The last two years have seen a lot of talk about the issues of science and scientific publishing – and how the incentives prevalent in science (publish or perish, preferably with high-impact stories with lots of news coverage) are actually bad for science. Corina Logan, a zoologist and part of a group of postdocs from the University of Cambridge is eager to push for a change in the publishing culture. They argue that the current way of publishing is hindering the progress of science. A recent column by Brian Martinson in Nature summarises the problem nicely: “[The fact that researchers need publications encourages] all manner of corner-cutting, sloppiness in research, and other degradations in the quality of publications, not to mention an obvious motive for plagiarism. A quest for high-profile papers leads researchers to favour a spectacular result, even if it is specious. Authors cite themselves to boost the impact of publications, and cite colleagues to curry favour.” Continue reading
Last summer saw the publication of the most important work in psychology in decades: the Reproducibility Project (Open Science Collaboration, 2015; see here and here for context). It stirred up the community, resulting in many constructive discussions but also in verbally violent disagreement. What unites all parties, however, is the call for more transparency and openness in research.
Eric-Jan “EJ” Wagenmakers has argued for pre-registration of research (Wagenmakers et al., 2012; see also here) and direct replications (e.g., Boekel et al., 2015; Wagenmakers et al., 2015), for a clearer demarcation of exploratory and confirmatory research (de Groot, 1954/2013), and for a change in the way we analyze our data (Wagenmakers et al., 2011; Wagenmakers et al., in press). Continue reading
Registered Reports (RRs) are a new publishing format pioneered by the journal Cortex (Chambers 2013). This publication format emphasises the process of rigorous research, rather than the results, in an attempt to avoid questionable research practices such as p-hacking and HARK-ing, which ultimately reduce the reproducibility of research and contribute to publication bias in cognitive science (Chambers et al. 2014). A recent JEPS post by Dablander (2016) and JEPS’ own editorial for adopting RRs (King et al. 2016) have given a detailed explanation of the RR process. However, you may have thought that publishing a RR is reserved for only senior scientists, and is not a viable option for a postgraduate student. In fact, 5 out of 6 of the first RRs published by Cortex have had post-graduate students as authors, and publishing by RR offers postgraduates and early career researchers many unique benefits. Continue reading
Does the mention of the word “statistics” strike fear into your heart and send shivers down your spine? The results section of your thesis seeming like that dark place one should avoid at all cost? Heteroscedasticity gives you nightmares? You dread having to explain to someone what degrees of freedom are? What is the point of using ANOVA if we can do a series of t-tests? If any of these remind you of the pain of understanding statistics, or the dread of how much more lies ahead during your studies, when all you really want is someone to explain it in a humanly understandable way—look no further. Quite a few fellow students might tell you “You should go and look at Andy Field’s books. Now, at least, I understand stats”. The “Discovering statistics using …” is a gentle, student friendly introduction to statistics. Principles are introduced at a slow pace, with plenty of workable examples so that anyone with basic maths skills will be able to digest it. Now add a lens of humor and sarcasm that will have you giggling about statistics in no time! Continue reading
Do you wish to publish your work but don’t know how to get started? We asked some of our student authors, Janne Hellerup Nielsen, Dimitar Karadzhov, and Noelle Sammon, to share their experience of getting published. Continue reading
People suffering from mental disorders face great difficulties in their daily lives and deserve all possible support from their social environment. However, their social milieus are often host to stigmatizing behaviors that actually serve to increase the severity of their mental disorders: People diagnosed with a mental disorder are often believed to be dangerous and excluded from social activities. Individuals who receive treatment are seen as being “taken care of” and social support is extenuated. Concerned friends, with all their best intentions, might show apprehensiveness when it comes to approaching someone with a diagnosis, and end up doing nothing (Corrigan & Watson, 2002). These examples are not of exceptional, sporadic situations—according to the World Health Organisation, nine out of ten people with a diagnosis report suffering from stigmatisation (WHO, 2016). Continue reading
Are you tired of SPSS’s confusing menus and of the ugly tables it generates? Are you annoyed by having statistical software only at university computers? Would you like to use advanced techniques such as Bayesian statistics, but you lack the time to learn a programming language (like R or Python) because you prefer to focus on your research?
While there was no real solution to this problem for a long time, there is now good news for you! A group of researchers at the University of Amsterdam are developing JASP, a free open-source statistics package that includes both standard and more advanced techniques and puts major emphasis on providing an intuitive user interface.
The current version already supports a large array of analyses, including the ones typically used by researchers in the field of psychology (e.g. ANOVA, t-tests, multiple regression).
In addition to being open source, freely available for all platforms, and providing a considerable number of analyses, JASP also comes with several neat, distinctive features, such as real-time computation and display of all results. For example, if you decide that you want not only the mean but also the median in the table, you can tick “Median” to have the medians appear immediately in the results table. For comparison, think how this works in SPSS: First, you must navigate a forest of menus (or edit the syntax), then, you execute the new syntax. A new window appears and you get a new (ugly) table.
In JASP, you get better-looking tables in no time. Click here to see a short demonstration of this feature. But it gets even better—the tables are already in APA format and you can copy and paste them into Word. Sounds too good to be true, doesn’t it? It does, but it works!
Interview with lead developer Jonathon Love
How did development on JASP start? How did you get involved in the project?
All through my undergraduate program, we used SPSS, and it struck me just how suboptimal it was. As a software designer, I find poorly designed software somewhat distressing to use, and so SPSS was something of a thorn in my mind for four years. I was always thinking things like, “Oh, what? I have to completely re-run the analysis, because I forgot X?,” “Why can’t I just click on the output to see what options were used?,” “Why do I have to read this awful syntax?,” or “Why have they done this like this? Surely they should do this like that!”
At the same time, I was working for Andrew Heathcote, writing software for analyzing response time data. We were using the R programming language and so I was exposed to this vast trove of statistical packages that R provides. On one hand, as a programmer, I was excited to gain access to all these statistical techniques. On the other hand, as someone who wants to empower as many people as possible, I was disappointed by the difficulty of using R and by the very limited options to provide a good user interface with it.
So I saw that there was a real need for both of these things—software providing an attractive, free, and open statistics package to replace SPSS, and a platform for methodologists to publish their analyses with rich, accessible user interfaces. However, the project was far too ambitious to consider without funding, and so I couldn’t see any way to do it.
Then I met E.J. Wagenmakers, who had just received a European Research Council grant to develop an SPSS-like software package to provide Bayesian methods, and he offered me the position to develop it. I didn’t know a lot about Bayesian methods at the time, but I did see that our goals had a lot of overlap.
So I said, “Of course, we would have to implement classical statistics as well,” and E.J.’s immediate response was, “Nooooooooooo!” But he quickly saw how significant this would be. If we can liberate the underlying platform that scientists use, then scientists (including ourselves) can provide whatever analyses we like.
And so that was how the JASP project was born, and how the three goals came together:
- to provide a liberated (free and open) alternative to SPSS
- to provide Bayesian analyses in an accessible way
- to provide a universal platform for publishing analyses with accessible user interfaces
What are the biggest challenges for you as a lead developer of JASP?
Remaining focused. There are hundreds of goals, and hundreds of features that we want to implement, but we must prioritize ruthlessly. When will we implement factor analysis? When will we finish the SEM module? When will data entry, editing, and restructuring arrive? Outlier exclusion? Computing of variables? These are all such excellent, necessary features; it can be really hard to decide what should come next. Sometimes it can feel a bit overwhelming too. There’s so much to do! I have to keep reminding myself how much progress we’re making.
Maintaining a consistent user experience is a big deal too. The JASP team is really large, to give you an idea, in addition to myself there’s:
- Ravi Selker, developing the frequentist analyses
- Maarten Marsman, developing the Bayesian ANOVAs and Bayesian linear regression
- Tahira Jamil, developing the classical and Bayesian contingency tables
- Damian Dropmann, developing the file save, load functionality, and the annotation system
- Alexander Ly, developing the Bayesian correlation
- Quentin Gronau, developing the Bayesian plots and the classical linear regression
- Dora Matzke, developing the help system
- Patrick Knight, developing the SPSS importer
- Eric-Jan Wagenmakers, coming up with new Bayesian techniques and visualizations
With such a large team, developing the software and all the analyses in a consistent and coherent way can be really challenging. It’s so easy for analyses to end up a mess of features, and for every subsequent analysis we add to look nothing like the last. Of course, providing as elegant and consistent a user-experience is one of our highest priorities, so we put a lot of effort into this.
How do you imagine JASP five years from now?
JASP will provide the same, silky, sexy user experience that it does now. However, by then it will have full data entering, editing, cleaning, and restructuring facilities. It will provide all the common analyses used through undergraduate and postgraduate psychology programs. It will provide comprehensive help documentation, an abundance of examples, and a number of online courses. There will be textbooks available. It will have a growing community of methodologists publishing the analyses they are developing as additional JASP modules, and applied researchers will have access to the latest cutting-edge analyses in a way that they can understand and master. More students will like statistics than ever before.
How can JASP stay up to date with state-of-the-art statistical methods? Even when borrowing implementations written in R and the like, these always have to be implemented by you in JASP. Is there a solution to this problem?
Well, if SPSS has taught us anything, you really don’t need to stay up to date to be a successful statistical product, ha-ha! The plan is to provide tools for methodologists to write add-on modules for JASP—tools for creating user interfaces and tools to connect these user interfaces to their underlying analyses. Once an add-on module is developed, it can appear in a directory, or a sort of “App Store,” and people will be able to rate the software for different things: stability, user-friendliness, attractiveness of output, and so forth. In this way, we hope to incentivize a good user experience as much as possible.
Some people think this will never work—that methodologists will never put in all that effort to create nice, useable software (because it does take substantial effort). But I think that once methodologists grasp the importance of making their work accessible to as wide an audience as possible, it will become a priority for them. For example, consider the following scenario: Alice provides a certain analysis with a nice user interface. Bob develops an analysis that is much better than Alice’s analysis, but everyone uses Alice’s, because hers is so easy and convenient to use. Bob is upset because everyone uses Alice’s instead of his. Bob then realizes that he has to provide a nice, accessible user experience for people to use his analysis.
I hope that we can create an arms race in which methodologists will strive to provide as good a user experience as possible. If you develop a new method and nobody can use it, have you really developed a new method? Of course, this sort of add-on facility isn’t ready yet, but I don’t think it will be too far away.
You mention on your website that many more methods will be included, such as structural equation modeling (SEM) or tools for data manipulation. How can you both offer a large amount of features without cluttering the user interface in the future?
Currently, JASP uses a ribbon arrangement; we have a “File” tab for file operations, and we have a “Common” tab that provides common analyses. As we add more analyses (and as other people begin providing additional modules), these will be provided as additional tabs. The user will be able to toggle on or off which tabs they are interested in. You can see this in the current version of JASP: we have a proof-of-concept SEM module that you can toggle on or off on the options page. JASP thus provides you only with what you actually need, and the user interface can be kept as simple as you like.
Students who are considering switching to JASP might want to know whether the future of JASP development is secured or dependent on getting new grants. What can you tell us about this?
JASP is currently funded by a European Research Council (ERC) grant, and we’ve also received some support from the Centre for Open Science. Additionally, the University of Amsterdam has committed to providing us a software developer on an ongoing basis, and we’ve just run our first annual Bayesian Statistics in JASP workshop. The money we charge for these workshops is plowed straight back into JASP’s development.
We’re also developing a number of additional strategies to increase the funding that the JASP project receives. Firstly, we’re planning to provide technical support to universities and businesses that make use of JASP, for a fee. Additionally, we’re thinking of simply asking universities to contribute the cost of a single SPSS license to the JASP project. It would represent an excellent investment; it would allow us to accelerate development, achieve feature parity with SPSS sooner, and allow universities to abandon SPSS and its costs sooner. So I don’t worry about securing JASP’s future, I’m thinking about how we can expand JASP’s future.
Of course, all of this depends on people actually using JASP, and that will come down to the extent that the scientific community decides to use and get behind the JASP project. Indeed, the easiest way that people can support the JASP project is by simply using and citing it. The more users and the more citations we have, the easier it is for us to obtain funding.
Having said all that, I’m less worried about JASP’s future development than I’m worried about SPSS’s! There’s almost no evidence that any development work is being done on it at all! Perhaps we should pass the hat around for IBM.
What is the best way to get started with JASP? Are there tutorials and reproducible examples?
For classical statistics, if you’ve used SPSS, or if you have a book on statistics in SPSS, I don’t think you’ll have any difficulty using JASP. It’s designed to be familiar to users of SPSS, and our experience is that most people have no difficulty moving from SPSS to JASP. We also have a video on our website that demonstrates some basic analyses, and we’re planning to create a whole series of these.
As for the Bayesian statistics, that’s a little more challenging. Most of our effort has been going in to getting the software ready, so we don’t have as many resources for learning Bayesian statistics ready as we would like. This is something we’ll be looking at addressing in the next six to twelve months. E.J. has at least one (maybe three) books planned.
That said, there are a number of resources available now, such as:
- Alexander Etz’s blog
- E.J.’s website provides a number of papers on Bayesian statistics (his website also serves as a reminder of what the internet looked like in the ’80s)
- Zoltan Dienes book is a great for Bayesian statistics as well
However, the best way to learn Bayesian statistics is to come to one of our Bayesian Statistics with JASP workshops. We’ve run two so far and they’ve been very well received. Some people have been reluctant to attend—because JASP is so easy to use, they didn’t see the point of coming and learning it. Of course, that’s the whole point! JASP is so easy to use, you don’t need to learn the software, and you can completely concentrate on learning the Bayesian concepts. So keep an eye out on the JASP website for the next workshop. Bayes is only going to get more important in the future. Don’t be left behind!
We all know these crucial moments while analysing our hard-earned data – the moment of truth – is there a star above the small p? Maybe even two? Can you write a nice and simple paper or do you have to bend your back to explain why people do not, surprisingly, behave the way you thought they would? It all depends on those little stars, below or above .05, significant or not, black or white. Continue reading
|Prof. Dermot Barnes-Holmes was a Foundation Professor at the Department of Psychology at National University of Ireland, Maynooth. He is known for his research in human language and cognition through the development of the Relational Frame Theory (RFT) with Steven C. Hayes, and its applications in various psychological settings.|
What I enjoy most about my job as a researcher … Supervising research students who are passionate about and genuinely interested in their research. Sharing what is often a voyage of intellectual discovery for both the student and me is still, after all these years, by far the most stimulating and enjoyable feature of what I do as an academic. Continue reading
What I enjoy most about my job as a researcher … Without a doubt what I enjoy most about my job as a researcher is the possibility to create and devise new experiments, to test new exciting ideas, to challenge pre-existing models with new hypotheses that I gather from discussions with people, but especially from a lot of reading and listening to insightful talks. Continue reading