Discriminant function analysis (DFA) and Permuted Discriminant Function Analysis (pDFA)

[TarandeepKang]

I propose the inclusion of both the Discriminant function analysis and the Permuted discriminant function analysis in JASP. At present, as far as I can tell, neither of these analysis functions are available.

The DFA is available in all major statistical packages And is used to classify unknown individuals into groups, and the probability of those individuals belonging in those groups. It is widely used across, biology, acoustics and animal behaviour. The tests are available in linear and quadratic forms using the lda and qda functions from the MASS package. Another major selling point is that the DFA has long been recommended (Enders, 2003) as the follow-up test of choice for MANOVA.

The pDFA has a number of advantages over the regular discriminant function analysis. This analysis was designed and proposed by Mundry and Sommer (2007). In the abstract of the paper The authors summarise the advantages of the test thus:

“ The discriminant function analysis (DFA) is a multivariate method that is frequently used in bioacoustic research to examine, for instance, whether calls from different species, contexts, or social groups can be distinguished by their acoustic properties. Most published studies include more than one call per subject into such an analysis. This, in fact, leads to a two-factorial data set that includes the factor ‘subject’ in addition to the factor of interest (e.g. species, context, or social group). The regular version of the DFA, however, does not allow for the analysis of such data sets without violating the assumption of independence. In this paper, we show that analysing factorial data sets using a conventional DFA is a case of pseudoreplication and tends to produce (sometimes grossly) incorrect results. In such a case the discriminability of species, contexts or groups etc. can be drastically overestimated. Furthermore, we provide a permutation-based procedure that copes with such data sets.“

Unfortunately, the underlying R code has not been made public. However, the author does provide the code if he is contacted via email.

I hope I have been clear, and provided The necessary information. Thanks for your efforts.

Enders, C. K. (2003). Performing Multivariate Group Comparisons Following a Statistically Significant MANOVA. Measurement and Evaluation in Counseling and Development, 36(1), 40–56. https://doi.org/10.1080/07481756.2003.12069079

Mundry, R., & Sommer, C. (2007). Discriminant function analysis with nonindependent data: consequences and an alternative. Animal Behaviour.

[EJWagenmakers]

Excellent suggestion, we're going to do this EJ

[TarandeepKang]

The pDFA, in particular can be a big learning curve when it comes to using the R code. It will be great to have a readily available alternative to the DFA, given it is not suitable for use, in many situations. I hope that Roger will be amenable. Thanks again!

[TarandeepKang]

Hi EJ,

I’d like to make clear that I was not only referring to the discriminant function analysis, but also to the Permuted discriminant function analysis. I emphasise that these are two different statistical tests. I have changed the title to clarify this.

Best,

Tarandeep

Sent from my iPhone

On 16 Aug 2019, at 15:05, EJWagenmakers notifications@github.com wrote:

Excellent suggestion, we're going to do this EJ

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[tinamaltezou]

Hey there, I'm really interested in the feature of discriminant function analysis too. Has any work started on it yet or still in the backlog?

[EJWagenmakers]

Backlog, but we'll get there

[EJWagenmakers]

Thanks for the reminder though!

[rsbalkin]

still wanting to see discriminant analysis so post hoc analysis for MANOVA (as opposed to univariate ANOVAs) can be performed

[rsbalkin]

I posted about the inclusion of DFA as a post hoc to MANOVA back in November 2021. I hope to see some traction on this.

[rsbalkin]

Backlog, but we'll get there

It’s been three years since the initial post! Is this really going to be added?

[tomtomme]

I would bet on it, yes. But it may need some time still. There are many many many feature requests in the queue. And the dev team is very busy implementing them one by one. Ther is only so much you can do with a hand full of fulltime devs and another handful of parttime devs. See all the features implemented during the last years here: https://jasp-stats.org/release-notes/

sorry that I can not bring more positive messages. Maybe @JohnnyDoorn can say more on the timeline.

[rsbalkin]

I understand. I teach research methods and statistics in the School of Education at The University of Mississippi to doctoral students. We use JASP. However, when I teach MANOVA, univariate post hoc is an inadequate post hoc and has been considered inadequate for quite some time (Enders, 2003). JASP has been a game-changer, and I love using it and teaching with it. But this one addition would allow me to be completely done with SPSS. Thank you for all you and your team do.RickRichard S. Balkin, Ph.D., LPC, NCCEditor-in-Chief, International Journal for the Advancement of CounsellingFellow, American Counseling AssociationProfessor & Chair, Department of Leadership and Counselor EducationSchool of EducationUniversity of @. from my iPadOn Dec 24, 2023, at 6:33 AM, Thomas Langkamp @.> wrote: I would bet on it, yes. But it may need some time still. There are many many many feature requests in the queue. And the dev team is very busy implementing them one by one. Ther is only so much you can do with a hand full of fulltime devs and another handful of parttime devs. See all the features implemented during the last years here: https://jasp-stats.org/release-notes/ sorry that I can not bring more positive messages. Maybe @JohnnyDoorn can say more on the timeline.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>