Me a-analysis/sys ema ic e iew
iScience
Measu ing en i onmen al (in)jus ices: Insigh s om
a sys ema ic li e a u e e iew on me hodological
app oaches
G aphical abs ac
Highligh s
• En i onmen al (in)jus ice esea ch is expanding, ye
me hodologically agmen ed
• Re iew o 421 s udies shows dominance o quan i a i e, GIS,
and seconda y-da a me hods
• Pa icipa o y and quali a i e, p ocedu al, and ecogni ion
jus ice a e unde used
• Resea ch skews owa d No h Ame ica; u u e wo k should
cen e on ma ginalized oices
Au ho s
Jacqueline Loos, Cha lo e Goh ,
Noelia Za a-Cal o,
Gonzalo Co e
´
s-Capano,
Anna Lena Tonninge ,
Hen ik on Weh den
Co espondence
jacqueline.loo[email p o ec ed]
In b ie
Ea h sciences; En i onmen al science;
Social sciences
Loos e al., 2025, iScience 28, 113889
Decembe 19, 2025 © 2025 The Au ho (s). Published by Else ie Inc.
h ps://doi.o g/10.1016/j.isci.2025.113889
ll
iScience
Me a-analysis/sys ema ic e iew
Measu ing en i onmen al (in)jus ices:
Insigh s om a sys ema ic li e a u e
e iew on me hodological app oaches
Jacqueline Loos,
1,6,
*
Cha lo e Goh ,
2
Noelia Za a-Cal o,
3
Gonzalo Co e
´
s-Capano,
4,5
Anna Lena Tonninge ,
1
and Hen ik on Weh den
2
1
Uni e si y o Vienna, Facul y o Li e Sciences, Depa men o Bo any and Biodi e si y Resea ch, Rennweg 14, 1030 Vienna, Aus ia
2
Leuphana Uni e si y, Cen e o Me hods and Facul y o Sus ainabili y, 21335 Lu¨ nebu g, Ge many
3
Basque Cen e o Clima e Change (BC3), 48940 Leioa, Spain
4
Depa men o Ag icul u al Economics and Ru al De elopmen , Uni e si y o Go
¨
ingen, Go
¨
ingen, Ge many
5
Facul y o O ganic Ag icul u al Sciences, Uni e si y o Kassel, Wi zenhausen, Ge many
6
Lead con ac
*Co espondence: [email p o ec ed]
h ps://doi.o g/10.1016/j.isci.2025.113889
SUMMARY
En i onmen al (in)jus ice esea ch uses a ious concep ual amewo ks and me hodological app oaches,
leading o agmen a ion ac oss con ex s and disciplines. Ou sys ema ic e iew p o ides a me hodological
o e iew o how en i onmen al (in)jus ice has been s udied in 421 English-language scien i ic a icles. Mos
s udies app oach en i onmen al (in)jus ice om a quan i a i e and in e disciplina y pe spec i e, p ima ily
using pu posi e sampling, seconda y da a, and GIS/ emo e sensing ools wi h an emphasis on dis ibu i e
jus ice. Al hough he e is a no able di e si ica ion o e ime in da a collec ion and analysis, he e is a s ong
geog aphic bias wi h sho - e m, locally ocused, and limi ed ac o in ol emen , hough ac o di e si y is
g owing o e ime. We iden i ied eigh hema ic clus e s wi h dis inc me hodological pa e ns: heal h, pollu-
ion, go e nance, clima e change, collabo a ion, access, and g een space. The lack o b oadly adop ed
me hodological app oaches o e alua ing en i onmen al (in)jus ices la gely s ems om he con ex -speci ic,
mul i-scala na u e o cases and he philosophical and no ma i e di e si y embedded in he EJ concep i sel .
INTRODUCTION
En i onmen al jus ice (EJ) e e s o emo ing he ba ie s ha
cause dispa i ies and inequali ies in a gi en eali y.
1,2
Al hough i
s a ed as a ci il igh s mo emen in esponse o disp opo iona e
pollu ion and heal h issues bea ed by ma ginalized people, such
as black communi ies,
3,4
o e he pas decades, EJ has signi i-
can ly expanded i s ocus o a global, mul i ace ed ma e ha ad-
d esses en i onmen al go e nance, pollu ion, land use, heal h,
biodi e si y, sus ainabili y ansi ions, and clima e change.
5–8
EJ
has been inc easingly backed up by scien i ic da a and e idence
since i s o igin, o en h ough communi y-led science,
9
he eby
un a eling sys emic disad an ages o ma ginalized people. These
disad an ages a e exp essed ei he in disp opo iona e exposu e
o en i onmen al haza ds o heal h isks, o in educed access and
a ailabili y o bene i s ha a e ela ed o en i onmen al asse s.
Cu en ly, EJ ep esen s bo h a b oad ange o g ass oo social
mo emen s and an in e disciplina y ield o academic esea ch,
engaging wi h di e se academic adi ions and disciplines (e.g.,
geog aphy, sociology, en i onmen al sciences) and ac i is com-
muni ies a ound he wo ld.
10–12
The concep ual expansion o EJ is well documen ed.
13,14
Exis -
ing e iews syn hesize impo an ad ances and deba es in ela-
ion o i s concep ualiza ion o ame and s udy inequali y issues
in clima e ulne abili y, g een spaces, pollu ion, heal h, and go e -
nance, as well as o in es iga e how o suppo all membe s o so-
cie y in p opo ion o wha hey ini ially ha e o wha hey need o
ackle inequali y and each jus ice; a concep known as eq-
ui y.
15,16
EJ is a mul idimensional concep , encompassing a ious
no ions and amewo ks. I is commonly unde s ood o include
dis ibu i e, p ocedu al, ecogni ional, and con ex ual ele-
men s,
17–19
bu has also been expanded o inco po a e addi ional
dimensions such as es o a i e o mul ispecies jus ice.
5,20–22
In addi ion, EJ has become an impo an p inciple in many pub-
lic p og ams and policies in di e en geog aphies and a di e en
scales, om local o na ional and global. Thus, he e is a g owing
need o e iews and oolki s ha cap u e and add ess how EJ
can be s udied empi ically in di e se geog aphies and con ex s.
Exis ing e iews ocus on speci ic analy ical ools (e.g., spa ial
analysis, pa icipa o y mapping) o domains (e.g., heal h, pollu-
ion),
23–25
bu he e is o en a lack o clea unde s anding and guid-
ance on how o empi ically measu e EJ. Ta an and Co dell
(1999)
26
al eady s a ed ea ly ha da a collec ion and analysis
echniques in en i onmen al jus ice esea ch may lead o con lic -
ing esul s. Howe e , o ou knowledge, a comp ehensi e me h-
odological o e iew o he ield is ye o come.
iScience 28, 113889, Decembe 19, 2025 © 2025 The Au ho (s). Published by Else ie Inc. 1
This is an open access a icle unde he CC BY-NC license (h p://c ea i ecommons.o g/licenses/by-nc/4.0/).
ll
OPEN ACCESS
Empi ical esea ch on EJ is agmen ed ac oss disciplines
such as geog aphy, sociology, poli ical ecology, and planning,
wi h di e en me hodological adi ions, employing quan i a i e,
quali a i e, and mixed me hods.
25,27
Each b ings dis inc he-
o ies and epis emological assump ions, shaping how injus ices
a e iden i ied, measu ed, and add essed. This agmen a ion
p oduces challenges such as how o iden i y sui able me hods
and measu emen s aligned wi h he jus ice dimensions s udied,
how me hods can be bo h con ex -sensi i e and compa able
ac oss scales and o wha ex en hey make isible hidden o
s uc u al o ms o ha m, and how hey e lec unde lying alues,
no ms, and powe dynamics.
28,29
The way EJ issues a e con-
s uc ed is deeply connec ed o he me hods used o e eal mul-
iple dimensions o ha m ac oss ime and space.
17,30
I also de-
pends on he capaci y o c i ically e lec on he assump ions,
limi a ions, and po en ial o hese me hods o make conscious
decisions when using hem and wo king wi h da a.
Likewise wi h he mul i ude o in e p e a ions and concep u-
aliza ions o wha can be conside ed ‘‘jus ,’’ he e is no me h-
odological uni y in how o assess, measu e, and communica e
en i onmen al (in)jus ice issues.
31
To ou knowledge, he e is
also no sys ema ic and compa a i e syn hesis o empi ical
me hods used o s udy and assess EJ ac oss mul iple dimen-
Figu e 1. PRISMA 2020 low diag am o he
li e a u e sc eening and selec ion p ocess
Adap ed om Page MJ e al. (2021), BMJ, 372:n71,
unde he CC BY 4.0 license (h ps://c ea i ecommons.
o g/licenses/by/4.0/).
sions and disciplines, no how hese
me hods ha e e ol ed alongside concep-
ual de elopmen s. Such an o e iew
could p o ide a clea e mapping o me h-
odological di e si y, e ealing dominan
app oaches as well as unde u ilized o
eme ging ones. This syn hesis would help
iden i y implici biases, neglec ed aspec s,
o example, endencies owa d ce ain
geog aphic egions o me hodological ap-
p oaches, he eby highligh ing oppo u-
ni ies o b oaden me hodological plu alism.
Mo eo e , by acking how me hods ha e
de eloped, his wo k could guide he in e-
g a ion o quan i a i e, quali a i e, and
mixed app oaches, os e ing mo e holis ic
and con ex -sensi i e esea ch. In doing
so, i would suppo in e disciplina y collab-
o a ion, en ich u u e esea ch agendas by
iden i ying knowledge gaps and inno a ion
spaces, and enhance he policy ele-
ance and communica ion o en i onmen al
(in)jus ice schola ship. Ul ima ely, his con-
ibu es o a mo e inclusi e and nuanced
unde s anding o en i onmen al (in)jus ice,
be e equipped o add ess plu ali y in en i-
onmen al (in)jus ice esea ch.
32
To his end, we conduc ed a sys ema ic li e a u e e iew o En-
glish academic publica ions (Figu e 1) o cha ac e ize he main
empi ical app oaches used o s udy en i onmen al (in)jus ice
ac oss disciplines, geog aphies, jus ice dimensions, and scales.
In pa icula we 1- iden i ied which empi ical me hods ha e been
used o s udy en i onmen al jus ice in published case s udies;
2- cha ac e ized which jus ice dimensions (e.g., con ex ual,
dis ibu i e, p ocedu al, ecogni ional, es o a i e, in e gene a-
ional, mul ispecies) hey add ess; 3- explo ed how hese
me hods e ol ed o e ime and ac oss geog aphies and 4-
which ac o s we e conside ed and how hey we e ela ed o
which sampling, da a collec ion and analysis me hods. O e all,
by p o iding an o e iew o exis ing me hods, hei uses and lim-
i a ions, and how hey ela e o di e en jus ice claims, we aimed
o suppo bo h academic esea ch and policy applica ions.
RESULTS
Desc ip i e o e iew o me hods used in en i onmen al
jus ice
A log-linea eg ession model e ealed a signi ican inc ease in
he numbe o published s udies o e ime, wi h an es ima ed
annual g ow h a e o app oxima ely 13% pe yea (β = 0.123,
2 iScience 28, 113889, Decembe 19, 2025
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
p < 0.001, R
2
= 0.81). A icles we e published in 190 di e en
jou nals, wi h he In e na ional Jou nal o En i onmen al
Resea ch and Public Heal h (n = 24), En i onmen al Jus ice
(n = 23), Sus ainabili y (n = 13), U ban Planning (n = 12) and Ge-
og aphy (n = 10) being he i e mos equen ed ones. Mos o he
s udies (n = 248) we e quan i a i e in es iga ions, ollowed by
mixed app oaches (n = 101) and quali a i e s udies (n = 72).
The p opo ion o quan i a i e and mixed-me hods a icles
dec eased signi ican ly o e ime (p < 0.05), while he sha e o
quali a i e s udies emained s able. Ou se o a icles en ails
210 in e disciplina y and 155 disciplina y s udies, wi h mul i-
and ansdisciplina y s udies only ha ing ound hei way in o
en i onmen al jus ice esea ch in he yea 2005. Since hen,
howe e , he p opo ional sha e o ansdisciplina y s udies
has signi ican ly declined (es ima e (− 0.132, p < 0.001).
Sampling s a egy
Almos 50% o all a icles used pu posi e sampling (n = 213),
ollowed by 161 s udies employing ull inclusion/census sam-
pling. While pu posi e sampling was associa ed wi h quali a-
i e and mixed in e - & ansdisciplina y s udies, ull inclusion/
census was ypically ela ed o quan i a i e, disciplina y
s udies. S a i ied sampling (n = 42) and con enience sampling
(n = 40) we e less common, wi h andom (n = 16), mul i-s age
(n = 15), and snowball sampling (n = 9) used e en less
equen ly. Full inclusion/census was mos commonly com-
bined wi h pu posi e sampling (n = 34), and con enience sam-
pling was also equen ly pai ed wi h pu posi e sampling
(n = 11). O e all, pu posi e sampling was he me hod mos
o en combined wi h o he s (n = 76), ollowed by ull inclusion/
census (n = 44). A e accoun ing o he inc easing annual num-
be o a icles, he e was no signi ican end in he use o single
e sus mul iple sampling s a egies o e ime. Howe e , snow-
ball sampling showed he s onges signi ican decline o e
ime (es ima e = − 0.233, p = 0.034), accompanied by no able
dec eases in basic andom (es ima e = − 0.104, p = 0.007)
and mul i-s age sampling (es ima e = − 0.103, p = 0.033), while
con enience sampling emained s able.
Da a collec ion echniques
A o al o 138 a icles applied GIS and emo e sensing o da a
collec ion, and 133 u ilized seconda y da a sou ces, mos
equen ly associa ed wi h a disciplina y, quan i a i e con ex .
Su eys, censuses, o ques ionnai es we e used in 119 s udies,
while 68 a icles conduc ed in e iews. These we e mainly em-
ployed in quali a i e s udies. Pa icipa o y me hods we e applied
o 41 ansdisciplina y, quali a i e, and mixed s udies. The e was
a signi ican inc ease in he di e si y o da a collec ion me hods
o e ime (β = 0.47, SE = 0.04, = 10.91, p < 0.001), wi h he
model explaining 82% o he a iance (R
2
= 0.82). Mos a icles
(58%) elied on a single da a collec ion me hod, p edominan ly
seconda y da a (n = 69), GIS/ emo e sensing (n = 60), o su -
eys/census/ques ionnai es (n = 57). O e ime, he ela i e
use o single me hods sligh ly dec eased while he use o mul i-
ple da a collec ion echniques inc eased. Among he me hods
analyzed, only su eys/census/ques ionnai es showed a signi i-
can decline (es ima e = − 0.056, p = 0.003). The mos equen
combina ions we e GIS/ emo e sensing wi h seconda y da a
(n = 34) and GIS/ emo e sensing wi h su eys/census/ques ion-
nai es (n = 26).
Da a analysis me hods
A o al o 249 a icles applied only one da a analysis me hod a a
ime, ou o which 76 disciplina y a icles solely applied uni a i-
a e s a is ics, 69 applied spa ial analysis, 23 a icles used quali-
a i e coding, and 21 used desc ip i e s a is ics, ou o which a
la ge sha e we e o an in e disciplina y na u e. 146 a icles
used se e al da a analysis me hods, wi h mos o hem (n = 62)
combining spa ial analysis wi h uni a ia e s a is ics, while 26
combined spa ial analysis wi h desc ip i e s a is ics. Ano he
ela i ely la ge ac ion o pai ed analyses me hods we e com-
bined desc ip i e and uni a ia e s a is ics (n = 21), as well as
desc ip i e and mul i a ia e s a is ics (n = 15). We de ec ed a
small bu s a is ically signi ican inc easing end in he numbe
o di e en me hods applied o e ime (Kendall’s au = 0.11,
p = 0.006), indica ing a g adual ise in he use o mul iple da a
analysis me hods ac oss a icles. No malized by he inc easing
numbe o publica ions o e ime, gene alized linea models e-
ealed a signi ican decline in he ela i e use o machine lea ning
and modeling me hods (p < 0.005). Uni a ia e s a is ical
me hods also showed a dec easing end, hough his was
ma ginally non-signi ican (p = 0.0076). No o he me hods
demons a ed signi ican changes in ela i e usage o e ime.
Linkage be ween sampling, da a collec ion, and analysis
me hods
Pea son’s chi-squa ed es s e ealed highly signi ican associa-
ions be ween sampling s a egies and da a collec ion me hods
(χ
2
= 540.49, d = 220, p < 0.001), sampling s a egies and da a
analysis me hods (χ
2
= 198.87, d = 66, p < 0.001), as well as be-
ween da a collec ion and da a analysis me hods (χ
2
= 386.26,
d = 120, p < 0.001), indica ing s ong in e dependencies among
hese me hodological choices. The esul s e eal clea in e de-
pendencies among sampling s a egies, da a collec ion
me hods, and analy ical app oaches. Pu posi e sampling was
closely associa ed wi h quali a i e echniques such as in e iews
and quali a i e o mixed me hods analysis. Full inclusion o
census-based sampling o en co-occu ed wi h seconda y
da a and GIS/ emo e sensing collec ion, linking u he o spa ial
and s a is ical analyses. Basic andom sampling was ied o
s uc u ed ins umen s such as su eys and was equen ly
analyzed using machine lea ning o s a is ical echniques.
Ac oss da a collec ion and analysis, quali a i e me hods such
as in e iews, ocus g oups, pa icipa o y echniques, and
e hnog aphic app oaches we e s ongly linked o quali a i e
and mixed me hods analysis. In con as , seconda y da a and
GIS/ emo e sensing we e associa ed mo e wi h s a is ical and
spa ial analyses (Figu e 2).
Jus ice dimensions add essed ac oss s udies
O he coded a icles, 68% (n = 287) explici ly engaged wi h
dis ibu i e jus ice, 35 a icles wi h p ocedu al jus ice, and 19
wi h ecogni ion jus ice; 91 a icles did no clea ly speci y a jus-
ice dimension. O he dimensions add essed included social
(n = 13), epis emic (n = 6), and spa ial jus ice (n = 6). A o al o
32 a icles engaged wi h mo e han one jus ice dimension.
iScience 28, 113889, Decembe 19, 2025 3
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
No ably, all a icles published be o e 2005 engaged only in a sin-
gle jus ice dimension. In 2005, one- hi d o he s udies inco po-
a ed mo e han one jus ice dimension. Since hen, he numbe
o mul idimensional jus ice s udies has inc eased signi ican ly,
wi h an a e age annual ise o 0.2 a icles, comp ising 1–12%
o annual publica ions.
Jus ice dimensions a e highly signi ican ly associa ed wi h he
me hodological app oaches employed in he s udies, including
sampling s a egies (χ
2
= 254.26, d = 96, p < 0.001), da a collec-
ion me hods (χ
2
= 703.38, d = 176, p < 0.001), and da a analysis
echniques (χ
2
= 437.83, d = 88, p < 0.001), indica ing ha he
choice o me hods a ies sys ema ically depending on he jus-
ice ocus o he esea ch. A icles ha conside ed dis ibu i e
jus ice mos ly used pu posi e sampling (n = 141) o ull inclu-
sion/census (n = 119) as a sampling s a egy. In a icles looking
in o p ocedu al and ecogni ion jus ice, pu posi e sampling
domina ed (n = 23; n = 15, espec i ely). Da a collec ion in
dis ibu i e jus ice s udies was mos ly ela ed o GIS/ emo e
sensing (n = 109) o seconda y da a collec ion (n = 100), whe eas
s udies wi h a p ocedu al and ecogni ion ocus used in e iews
(n = 17; n = 14, espec i ely). Da a analysis in dis ibu i e jus ice
s udies p edominan ly employed spa ial analysis (n = 127) and
uni a ia e s a is ics (n = 122), bo h showing s ong posi i e asso-
cia ions (s d. esiduals >5). Desc ip i e s a is ics (n = 58) and
mul i a ia e s a is ics (n = 33) we e also commonly used in hese
s udies. In con as , p ocedu al and ecogni ion jus ice s udies
p ima ily u ilized quali a i e coding (p ocedu al n = 8, ecogni ion
n = 6) and con en analysis (p ocedu al n = 6, ecogni ion n = 6),
wi h signi ican posi i e associa ions indica ed by esiduals
abo e 5 and 6, espec i ely. Recogni ion s udies u he ea u ed
con ex ual analysis and machine lea ning/modeling me hods,
sugges ing a b oade me hodological di e si y wi hin his ca e-
go y. Pa icipa o y me hods o analyses we e used in 25 a icles,
ou o which 14 engaged wi h dis ibu i e jus ice, 6 on p oce-
du al, 2 on epis emic, and 1 on clima e, ecogni ion, and/o so-
cial jus ice.
The concep ualiza ion o EJ has e ol ed om ocus on dispa -
i ies o a deepe engagemen owa d assessing equi y conce ns
and b oadly he oo causes o (in)jus ice. The e was a signi ican
inc ease o e ime in he use o equali y- ela ed e minology
(p = 0.03), while o he e minologies, such as equi y, ai ness,
dispa i y, and une enness, showed no signi ican change. The
use o jus ice- ela ed e ms showed a posi i e bu ma ginally
non-signi ican upwa d end (p = 0.096).
Geog aphic and spa io empo al pa e ns ac oss s udies
The majo i y o he n = 421 e iewed a icles explo ed en i on-
men al (in)jus ice cases in No h Ame ica (n = 245, Figu e 3A),
emb acing he USA (n = 227) and Canada (n = 18). S udies
om Eu ope (n = 65) en ailed cases om Spain (n = 10), Ge many
(n = 9), F ance (n = 8), and he UK (n = 8). Mos s udies om Asia
(n = 53) came om China (n = 36). Rema kably ew s udies ep e-
sen Oceania (n = 3), and A ica (n = 14), and 17 s udies employed
cases on a ious con inen s. Al hough ela i ely unde ep e-
sen ed, li e a u e om Sou h Ame ica and A ica o en ocuses
on issues a ound go e nance, ecogni ion, pa icipa ion, li eli-
hood, and conse a ion. A ican s udies had he highes sha e
o in e disciplina y esea ch app oaches (71%), whe eas s udies
Figu e 2. Sankey diag am showing 927 connec ions be ween da a
collec ion and analysis me hods ac oss 421 a icles, e lec ing
mul iple me hods pe a icle
4 iScience 28, 113889, Decembe 19, 2025
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
om Sou h Ame ica we e mos ly disciplina y (61%). In he case
o Sou h Ame ica, EJ esea ch include ou ism in p o ec ed a ea
managemen ,
33
ex ac i ism,
34
ood insecu i y in indigenous
communi ies
35
and human igh s iola ions.
36
Me hods ange
om geospa ial ools and su eys o e hnog aphic and pa icipa-
o y me hods. In he case o A ica, he ocus is se on exclusion
om en i onmen al decision-making o poo and ma ginalized
people, who emain excluded om en i onmen al decision-mak-
ing,
37
PAs impac on en i onmen al jus ice and human well-be-
ing
38
and women’s ulne abili y o clima e change.
39
Rega ding
me hods, A ica-based s udies p esen in ou da abase ocus
mo e on s uc u ed da a collec ion and less on imme si e qual-
i a i e ieldwo k.
Mos s udies in es iga ed en i onmen al jus ice issues on a
local (n = 184), egional (n = 170), o na ional (n = 50) scale
(Figu e 3B). O e ime, he p opo ion o s udies conduc ed a
local and egional scales inc eased sligh ly, while he ela i e
sha e o global-scale s udies declined signi ican ly (p < 0.001).
Rega ding empo al co e age, he majo i y o s udies (n = 177)
ocused on pe iods o one yea o less. A o al o 91 s udies
examined du a ions o 1–5 yea s, 95 s udies ocused on 5–25
yea s, and 34 s udies add essed pe iods exceeding 25 yea s.
An o dinal logis ic eg ession showed ha o e ime, s udies
ended o ha e sho e du a ions (es ima e = − 0.001,
p < 0.001). The empo al scale o analysis was signi ican ly asso-
cia ed wi h he spa ial scale (χ
2
= 37.90, d = 16, p = 0.0016).
Sho - e m s udies (≤1 yea ) p ima ily ocused on local scales
(∼50%), whe eas long- e m s udies (>25 yea s) we e mo e
e enly dis ibu ed ac oss local (41%) and b oade egional o na-
ional scales (44%).
Ac o engagemen and epis emic p io i iza ion
Mos s udies (n = 282) engaged wi h a single ac o , while 139
s udies engaged wi h se e al ac o s. Mos equen ly, commu-
ni y ac o s (n = 315) and ulne able communi ies (n = 178)
we e subjec o he s udy, wi h ma kedly less ocus on ins i u-
ional ac o s (n = 37) and go e nance ac o s (n = 28). Al hough
he e was a end owa d mo e ac o g oups being in ol ed
pe s udy o e he yea s, his end was non-signi ican . While
communi y ac o s and ulne able people ha e been in ol ed
h oughou he en i e in es iga ed pe iod (1997–2025), o he ac-
o s we e only conside ed la e , s a ing wi h ins i u ional ac o s
in 2002, expe s and go e nance ac o s in 2006, and policy
shape s and ac o s om he p i a e sec o in 2010. While he
numbe o ac o s conside ed in s udies inc eased o e all, he
p opo ional sha e o each indi idual ac o g oup declined
excep o ex e nal ac o s, and his was signi ican o go e -
nance ac o s (es ima e − 0.016, p < 0.005), ins i u ional ac o s
(es ima e − 0.008, p < 0.01), and policy shape s (es ima e
− 0.011, p < 0.01). S udies including communi y ac o s mos ly
engaged a he local (n = 147) and a he egional (n = 122) scale,
which is a pa e n ha was also isible o ulne able communi y
g oups (wi h n = 76 s udies a he local and n = 71 a he egional
scale). In con as , ins i u ional ac o s we e associa ed wi h
s udies co e ing he local o global scale. Ac o inclusion
showed o be scale-sensi i e in he scien i ic li e a u e (χ
2
=
60.99, d = 32, p < 0.005). The ac o s in es iga ed in he s udies
we e as well highly signi ican ly associa ed o he choices o
sampling (χ
2
= 381.86, d = 104, p < 0.005), da a collec ion
(χ
2
= 431.76, d = 144, p < 0.005) and analysis me hods (χ
2
=
556.86, d = 72, p < 0.005), sugges ing epis emic p io i iza ion
in hese s udies: While expe s we e ypically in ol ed h ough
pu posi e sampling, communi y ac o s mos equen ly we e
in es iga ed by s a i ied sampling, and ulne able ac o s
h ough ull inclusion/census and go e nance ac o s h ough
snowball sampling. In e ms o da a collec ion, in o ma ion
om ex e nal in luence s was associa ed wi h quali a i e
coding, whe eas expe s we e in es iga ed mos ly h ough
case s udies and in e iews, ulne able communi y membe s
mos ly h ough su eys, census, and ques ionnai es, and a ely
h ough pa icipa o y app oaches. Ins ead, pa icipa o y da a
collec ion me hods mos ly occu ed in ela ion o go e nance
ac o s. Da a analysis in s udies ha included ins i u ional ac o s
was p ima ily h ough con en analysis, while s udies consid-
e ing expe s, go e nance ac o s, and/o policy shape s em-
ployed con ex ual analysis. Ac o s om he p i a e sec o we e
included in s udies using quali a i e coding. Ac o g oups
di e ed signi ican ly ac oss he conside ed jus ice dimensions
(χ
2
= 162.43, d = 64, p < 0.001). A icles including communi y
and ulne able ac o s mainly add essed dis ibu i e jus ice,
Figu e 3. De ended Co espondence Analysis O dina ion showing indica o wo d clus e s
(A) Sha e o a icles pe con inen , based on he 421 in es iga ed a icles.
(B) Coun s o published a icles so ed by hei spa ial ocus, anging om local o global in es iga ions.
iScience 28, 113889, Decembe 19, 2025 5
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
while hose ea u ing go e nance and ins i u ional ac o s
engaged mo e e enly wi h p ocedu al and ecogni ion
dimensions.
Clus e s/seman ic map o en i onmen al jus ice
clus e s and claims
Ou clus e analysis (Figu e S1) and he mul i a ia e analysis o
he 409 a icles used in he wo d analysis sugges ed eigh
di e en esea ch clus e s p esen in he en i onmen al jus ice
li e a u e ha we e iewed (Figu e 4). Gi en hei mos abundan
e ms, we named hese esea ch clus e s go e nance, pollu ion,
clima e change, heal h, was e, access, collabo a ion and g een
space (Table 1). The clus e comp ising he la ges numbe o a -
icles in his e iew is he go e nance clus e (n = 73) which
shows a dis inc wo d usage p o ile compa ed o he closely
o e lapping clus e s on pollu ion (n = 72), was e (n = 52), and
heal h (n = 53); he o e lapping clus e s o clima e change
(n = 62), access (n = 43), and g een space (n = 52); and he sepa-
a e clus e on collabo a ion (n = 73) (see Figu e 4). These clus-
e s a e dis ibu ed along he wo DCA axes in ways ha e lec
b oade hema ic dis inc ions. The clus e s on pollu ion, was e,
and heal h cen e on en i onmen al ha ms and he disp opo -
iona e exposu e o socially disad an aged people, o en de ined
by ace, e hnici y, o income. These a icles p ima ily aim o iden-
i y, desc ibe, and measu e en i onmen al injus ices, empha-
sizing diagnosis and p oblem cha ac e iza ion. In con as , he
go e nance and collabo a ion clus e s a e mo e o ien ed owa d
ins i u ional esponses, pa icipa o y s a egies, and solu ions
o add essing injus ices. Based on his con as , he i s DCA
axis can be in e p e ed as cap u ing a g adien om diagnosing
ha ms o de eloping s a egies o imp o emen . A second he-
ma ic dis inc ion eme ges be ween clus e s ocusing on en i on-
men al ha ms (pollu ion, was e) and hose add essing he dis i-
bu ion o en i onmen al goods and oppo uni ies (access, g een
space). While he o me highligh s disp opo iona e bu dens
and ulne abili ies, he la e emphasize jus access o en i on-
Figu e 4. Eigh esea ch clus e s de i ed
om 409 empi ical a icles on en i on-
men al jus ice (12 excluded due o o e -
leng h)
Colo ed nouns indica e clus e membe ship.
‘‘Accessibili y’’ is p ominen bo h in he ligh g een
and iole clus e s. P oximi y be ween clus e s
e lec s hema ic simila i y.
men al bene i s and adap a ion s a e-
gies. Acco dingly, he second DCA axis
may be in e p e ed as ep esen ing a
con inuum om exposu e o en i on-
men al ha ms o access o en i onmen al
bene i s.
Be o e 2000, he e we e mainly s udies
belonging o he clus e s o was e and
pollu ion (Figu e 5A). In he yea 2000, ac-
cess o med a new clus e in EJ esea ch,
wi h collabo a ion ollowing in 2002, heal h
in 2004 and go e nance in 2005. The
clima e change clus e occu ed o he i s ime in 2012, and
g een space li e a u e eme ged in 2018. S udies in he clus e
was e peaked in 2013 wi h 6 publica ions, and pollu ion s udies
peaked in 2024 wi h 12 s udies. 2024 was also he yea wi h he
highes numbe o publica ions o all o he clus e s, excep o ac-
cess, which eached he highes numbe o 7 publica ions in 2023.
O e ime, he was e clus e showed a signi ican declining end
(es ima e = − 0.119, p < 0.001), indica ing educed ela i e ocus
despi e inc easing o al publica ions. The pollu ion, access, go e -
nance, and collabo a ion clus e s also exhibi ed dec easing bu
s a is ically non-signi ican ends. Con e sely, clima e change
and g een space clus e s showed posi i e ye non-signi ican in-
c eases, sugges ing eme ging in e es . The heal h clus e e-
mained ela i ely s able o e ime. Ac oss all con inen s, esea ch
in A ica is mainly ela ed o he go e nance clus e (n = 9), in Asia
o g een space (n = 16), in Eu ope o pollu ion (n = 28), and in No h
Ame ica o heal h (n = 50). The go e nance clus e also domina ed
in s udies in Oceania, Sou h Ame ica, and global o mul iple
geog aphic oci. The e was no esea ch on heal h o was e in A -
ica, Oceania, o Sou h Ame ica, al hough hese we e he domi-
na ing clus e s in No h Ame ica. Ano he s a k con as is in he
numbe o s udies in he g een space clus e , which domina ed
in Asia bu was only p esen wi h 1 publica ion in No h Ame ica,
and no p esen in any o he egion.
A Pea son’s Chi-Squa e es e ealed signi ican associa-
ions be ween clus e s and me hodological choices, wi h sam-
pling me hods (χ
2
= 126.87, d = 77, p = 0.0003), da a collec ion
me hods (χ
2
= 375.00, d = 147, p < 2.2e-16), and analysis
me hods (χ
2
= 257.52, d = 70, p < 2.2e-16) all showing non-
andom dis ibu ions ac oss clus e s. The domina ing da a
collec ion and analysis me hods pe g oup a e displayed in
Table 1. In he go e nance and collabo a ion clus e s, s udies
p edominan ly used quali a i e and pa icipa o y da a collec-
ion me hods such as in e iews and communi y engagemen ,
emphasizing p ocedu al jus ice and ad ocacy. S udies om
he go e nance clus e s we e also less likely o use mul iple
6 iScience 28, 113889, Decembe 19, 2025
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
Table
1.
G oup
cha ac e is ics
o de ed
by
size,
showing
a icle
coun s
and
he
op
75%
mos
equen
da a
collec ion
and
analysis
me hods
pe
g oup
by
cumula i e
occu ence
G oup #
A icles Fi e
mos
abundan
wo ds Desc ip ion Da a
collec ion
me hod Da a
analysis
me hods
3-Go e nance 73 ecogni ion,
go e nance,
li elihood,
pa icipa ion,
conse a ion
Mos
s udies
conduc ed
in
No h
Ame ica
(n
=
16).
Clus e
wi h
he
mos
global
s udies
(n
=
5).
Mos
s udies
use
a
quali a i e
o
mixed
app oach
and
in e iews
ocusing
on
p ocedu al
and
ecogni i e
(in)jus ice.
Topics
e ol e
a ound
conse a ion,
capaci y
building,
ecosys em
se ices,
and
land
igh s.
In e iews
(26%);
Pa icipa o y
me hods
(12%);
e iew
(11%);
GIS/ emo e
sensing
(10%);
su eys,
census,
and
ques ionnai es
(10%)
Quali a i e
coding
(21%);
Con en
analysis
(39%);
Desc ip i e
s a is ics
(55%);
Spa ial
analysis
(70%)
7-Pollu ion 72 dioxide,
emission,
ni ogen,
concen a ion,
pa icula e
Mos
s udies
conduc ed
in
Eu ope
(n
=
28).
The
majo i y
o
a icles
in
his
clus e
ocus
on
u ban
a eas
and
connec
ai
pollu ion
om
a ic
o
he
socio-economic
s a us
o
esiden s.
O he
a icles
e alua e
heal h
e ec s
o
p oximi y
o
indus ial
si es.
Mos
s udies
use
a
quan i a i e
app oach
and
spa ial
analysis
ools
ocusing
on
dis ibu ional
(in)jus ice.
Seconda y
da a
(33%);
GIS/ emo e
sensing
(23%);
su eys,
census,
and
ques ionnai es
(18%)
Uni a ia e
analysis
(39%);
Spa ial
analysis
(74%)
6-Clima e
change 62 hea ,
mi iga ion,
clima e,
co e ,
esilience
Mos
s udies
we e
conduc ed
in
No h
Ame ica
(n
=
38).
A icles
in
his
clus e
co e
ex eme
wea he
e en s
in
u ban
and
u al
a eas,
such
as
hea ,
loods,
and
d ough s,
seismic
isks.
A icles
de elop
decision
suppo
sys ems
o
educe
ha m
h ough
en i onmen al
haza ds
ia
isk
indices
o
imp o e
communi y
esilience.
GIS/ emo e
sensing
(36%);
Seconda y
da a
(25%)
Spa ial
analysis
(42%)
2-Heal h 53 bu den,
disease,
exposu e,
po e y,
ac
Mos
s udies
we e
conduc ed
in
No h
Ame ica
(n
=
50),
and
h ee
s udies
we e
om
Asia.
Mos
s udies
use
a
quan i a i e
app oach
using
da abases
and
census
ac
in o ma ion,
ocusing
on
dis ibu ional
(in)jus ice.
The
analyses
e ol e
a ound
heal h
isks,
p ima ily
cance ,
as hma,
and
hea
diseases,
and
h ough
ai
pollu ion
and
soil
con amina ion
ela ed
o
socio-
economic
s a us.
Seconda y
da a
(30%);
su eys,
census,
and
ques ionnai es
(20%);
GIS/ emo e
sensing
(16%);
clinical
ials
and
medical
da a
collec ion
(7%)
Uni a ia e
s a is ics
(55%)
1-Was e 52 mino i y,
ace,
a iable,
acili y,
acism
Mos
s udies
we e
conduc ed
in
No h
Ame ica
(n
=
45).
A icles
in
his
clus e
analyze
spa ial
inequali ies
om
haza dous
was e
dump
acili ies,
b own ields,
and
ai
oxics.
Mos
s udies
use
a
quan i a i e
app oach
and
spa ial
analysis
ools
ocusing
on
dis ibu ional
(in)jus ice.
Su eys,
census,
and
ques ionnai es
(29%);
GIS/ emo e
sensing
(26%)
Uni a ia e
s a is ics
(46%)
(Con inued
on
nex
page)
iScience 28, 113889, Decembe 19, 2025 7
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
Table
1. Con inued
G oup #
A icles Fi e
mos
abundan
wo ds Desc ip ion Da a
collec ion
me hod Da a
analysis
me hods
5-Access 43 accessibili y,
ec ea ion,
sa e y,
ac i i y,
anspo a ion
Mos
s udies
we e
conduc ed
in
No h
Ame ica
(n
=
26).
S udies
ocus
on
sa e y
in
u ban
en i onmen s,
access
o
g een
spaces,
and
he
e ec s
o
highway
cons uc ions.
Mos
s udies
use
a
quan i a i e
o
mixed
app oach
and
spa ial
analysis
ools
ocusing
on
dis ibu ional
(in)jus ice.
GIS/ emo e
sensing
(30%);
su eys,
census,
and
ques ionnai es
(28%)
Spa ial
analysis
(37%);
Uni a ia e
s a is ics
(64%)
4-Collabo a ion 37 pa ne ship,
ad ocacy,
ac ion,
collabo a ion,
eam
Mos
s udies
we e
conduc ed
in
No h
Ame ica
(n
=
36).
Mos
s udies
use
a
quali a i e
app oach
and
in e iews
ocusing
on
dis ibu i e
and
p ocedu al
(in)jus ice.
A icles
in
his
clus e
wo k
wi h
esea ch
app oaches
o
enhance
capaci y
building
using
ci izen
science
pa icipa ion,
pa icipa o y
esea ch,
and
communi y-
engaged
esea ch
me hods.
In e iews
(17%);
pa icipa o y
me hods
(14%);
su eys,
census,
and
ques ionnai es
(14%);
e iew
(10%);
seconda y
da a
(8%);
ocus
g oups
(7%)
Con en
analysis
(24%);
Spa ial
analysis
44%);
Desc ip i e
s a is ics
(62%)
8-G een
space 17 walking,
accessibili y,
a el,
supply,
demand
Mos
s udies
we e
conduc ed
in
Asia
(n
=
16),
and
one
s udy
was
om
he
US.
Mos
a icles
ocus
on
accessibili y
o
g een
spaces
in
u ban
a eas.
They
use
a
quan i a i e
app oach
and
spa ial
analysis
ools
ocusing
on
dis ibu ional
(in)jus ice.
GIS/ emo e
sensing
(40%);
Seconda y
da a
(20%)
Spa ial
analysis
(39%);
Uni a ia e
s a is ics
(59%);
Mul i a ia e
s a is ics
(75%)
8 iScience 28, 113889, Decembe 19, 2025
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
70. Pulido, L., and De La a, J. (2018). Reimagining ‘jus ice’in en i onmen al
jus ice: Radical ecologies, decolonial hough , and he Black Radical
T adi ion. En i on. Plan. E Na . Space 1, 76–98.
71. Aze edo, F., Jos , J.T., Ro hmund, T., and S e ling, J. (2019). Neolibe al
ideology and he jus i ica ion o inequali y in capi alis socie ies: Why so-
cial and economic dimensions o ideology a e in e wined. J. Soc. Issues
75, 49–88.
72. Ve meylen, S. (2019). En i onmen al Jus ice and Epis emic Violence (Tay-
lo & F ancis).
73. Ramcilo ic-Suominen, S. (2023). En isioning jus ans o ma ions in and
beyond he EU bioeconomy: inspi a ions om decolonial en i onmen al
jus ice and deg ow h. Sus ain. Sci. 18, 707–722.
74. A
´
l a ez, L., and Coolsae , B. (2020). Decolonizing en i onmen al jus ice
s udies: a La in Ame ican pe spec i e. Capi alism na u e socialism
31, 50–69.
75. Tempe , L. (2019). F om boome angs o mine ields and ca apul s: dy-
namics o ans-local esis ance o land-g abs. J. Peasan S ud. 46,
188–216.
76. Wong, P. (2020). Linking ‘‘local’’ o ‘‘global’’: aming En i onmen al Jus-
ice mo emen s h ough p og essi e con ex ualiza ion. In e ace 12,
215–243.
77. Tsing, A.L. (2024). F ic ion: An E hnog aphy o Global Connec ion (P ince-
on Uni e si y P ess).
78. Men on, M., La ea, C., La o e, S., Ma inez-Alie , J., Peck, M., Tempe ,
L., and Wal e , M. (2020). En i onmen al jus ice and he SDGs: om syn-
e gies o gaps and con adic ions. Sus ain. Sci. 15, 1621–1636.
79. Lede , S. (2024). T anslocal esou ce go e nance, social ela ions and as-
pi a ions: Linking anslocali y and Feminis Poli ical Ecology o explo e
a me -managed i iga ion sys ems and mig a ion in Nepal. Geo o um
148, 103905.
80. Con a do, T., and Loppi, S. (2024). Assessing En i onmen al Jus ice a he
U ban Scale: The Con ibu ion o Lichen Biomoni o ing o O e coming he
Dicho omy be ween P oximi y-Based and Dis ibu ion-Based Ap-
p oaches. A mosphe e 15, 275.
81. Ba le , L., and Va us, F. (2016). Re hinking Case S udy Resea ch: A
Compa a i e App oach (Rou ledge).
82. Thale , G.M. (2021). E hnog aphy o en i onmen al go e nance: Towa ds
an o ganiza ional app oach. Geo o um 120, 122–131.
83. Wilson, A.M., Polk, E., Field, C.B., and Fendo , S. (2025). Towa ds en i-
onmen al jus ice: A amewo k and s a egic app oach o implemen ing
communi y based pa icipa o y esea ch in he ea h and en i onmen al
sciences. En i on. Sci. Pol. 167, 104036.
84. Tassan, M. (2022). Re hinking en i onmen al jus ice in he An h opocene:
An an h opological pe spec i e. An h opol. Today 38, 13–16.
85. Po ada, H., Boelens, R., and Vos, J. (2024). B idging Jus ice S uggles: A
Poli ical Ecology o T anslocal Alliance Building agains Ex ac i e Indus-
ies. Al e nau as 11, 1–34.
86. Backhaus, J., and John, S. (2025). Gene aliza ion as local and anslocal
embedding: in e oga ing go e nance and decons uc ing democ a iza-
ion in li ing labs. Sus ain. Sci. P ac . Pol. 21, 2450856.
87. Sul ana, F. (2022). The unbea able hea iness o clima e coloniali y. Poli .
Geog . 99, 102638.
88. Dei anayagam, T.A., English, S., Hickel, J., Boni acio, J., Guin o, R.R., Hill,
K.X., Huq, M., Issa, R., Mulindwa, H., Nagginda, H.P., e al. (2023). En i-
sioning en i onmen al equi y: clima e change, heal h, and acial jus ice.
Lance 402, 64–78.
89. Massey, D.B. (2005). Fo space (SAGE Publica ions).
90. Walke , G. (2009). Globalizing en i onmen al jus ice: The geog aphy and
poli ics o ame con ex ualiza ion and e olu ion. Glob. Soc. Policy 9,
355–382.
91. Schlosbe g, D. (1999). En i onmen al Jus ice and he New Plu alism: The
Challenge o Di e ence o En i onmen alism (Ox o d Uni e si y P ess).
92. Tu nhou , E. (2024). A be e knowledge is possible: T ans o ming en i on-
men al science o jus ice and plu alism. En i on. Sci. Pol. 155, 103729.
93. Ahmed, S.K. (2024). Resea ch Me hodology Simpli ied. How o Choose
he Righ Sampling Technique and De e mine he App op ia e Sample
Size o Resea ch. O al Oncol. Rep. 12, 100662.
94. Abson, D.J., Von Weh den, H., Baumga
¨
ne , S., Fische , J., Hanspach, J.,
Ha
¨
d le, W., Hein ichs, H., Klein, A.-M., Lang, D.J., Ma ens, P., and
Walmsley, D. (2014). Ecosys em se ices as a bounda y objec o sus ain-
abili y. Ecol. Econ. 103, 29–37.
95. Du e
ˆ
ne, M., and Legend e, P. (1997). Species assemblages and indica o
species: he need o a lexible asymme ical app oach. Ecol. Monog . 67,
345–366.
iScience 28, 113889, Decembe 19, 2025 15
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
STAR★METHODS
KEY RESOURCES TABLE
METHOD DETAILS
We conduc ed a sys ema ic li e a u e e iew using he Scopus da abase on No embe 29, 2024, applying he ollowing sea ch s ing:
(TITLE-ABS(‘‘en i onmen al equi y’’ OR ‘‘en i onmen al inequi y’’ OR ‘‘en i onmen al jus ice’’ OR ‘‘en i onmen al injus ice’’ OR
‘‘en i onmen al dispa i ies’’ OR ‘‘en i onmen al inequali y’’ OR ‘‘en i onmen al equali y’’) AND TITLE-ABS(‘‘case s udy’’ OR ‘‘a ea’’
OR ‘‘s udy egion’’) AND TITLE-ABS(‘‘me hod’’ OR ‘‘assessmen *’’ OR ‘‘ ool*’’)).
We limi ed he sea ch o English-language, pee - e iewed a icles, esul ing in 730 ini ial eco ds. Sc eening occu ed in h ee
s ages: 1. Ti le sc eening excluded 256 a icles ha we e clea ly un ela ed o en i onmen al (in)jus ice o lacked an empi ical ocus.
2. Abs ac sc eening emo ed an addi ional 26 a icles ha did no mee ou inclusion c i e ia: empi ical ocus and engagemen wi h
en i onmen al (in)jus ice. 3. Full- ex sc eening excluded 27 mo e a icles due o insu icien ele ance o lack o me hodological
de ail.
EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAIL
The inal da ase included 421 a icles o ull analysis (Figu e 1).
We classi ied da a collec ion empo al scales in o i e ca ego ies: sho (<1 yea ), medium (1–5 yea s), long (>5 yea s), and e y long
(>30 yea s). S udies lacking empo al in o ma ion we e coded as missing. Da a ypes we e ca ego ized as quan i a i e, mixed, o
quali a i e. Sampling s a egies we e classi ied as p obabili y-based (e.g., simple andom, s a i ied, clus e , sys ema ic, mul i-s age)
o non-p obabili y-based (e.g., quo a, snowball, pu posi e, con enience) ollowing Ahmed (2024).
93
Full inclusion o census ap-
p oaches we e also no ed when en i e da ase s we e analyzed.
Da a collec ion me hods we e g ouped in o he ollowing b oad ca ego ies based on me hodological ai s: In e iews, su eys/
census/ques ionnai es, GIS/ emo e sensing, clinical ials o medial da a collec ion, simula ion & modelling, s a is ical designs,
ins umen al o labo a o y measu emen s, ield sampling, e iews, pa icipa o y me hods, obse a ion, ocus g oups, case s udies,
e hnog aphic me hods, socio-legal sou ces o seconda y da a sou ces. Da a analysis me hods we e coded in o uni a ia e s a is ics,
mul i a ia e s a is ics, desc ip i e s a is ics, spa ial analysis, machine lea ning & modelling, con ex ual analysis, pa icipa o y
me hods o analysis, quali a i e coding o con en analysis. Finally, we classi ied s udies by hei disciplina y o ien a ion (disciplina y,
in e disciplina y, mul idisciplina y o ansdisciplina y), and he dimensions o jus ice add essed.
We conduc ed wo ounds o in e code eliabili y checks du ing he i le and abs ac sc eening phases o ensu e consis ency. Fo
ull- ex coding, a iable de ini ions we e collabo a i ely de eloped and e ined h ough eam discussions. Each a icle was coded by
one esea che and e iewed o supplemen ed by a second. Al hough we did no calcula e o mal in e code ag eemen sco es,
collabo a i e e iew aimed o minimize inconsis encies.
QUANTIFICATION AND STATISTICAL ANALYSIS
We analyzed ou da ase a wo le els. We pe o med desc ip i e s a is ical analysis on all 421 a icles. This included uni a ia e an-
alyses o explo e associa ions be ween a iables using Chi-squa e es s, gene alized linea models (GLMs), log-linea models o
in es iga e p opo ions, Fishe ’s exac es s, and Kendall’s au ank co ela ions, depending on he ype and dis ibu ion o he
da a. Following Abson e al. (2014),
94
we conduc ed a mul i a ia e ull- ex wo d analysis based on nouns ex ac ed om he 409
REAGENT o RESOURCE SOURCE IDENTIFIER
So wa e and algo i hms
Py hon ( 3.12) Py hon So wa e Founda ion h ps://py hon.o g
R ( 4.4.2) R p ojec h ps://www. -p ojec .o g
O he
Li e a u e da abase o sys ema ic e iew Scopus h ps://www.scopus.com
Analy ical p ocedu es: Chi-squa e, GLM,
log-linea models, Fishe ’s exac es ,
Kendall’s au, hie a chical clus e analysis,
indica o species analysis, DCA
Analy ical p ocedu es Abson e al.
93
; This pape
Sys ema ic e iew epo ing guideline PRISMA 2020 h ps://p isma-s a emen .o g
e1 iScience 28, 113889, Decembe 19, 2025
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
a icle pd s. Assuming ha esea ch communi ies end o use dis inc ocabula ies, we applied hie a chical clus e analysis o g oup
a icles by simila i y in wo d use. Clus e ing was based on minimizing wi hin-g oup a iance and maximizing be ween-g oup di e -
ences in wo d equency dis ibu ions. To cha ac e ize each clus e , we applied an indica o species analysis,
95
ea ing wo ds as
‘‘species’’ and clus e s as ‘‘habi a s’’. This allowed us o iden i y s a is ically signi ican indica o wo ds, i.e., e ms ha a e disp o-
po iona ely associa ed wi h speci ic clus e s, o ming he seman ic co e o each esea ch clus e . Finally, we employed a de ended
co espondence analysis (DCA) o educe dimensionali y and isualize he p ima y axes o a ia ion in wo d use ac oss a icles. The
esul ing o dina ion plo shows indica o wo ds along wo p ima y axes, enabling in e p e a ion o he ela ionships be ween hema ic
clus e s in he en i onmen al jus ice li e a u e. All analyses we e conduc ed in R,Ve sion 4.4.2 (R Co e Team 2021).
iScience 28, 113889, Decembe 19, 2025 e2
iScience
Me a-analysis/sys ema ic e iew
ll
OPEN ACCESS
