Discussion
Wi h Discui you can
• spli you lis o ma e ials in mul iple balanced se s - based on clus e ing algo i hms;²
• ob ain au oma ic se compa isons (s a is ics);
• choose be ween a web app o a Py hon package.
Wi h he desc ibed use case, we show ha his also wo ks o ac ual/complex da a.
Con ac me i you ha e da a-se s I can un his on as u he use cases o an upcoming pape !
Use case - ou pu
Discui p oduces he ollowing ou pu :
• File wi h added se dis ibu ion (see below);
• S a is ics o all a iables (se compa ison, see igh ).
Ou pu ile:
• Discui p oduced wo se s ha di e in none o he 10 a iables con olled o .
• I ook ~ 1 minu e o do so.
Ve b Log10
eq4
AoA5Conc .6Ins . T ansi . I eg.
Pas
P e1
Pas
P e1
Fu u e
P e2
Pas
P e2
Fu u e
Se
Box 2.20 4.60 3.81 0 0 0 1 0 1 1 2
B eak 2.35 5.94 3.71 0 1 1 0 0 1 1 2
B ush 1.15 4.12 4.54 1 1 0 0 0 0 1 1
Build 2.47 4.64 3.71 0 1 1 1 0 1 1 1
Ca ch 2.32 5.20 4.11 0 1 1 0 0 0 1 2
Chee 1.35 5.75 2.77 0 0 0 0 0 1 0 2
Clean 1.84 4.22 3.81 1 1 0 1 0 0 1 1
Cook 2.15 4.88 4.32 1 0 0 0 1 0 0 2
…
Au oma ed gene a ion o balanced se s using Discui :
A ool desc ip ion and usage example
Dö e de Kok
Cen e o Language and Cogni ion, Uni e si y o G oningen, The Ne he lands
✉ d.a.de.k[email p o ec ed]
Ge his
pos e !
In oduc ion
P oblem:
• Fo expe imen s o ea men s udies balanced se s o i ems a e needed.
• Con olled o mul iple linguis ic a iables…
• Fo each pa icipan indi idually.
• Don’ educe con inuous a iables o ca ego ies (use e.g. ac ual equency
alues).
Solu ion so a :
• Make 1 long lis .
• Use ca ego ies such as high s low equency.
• Spli by hand.
P oposed solu ion:
• New ool: Dynamic i em se clus e ing UI ool (Discui ).¹
• Wo ks as Py hon package and as a web app.
• Based on k-means and k-mode clus e ing ( om he sciki -lea n package).²
Use case - me hod
S udy in o ea men e icacy:³
• Pa icipan DTR: mild aphasia, ea men o e b in lec ion.
• 107 expe imen al i ems ( illing in e bs in sen ences: pas & u u e ense).
• Needed o be spli in ea ed and un ea ed i ems.
• Balanced a iables: Log10 equency4, AoA a ings5, conc e eness a ings6,
ansi i i y, ins umen ali y, egula i y o pas ense o m, accu acy in 2
sessions be o e ea men ( o bo h pas and u u e).
• File wi h all i ems and a iables passed o Discui .
Inpu ile:
Ve b Log10
eq4
AoA5Conc .6Ins . T ansi . I eg.
Pas
P e1
Pas
P e1
Fu u e
P e2
Pas
P e2
Fu u e
Box 2.20 4.60 3.81 0 0 0 1 0 1 1
B eak 2.35 5.94 3.71 0 1 1 0 0 1 1
B ush 1.15 4.12 4.54 1 1 0 0 0 0 1
Build 2.47 4.64 3.71 0 1 1 1 0 1 1
Ca ch 2.32 5.20 4.11 0 1 1 0 0 0 1
Chee 1.35 5.75 2.77 0 0 0 0 0 1 0
Clean 1.84 4.22 3.81 1 1 0 1 0 0 1
Cook 2.15 4.88 4.32 1 0 0 0 1 0 0
…
Discui
Ei he use Discui wi h he web applica ion… … o ins all he Py hon package and un locally.
h ps://discui .s eamli .app h ps://pypi.o g/p ojec /discui
T y
Discui app
T y
Discui package
Re e ences
1 De Kok, D. (2023). Discui ( 0.2.1). Zenodo. h ps://doi.o g/10.5281/zenodo.7839874
2 Ped egosa, F., Va oquaux, G., G am o , A., Michel, V., Thi ion, B., G isel, O., Blondel, M., P e enho e , P., Weiss, R., Dubou g, V., Vande plas, J., Passos, A., Cou napeau, D., B uche , M.,
Pe o , M., & Duchesnay, E. (2011). Sciki -lea n: Machine Lea ning in Py hon. Jou nal o Machine Lea ning Resea ch, 12(85), 2825-2830.
3 Cupe us, P. (2023). Aphasia he apy so wa e: Resea ch, de elopmen , and implemen a ion [Doc o al Disse a ion]. Uni e si y o G oningen. h ps://hdl.handle.ne /11370/419a2d92-877e-
49d1-8b69-089383800088
4 an Heu en, W. J. B., Mande a, P., Keulee s, E., & B ysbae , M. (2014). Sub lex-UK: A New and Imp o ed Wo d F equency Da abase o B i ish English. Qua e ly Jou nal o Expe imen al
Psychology, 67(6), 1176–1190. h ps://doi.o g/10.1080/17470218.2013.850521
5 Kupe man, V., S ad hagen-Gonzalez, H. & B ysbae , M. (2012). Age-o -acquisi ion a ings o 30,000 English wo ds. Beha io Reseach Me hods, 44, 978–990. h ps://doi.o g/10.3758/
s13428-012-0210-4
6 B ysbae , M., Wa ine , A.B., & Kupe man, V. (2014). Conc e eness a ings o 40 housand gene ally known English wo d lemmas. Beha io Resea ch Me hods, 46, 904-911. h ps://doi.
o g/10.3758/s13428-013-0403-5
Use case - s a is ics
a: Pea son Chi-squa e es s o independence
b: K usall-Wallis Ano a
Va iable Se 1 Se 2 Tes s a is ic p
T ansi i i yaIn ansi i e
T ansi i e
26
28
26
27 X²(1) = 0.000 1.000
Ins umen ali ya
Non-ins .
Ins umen al
Missing da a
39
14
1
40
12
1
X²(2) = 0.157 .942
Pas ense o maRegula
I egula
35
19
31
22 X²(1) = 0.225 .636
P e es 1 - Pas aInco ec
Co ec
32
22
32
21 X²(1) = 0.000 1.000
P e es 1 - Fu u eaInco ec
Co ec
44
10
43
10 X²(1) = 0.000 1.000
P e es 2 - Pas aInco ec
Co ec
15
39
14
39 X²(1) = 0.000 1.000
P e es 2 - Fu u eaInco ec
Co ec
27
27
27
26 X²(1) = 0.000 1.000
F equencybMean
(SD)
1.947
(0.592)
1.998
(0.653) X²(1) = 0.058 .810
AoAbMean
(SD)
5.091
(1.132)
5.283
(1.314) X²(1) = 0.845 .358
Conc e enessbMean
(SD)
3.797
(0.591)
3.796
(0.564) X²(1) = 0.049 .824
