science. Of course, change can, does, and must take place, and so the challenge is putting all the pieces in place so that new ideas, new paradigms, and new pieces of software can be adopted. In my view, this is almost always the biggest challenge, but it must be overcome — progress depends upon it!

So it’s been pretty exciting seeing the level of support coming from the community. We’ve had a surprising number of very promising talks with authors and publishers. I think we’ll have some pretty exciting announcements in the coming months, and it looks like we’re well on the way to hatching that chicken … or egg … or whatever.

How is jamovi being funded? How can users be sure of its continuing existence?

So at the moment jamovi is still in the early stages, and our emphasis has been on demonstrating  that we have the sort of trajectory that people can get behind, and so we currently don’t have a lot of funding. I work for the university of Newcastle, and volunteer my time on jamovi, and the same applies to the other core developers. However, people can still feel confident in the future of jamovi.

We expect to provide a complete and practical alternative to SPSS by August – with full data-editing, filtering, restructuring, the works. At that time, jamovi could be considered “complete”. We don’t intend on stopping developing then, but if we did, jamovi would still be (in our view) one of the best tools available for social scientists, probably for years to come. It won’t require a lot of effort to continue to maintain jamovi into the future, and people can feel confident that jamovi will be here for years to come. (There’s a persistent myth, that the maintenance of software once written requires substantial resources to maintain. Indeed, in proprietary software it’s often a problem that old software “just keeps working”, and it’s hard to persuade customers to pay for newer versions!)

Having said all that, we are keen to develop funding and business models to support additional development of jamovi – and we have big plans going into the future. In the short-term, our efforts are concentrated on creating a viable alternative to SPSS, but longer term we want to provide a range of additional paid services that make the lives of researchers easier. jamovi itself will always (and must!) be free and open-source, but there’s a range of areas where we think we can provide services to make researchers more productive, and where it would be reasonable to charge a fee.

We’re also keen for benefactors, so if you or your institution benefit or stand to benefit from the work of the jamovi team, you could consider making a financial contribution to our work. Such a contribution would allow us to ramp up development, and provide a greater range of features. If there are particular features and analyses which are important to you or your institution, you could sponsor their development (e.g.,  reproducibility in a spreadsheet? We’d love to do that!). Do drop us a line.

How does the curious reader get started with jamovi?

jamovi is pretty straight forward to use, and it contains several example data-sets that make it easy  to get up and running. I’d recommend downloading and installing jamovi, and just playing around with it. We also have a user-guide, complete with neat little videos demonstrating the basic features. If you’ve used SPSS before, you should find the user interface concepts quite familiar; like the dragging and dropping of variables for an analysis. It’s designed to be easy and straight-forward to use, and if you find this not to be the case, do drop us a line in the forums. We’re very keen for feedback, and to make jamovi the best it can be!

 

Peter Edelsbrunner

Peter Edelsbrunner

Peter is currently doctoral student at the section for learning and instruction research of ETH Zurich in Switzerland. He graduated from Psychology at the University of Graz in Austria. Peter is interested in conceptual knowledge development and the application of flexible mixture models to developmental research. Since 2011 he has been active in the EFPSA European Summer School and related activities.

More Posts

Facebooktwitterrss

Structural equation modeling: What is it, what does it have in common with hippie music, and why does it eat cake to get rid of measurement error?

Do you want a statistics tool that is powerful; easy to learn; allows you to model complex data structures; combines the test, analysis of variance,  and multiple regression; and puts even more on top? Here it is! Statistics courses in psychology today often cover structural equation modeling (SEM), a statistical tool that allows one to go beyond classical statistical models by combining them and adding more. Let’s explore what this means, what SEM really is, and SEM’s surprising parallels with the hippie culture! Continue reading

Peter Edelsbrunner

Peter Edelsbrunner

Peter is currently doctoral student at the section for learning and instruction research of ETH Zurich in Switzerland. He graduated from Psychology at the University of Graz in Austria. Peter is interested in conceptual knowledge development and the application of flexible mixture models to developmental research. Since 2011 he has been active in the EFPSA European Summer School and related activities.

More Posts

Facebooktwitterrss

Bayesian Statistics: What is it and Why do we Need it?

prlipohellThere is a revolution in statistics happening: The Bayesian revolution. Psychology students who are interested in research methods (which I hope everyone is!) should know what this revolution is about. Gaining this knowledge now instead of later might spare you lots of misconceptions about statistics as it is usually instructed in psychology, and it might help you gain a deeper understanding of the foundations of statistics. To make sure that you can try out everything you learn immediately, I conducted analysis in the free statistics software R (www.r-project.org; click HERE for a tutorial how to get started with R, and install RStudio for an enhanced R-experience) and I provide the syntax for the analysis directly in the article so you can easily try them out. So let’s jump in: What is “Bayesian Statistics”, and why do we need it? Continue reading

Peter Edelsbrunner

Peter Edelsbrunner

Peter is currently doctoral student at the section for learning and instruction research of ETH Zurich in Switzerland. He graduated from Psychology at the University of Graz in Austria. Peter is interested in conceptual knowledge development and the application of flexible mixture models to developmental research. Since 2011 he has been active in the EFPSA European Summer School and related activities.

More Posts

Facebooktwitterrss

Advice for the Next Generation of Researchers in Psychology from an Experienced Editor

In April 2012, at the conference of the Austrian Society for Psychology (ÖGP) at the University of Graz, Robert Kail – experienced researcher and editor for one of the flagship journals in Psychology, Psychological Science – gave an insightful presentation and discussion targeted to give advice about manuscript preparation and the submission process to junior researchers in psychology. His presentation was organized around several key questions taken from a survey that students of the association for psychological science (APSSC) had conducted. The following main topics of his presentation will be discussed in this post: turning a thesis into a paper, writing a clear introduction, choosing the right title for a paper, and what to consider during the submission phase.

Continue reading

Peter Edelsbrunner

Peter Edelsbrunner

Peter is currently doctoral student at the section for learning and instruction research of ETH Zurich in Switzerland. He graduated from Psychology at the University of Graz in Austria. Peter is interested in conceptual knowledge development and the application of flexible mixture models to developmental research. Since 2011 he has been active in the EFPSA European Summer School and related activities.

More Posts

Facebooktwitterrss

Research as an international project

As a psychology student you have to face certain barriers, when you have the possibility to do research. Those barriers mostly concern the university you are studying at. If your university does not provide you the opportunity to research in a field of your interest, the chance of working in an international research team on a joint project might be a good option to develop your research skills and discover the world of academia on an international level. If you can find such a team, then you are lucky, because 1) it is great to work together with people from different cultures who are all interested in the same topic and can share their expertise 2) it gives you possibilities to work on aspects of a topic that you could not do within the cultural and methodological framework of just one university.

But the opportunity of an international research team reveals some challenges you will have to face at several critical points throughout your work. In the following post, those challenges are summarized.

Continue reading

Peter Edelsbrunner

Peter Edelsbrunner

Peter is currently doctoral student at the section for learning and instruction research of ETH Zurich in Switzerland. He graduated from Psychology at the University of Graz in Austria. Peter is interested in conceptual knowledge development and the application of flexible mixture models to developmental research. Since 2011 he has been active in the EFPSA European Summer School and related activities.

More Posts

Facebooktwitterrss