EcoHealth

https://doi.org/10.1007/s10393-024-01673-y

Ó 2024 The Author(s)

Original Contribution

The Brief Solastalgia Scale: A Psychometric Evaluation

and Revision

Bruce K. Christensen,1 Conal Monaghan,1 Samantha K. Stanley,1 Iain Walker,1,2

Zoe Leviston,1,3 Emily Macleod,1 Rachael M. Rodney,4 Lisa-Marie Greenwood,1
Timothy Heffernan,1,5 Olivia Evans,1 Stewart Sutherland,1 Julia Reynolds,1
Alison L. Calear,6 Tim Kurz,7 and Jo Lane8
1
School of Medicine and Psychology, Australian National University, Canberra, Australia
2
Melbourne Centre for Behaviour Change, University of Melbourne, Melbourne, Australia
3
School of Arts and Humanities, Edith Cowan University, Joondalup, Australia
4
Centre for Entrepreneurial Agri-Technology, Australian National University, Canberra, Australia
5
School of Built Environment, University of New South Wales, Kensington, Australia
6
Centre for Mental Health Research, Australian National University, Canberra, Australia
7
School of Psychological Science, University of Western Australia, Perth, Australia
8
National Centre for Epidemiology and Population Health, Australian National University, Canberra, Australia

Abstract: Witnessing degradation and loss to one’s home environment can cause the negative emotional expe-
rience of solastalgia. We review the psychometric properties of the 9-item Solastalgia subscale from the Envi-
ronmental Distress Scale (Higginbotham et al. (EcoHealth 3:245–254, 2006)). Using data collected from three
large, independent, adult samples (N = 4229), who were surveyed soon after the 2019/20 Australian bushfires,
factor analyses confirmed the scale’s unidimensionality, while analyses derived from Item Response Theory
highlighted the poor psychometric performance and redundant content of specific items. Consequently, we
recommend a short-form scale consisting of five items. This Brief Solastalgia Scale (BSS) yielded excellent model fit
and internal consistency in both the initial and cross-validation samples. The BSS and its parent version provide
very similar patterns of associations with demographic, health, life satisfaction, climate emotion, and nature
connectedness variables. Finally, multi-group confirmatory factor analysis demonstrated comparable construct
architecture (i.e. configural, metric, and scalar invariance) across validation samples, gender categories, and age.
As individuals and communities increasingly confront and cope with climate change and its consequences,
understanding related emotional impacts is crucial. The BSS promises to aid researchers, decision makers, and
practitioners to understand and support those affected by negative environmental change.

Bruce K. Christensen, Conal Monaghan and Samantha K. Stanley have contributed

equally to the preparation of this article.

Supplementary Information: The online version contains supplementary material

available at https://doi.org/10.1007/s10393-024-01673-y.

Correspondence to: Bruce K. Christensen, e-mail: [email protected]

B. K. Christensen et al.

Keywords: Bushfire, Solastalgia, Environmental Distress Scale, Psychometric, Factor analysis, Short form

A growing body of research examines the emotional impact solastalgia. For example, the EDS-S has revealed that Tex-
of climate change and ecological disasters (e.g. Clayton & ans living in areas with more oil and gas wells experience
Karaszia, 2020; Hogg et al., 2021; Stanley et al., 2021). heightened solastalgia (Elser et al., 2010), that most resi-
Solastalgia is one response and is characterised by sadness, dents living through rapid urbanisation in Pakistan expe-
grief, and powerlessness caused by the transformation and rienced solastalgia (Khan et al., 2012), and that solastalgia
degradation of one’s environment (for a review, see Galway in a community in Ireland affected by coastal erosion was
et al., 2019). Solastalgia has been likened to a sense of highest amongst long-term residents (Phillips & Murphy,
‘‘homesickness one gets when one is still at ‘home’’’ (Al- 2021). In another context, Eisenman et al. (2015) observed
brecht, 2005, p. 45). solastalgia amongst those affected by wildfires in Arizona
Albrecht and colleagues developed the construct and that greater experiences of solastalgia predicted more
through their work with residents of the Upper Hunter severe psychological distress one year after the fires.
Region of New South Wales, Australia, who experienced Therefore, not only does environmental degradation ap-
significant distress living near open cut coal mines (Al- pear to contribute to the experience of solastalgia, but it
brecht, 2005; Albrecht et al., 2007; Connor et al., 2004). may also place people at risk of poorer psychological
Similar case studies identify solastalgia amongst interviews wellbeing in the future.
with women in the Torres Strait talking about climate The consistent performance of the EDS-S notwith-
change (McNamara & Westoby, 2011), farmers in rural standing, psychometric validation of this subscale remains
Australia experiencing mental health effects from drought limited. Although previous studies have demonstrated
(Satore et al., 2008), and Inuit people in Canada describing adequate internal consistency, with Conbrach’s alpha val-
the effects of the changing climate on their lives and mental ues ranging from 0.75 (Warsini et al, 2014) to 0.93
health (Cunsolo et al., 2012). While most research on (Higgenbotham et al., 2006), the scale developers did not
solastalgia has been qualitative in nature (Galway et al., conduct a principal components or factor analysis of the
2019), Higginbotham et al. (2006) developed the Envi- solastalgia items themselves—instead verifying that the
ronmental Distress Scale (EDS), with a subscale measuring mean scores of each EDS subscale loaded onto a single
solastalgia. Developed from the content of interviews with factor indexing overall environmental distress. Others have
the Upper Hunter community and grounded in Environ- variously interpreted the EDS-S as unidimensional
mental Stress and Risk Theory, the EDS includes six sub- (Eisenman et al., 2015; Luce, 2021) or multidimensional
scales, namely: perceptions of environmental hazard (Warsini et al., 2014) and, thus, further investigation of the
(including both frequency and observation of hazard underlying dimensionality is needed.
events), appraisal of threat, felt impact of environmental Furthermore, it is difficult to draw comparisons be-
change, solastalgia, and environmental action. Higgin- tween past studies because each employed slightly different
botham and colleagues (2006) found that solastalgia—and versions of the EDS-S items. Higginbotham et al. (2006)
overall EDS scores—were higher amongst residents of the validated the EDS within a mining context, although they
Upper Hunter compared to those in a farming community recommended that researchers adapt the measure to ap-
living at a distance from the environmental degradation of praise environmental distress in the face of other envi-
the open cut mine. In this study, the solastalgia subscale ronmental and human challenges, including natural
returned excellent internal consistency estimates (a = 0.93) disasters and human conflict/war. Subsequent users of the
and acceptable test–retest reliability scores (intra-class scale have varied the items to refer to the specific envi-
correlation = 0.73). ronmental disaster of interest. For example, the item ‘‘I am
Researchers generally describe the EDS solastalgia saddened by unwelcomed change I see in my landscape’’
subscale (EDS-S) as effective (Eisenman et al., 2015; Elser was adjusted to ‘‘I feel sad when I look at the landscapes
et al., 2020; Khan et al., 2012; Phillips & Murphy, 2021) and damaged by the Wallow Fire’’ in Eisenman et al.’s (2015)
have used this measure to show that those living near de- research following a wildfire, to ‘‘I feel saddened by the loss
graded landscapes experience higher levels of self-reported of the beach at Courtown’’ in Phillips and Murphy’s (2021)
 The Brief Solastalgia Scale

study on coastal erosion, and to ‘‘I am saddened when I levels of solastalgia related to poorer self-rated physical
look at degraded landscapes and open-cut mine voids’’ in health.
Elser et al.’s (2020) research on the impacts of mining.
Other minor changes to the emotional response of
environmental change have also been reported. For exam- METHOD
ple, the item ‘‘I am ‘upset’ at the way this area looks now’’
has been variously replaced with ‘ashamed’ (Elser et al., We analyse data from three samples that included the EDS-
2020) or ‘disappointed’ (Phillips & Murphy, 2021; Warsini S. As these data were each collected for different primary
et al., 2014). Additionally, authors may select only a subset purposes, there are small differences in survey methods,
of the nine original items, either based on the items they which are highlight below.
felt were relevant (e.g. four items were used in Khan et al.,
2012), or for unstated reasons (e.g. seven items were used Participants and procedure
in Phillips & Murphy, 2021). These adaptations make it
Sample 1 was collected between March and April of 2020.
unclear whether studies are capturing a similar experience
The sample included 1,776 participants aged between 18
of solastalgia across contexts, raising significant challenges
and 87 years (M = 49.80, SD = 16.51, median = 51 years;
regarding how best to compare experiences of solastalgia
59.10% female, 40.51% male, 0.23% ‘other’, 0.17% pre-
across environmental events (Supplementary Material A
ferring not to say). Participants were recruited from the
shows each iteration of the scale items that we have iden-
Canberra and surrounding regions as part of a study de-
tified in the literature). For the purposes of the current
signed to examine the effects of the 2019–2020 bushfires on
study, wording for each item from the initial EDS-S was
the health and wellbeing of residents affected by heavy
used. Although bespoke item wording may increase the
smoke (Rodney et al., 2021), using postal invitations, a
relevance of individual items to a given context, such
Qualtrics panel sample, and convenience sampling methods
variability makes psychometric comparisons difficult and,
(social media and radio advertising, and local media sto-
therefore, the current study endorses consistent wording to
ries). Participants were eligible for inclusion in the study if
bolster comparability and measurement rigour.
they were living in Canberra or surrounding regions be-
Collectively, these results and observations underscore
tween 15 December 2019 and 15 February 2020 during the
the need for more focussed investigation of the dimen-
2019–2020 bushfire season. Although 2,095 completed the
sionality, validity, and reliability of the solastalgia subscale,
survey, participants were excluded if they were out of the
which is the primary goal of the current research. We also
area of interest or did not provide a valid postcode
consider solastalgia’s associations with demographic vari-
(n = 11), did not respond to all EDS-S items (n = 276), or
ables (age, gender, income, education, and perceived social
if they were identified as multivariate outliers on EDS-S
status) and with the theoretically-relevant constructs of
items (Mahalanobis, v2(9) = 27.88, p < 0.001; n = 32), to
perceived environmental change, nature connectedness,
reach our final sample of 1776.
climate emotions, identification as an environmentalist,
Sample 2 was collected between August 20 and
and life satisfaction. We expected those who believe their
September 20 of 2020. The sample included 1,651 partici-
environment is substantially worsening and who are more
pants aged between 18 and 85 years (M = 44.89, SD =
connected to nature and/or identify as environmentalists
17.62; median = 43 years; 46.94% female, 52.63% male,
are likely more sensitive to environmental change and,
0.18% ‘other’, 0.24% preferring not to say). Participants
thus, may experience higher self-reported levels of
were recruited for a national survey that was designed to
solastalgia. Higher levels of solastalgia should also relate to
examine Australians’ attitudes, emotions, and behaviours
greater emotional responses to climate change. Lastly, and
related to climate change. Participants were recruited
given the small associations with wellbeing in past literature
nationwide via a Qualtrics panel sample, with quotas in
(Eisenman et al., 2015), we expected solastalgia to be
place so that the final sample was representative of the adult
negatively related to life satisfaction, and positively related
Australian population (based on the 2016 census) in age,
to ecologically related anxiety and stress. Based on Al-
gender, and location. Participants were eligible for inclu-
brecht’s (2005) speculation that solastalgia can escalate into
sion in the national survey if the quota for their demo-
physical symptoms, we also investigated whether higher
graphic group was not yet full, and if they met quality
B. K. Christensen et al.

checks (i.e., passed attention checks presented in the first test and parallel analysis (95th percentile, 2000 simulated
third of the survey, did not take less than half the median samples), was used to reduce the likelihood of overdi-
time to complete the full survey, or respond with a pattern mensionalisation (van der Eijk & Rose, 2015) and to
or unrealistic responses such as unintelligible text in open- identify inadequately loading items (< 0.50 was utilised
ended text boxes). In total, 5,110 participants met these given the focus on a robust final scale). Item Response
quality requirements, but only those who responded to the Theory (IRT; Graded Response Model [GRM]; Samejima,
question: ‘‘Do you think that over the last few years the 1969) then estimated where (known as the location
quality of your local environment is getting better, staying parameter) on the solastalgia continuum items provided
the same, or getting worse’’ with response options ‘getting a information about participants. We focussed on balancing
bit worse’ or ‘getting much worse’ (n = 1,749) were pre- information while ensuring measurement across the entire
sented with the EDS-S items. Of this subset of participants, - 2.5 to + 2.5 logit range (interpreted similarly to Z-
we further excluded those who did not respond to all items scores). Poorer items were iteratively removed during the
(n = 30) or were identified as a multivariate outlier (sig- IRT stage based on distributional (e.g. censored / truncated
nificant Mahalanobis distances, p < 0.001; n = 68), to or bimodal distributions) and IRT parameters. In partic-
reach our final sample of 1,651. ular, items were retained if they yielded low to moderate
Sample 3 was collected between January and June estimates of local dependence, robust information curves,
2021. The sample included 802 participants aged between and unique positional information on the solastalgia con-
18 and 86 years (M = 42.59, SD = 16.26, median = 41 tinuum (see Results below).
years; 43.02% female; 56.23% male, 0.50% ‘other’, 0.25% Confirmatory Factor Analysis (CFA) on the confir-
prefer not to say). Participants were recruited as part of a matory subsample evaluated the factor structure’s replica-
nationwide survey designed to investigate impacts of Aus- bility and structural validity (i.e., that conclusions from the
tralia’s 2019/2020 bushfires on mental health and wellbe- exploratory subsample were not measurement artefacts or
ing, using postal invitations, a Qualtrics paid sample, and Type I errors). The robustness of the structure was then
convenience sampling methods (social media and radio further tested through a series of CFA measurement
advertising, and local media stories). Participants were invariance analyses (sample, age, and gender identity)
eligible for inclusion in the larger study if they were 18- across all three samples. Finally, the brief scale was sub-
years or over, had been living in Australia since the 2019– jected to external validity analysis to provide evidence for
2020 bushfire season, and passed the quality checks out- its construct validity.
lined for Sample 2. In total, 3,083 participants met these
criteria, and those who reported living in a bushfire-af-
fected postcode (identified by the Australian Government) RESULTS
at the start of the bushfire season (n = 569) or being in-
volved in fighting the fires (n = 343) were deemed bush- Item Reduction
fire-affected and thus presented with the EDS-S. Of this An investigation of item distributions suggested that several
group, multivariate outliers (n = 18) and those who re- items had skewed distributions: specifically, items 1 (posi-
sponded with ‘‘does not apply’’ to any item, were removed tive), 2 (negative), and 3 (negative). Only item 1’s skew was
prior to analysis, to reach a final sample of 802. considered incorrigible (censored), with all other items
having appropriate variation in relation to their mean,
Analytic Strategy thereby allowing adequate sampling variability. (Item dis-
tributions can be seen in Supplementary Material B.) All
Data were analysed in RStudio (2022.02.3; R version 4.1.1). items were subjected to a polychloric EFA (given that the 5-
Sample 1 data were randomly split into exploratory point response options were not considered adequate for
(n = 855) and confirmatory (n = 921) subsamples. Ex- underlying continuous data). Data were considered
ploratory analyses subjected the nine items from the appropriate for EFA based on the Kaiser–Meyer–Olkin test
Solastalgia scale to a schedule of item-level property anal- (0.91), and significant Barlett’s test, v2(36) = 12,531.64,
ysis, with each approach providing important and com- p < 0.001. The Scree plot and parallel analysis strongly
plementary information about item performance. supported a single dominant factor, with parallel analysis
Exploratory Factor Analysis (EFA), utilising Cattell’s elbow
 The Brief Solastalgia Scale

allowing for a very weak secondary factor. Since the DCFI > 0.01, suggests these parameters are not equivalent
Eigenvalue of the second factor was well below 1 (i.e. 0.38), between the groups (Table 2).
the single factor solution was adopted, with item loadings Three series of invariance analyses were conducted,
ranging from 0.46 (item 1) to 0.89 (item 8). (Item loadings comparing samples, binary gender identities, and then
are available in Supplementary Material C.) At this stage, across age cohort. As age is a continuous variable, we
Item 1 was removed from further analyses, given the created three age groups, < -1SD, -1SD to 1SD, and >
strength of the loadings of all other items (> 0.68) and the 1SD. Across all comparisons, the solastalgia scale was
incorrigible skew identified during the previous item- invariant at the configural, metric, and scalar levels (Chen,
analysis. 2007; Cheung & Rensvold, 2002), indicating that these
An initial round of IRT was conducted on the eight factor structures are equivalent without any systematic bias
candidate items. Approximately 98% of participants fitted that might skew mean level comparisons (scalar invariance)
the model based on Zh values > - 1.96, and non-fitting (Table 3).
participants were removed from further IRT analyses Finally, IRT analyses were run on the final scale (Fig. 1)
(n = 36) to reduce the introduction of unnecessary error using Sample 2 to maximise the sample size of participants
variance. The assumption of local dependence (LD) was not already used in scale development and to ensure the
also evaluated; however, these values can be difficult to robustness of the IRT parameters across samples and re-
interpret with a small number of items. Four pairs of items sponse scales. The results demonstrated that all items
showed low to moderate LD estimates ( >|.30|), and provided strong information about the middle of the latent
potentially inflated information parameters were consid- trait, with relatively clear distinctions between each of the
ered when selecting the final items (Yen, 1984). response category information curves (information ranged
Three items that yielded substantially higher informa- from 2.13 to 4.41; See Supplementary Material D.II for IRT
tion curves than the remaining items were retained (see parameters). In addition, the final scale captured the
Supplementary Material D for the original item response majority of its information in the - 2 to 2 logit range, ideal
curves). One additional item provided moderate informa- for studies on the general population (Fig. 2).
tion; however, because this was primarily situated below -
1.5 logits (non-solastalgia end of the continuum) it was External Validity
removed. Two items provided unique information above
1.5 logits (solastalgia end of the continuum), and we chose External validity analysis correlated BSS scores with exter-
to retain one of these items to ensure the final set contained nal variable scores in Sample 2 (Table 4). Demographic
sufficient content breadth. We named the final 5-item scale groups experienced solastalgia at similar levels, though
the Brief Solastalgia Scale (BSS) (Table 1; Appendix 1), and solastalgia is experienced to a slightly lower degree amongst
then subjected it to confirmatory analyses. those who tend to perceive themselves as higher in sub-
jective social status. Solastalgia is also related to lower life
satisfaction, though unrelated to subjective health. We also
Final Scale Evaluation and Confirmatory Analyses
found that people experiencing solastalgia tended to feel
We evaluated measurement invariance, sequentially con- more connected to nature, experience more intense nega-
straining parameters across each group to equality. This tive emotions about climate change, and were more likely
places a higher standard of equivalence between the models to report that their environment is getting much worse than
in each sample. After initially fitting the model to all a bit worse.
samples without constraints (configural invariance), we
sequentially constrained the factor loadings (metric),
intercepts (scalar), and then means (strict) across each
group to equality. A substantial reduction in model fit,
B. K. Christensen et al.

Table 1. Mean (Standard Deviation) Responses to Solastalgia Items Across Samples.

Item Sample 1 Sample 2 Sample 3

1. My sense of belonging to this place has been undermined by recent changes* 2.21(1.18) 4.00(1.37) 3.00(1.61)
2. I am sad that familiar aspects of this place are disappearing (e.g. animals, plants, landmarks, open 4.00(1.02) 5.43(1.28) 3.98(1.63)
space)
3. I am worried that aspects of this area that I value are being lost 3.73(1.08) 5.28(1.25) 3.90(1.58)
4. I miss having the peaceful feeling that I once enjoyed by being in this place 3.16(1.22) 4.89(1.38) 3.77(1.63)
5. I am upset at the way this area looks now 2.98(1.14) 4.63(1.47) 3.63(1.62)
6. My lifestyle is being threatened by change in my local area 2.74(1.19) 4.16(1.47) 3.46(1.58)
7. Unique aspects of nature that made this place special are being lost forever 3.37(1.20) 4.90(1.38) 3.80(1.60)
8. I am saddened by unwelcome change I see in my landscape 3.42(1.14) 4.96(1.33) 3.79(1.58)
9. I feel powerless to stop unwanted changes to this place 3.33(1.13) 5.03(1.38) –
Mean solastalgia score (of 9 items) 3.22 (.85) 4.81(1.04) 3.67(1.31)
Mean solastalgia score (of 5 items) 3.24 (.94) 4.79(1.14) 3.72(1.37)
Response scale 1–5 1–7 1–5

Item 1 varied slightly by sample (see Method section for details). Bolded items are those retained in the Brief Solastalgia Scale.
Item 9 not included in Sample 3. Standard deviation in brackets.

Table 2. CFA Fit Statistics, Internal Consistency, and Fidelity for the Solastalgia Factor in Each Sample.

Sample Test Statistic Incremental Fit Absolute Fit Internal Consistency Correlation with 9-item scale

v2 p CFI NNFI SRMR RMSEA RMSEA 90% CI Alpha Omega Pearson

1a .89 .89 .97
1b 2.719 .743 1.000 1.002 .015 .000 [.000, .033] .88 .89 .97
2 9.716 .084 .999 .997 .025 .024 [.000, .046] .89 .89 .97
3 7.779 .169 .999 .998 .027 .026 [.000, .060] .91 .91 .98

Sample 1a is the exploratory subsample, and Sample 1b is the confirmatory subsample. Model 1 has all five items loading onto a single.
solastalgia factor, models estimated using Diagonally weighted Least Squares (DWLS). Baseline df = 10, and model df = 5. Maximum likelihood estimation
also returned similar fit estimates (Hu & Bentler, 1999) (see Supplementary Material D).

DISCUSSION formed well regardless of whether it was operationalised

using a 5-point or 7-point Likert scale.
This is the first study to systematically evaluate solastalgia Interestingly, exploratory and confirmatory factor
measurement through factor analytic and item response analyses strongly supported a single underlying dimension
approaches. Here, a refined 5-item BSS is presented. The of solastalgia, clarifying previous uncertainty surrounding
BSS performed well, correlating strongly with the longer 9- its unidimensional (Eisenman et al., 2015) or multidi-
item scale, with high internal consistency, and a robust mensional (Warsini et al., 2014) nature. Therefore, indi-
single factor structure. Its outstanding psychometric per- viduals do not appear to differentiate between core aspects
formance, even with just over half the number of items of of solastalgia (e.g. worry, loss, powerlessness), instead see-
the original scale, enables expedient measurement of ing each aspect contributing towards the greater negative
solastalgia to facilitate the quantitative study of distress emotional experience. Measurement invariance also indi-
resulting from environmental degradation. The BSS per- cated that this structure was consistent across ages, samples,
 The Brief Solastalgia Scale

Table 3. CFA invariance fit statistics for the Solastalgia factor across samples and across age groups.

Analysis Model Test Statistic Alternative Fit

2
df p CFI RMSEA
Sample Configural 23.504 15 .974 .999 .020
Metric 35.900 23 .042 .999 .020
Scalar 133.421 31 < .001 .992 .049
Binary Gender Identity Configural 16.550 10 .85 1.000 .018
Metric 21.905 14 .081 .999 .016
Scalar 35.594 18 .008 .999 .022
Age Configural 18.326 15 .246 1.000 .013
Metric 26.277 23 .288 1.000 .010
Scalar 38.599 31 .164 .999 .013

Model invariance tested without estimated residual covariance for parsimony. All models estimated using Diagonally weighted Least Squares (DWLS).
Unfortunately, only binary gender identities were included in the analyses due to sample size constraints.

Figure 1. Item information curves and item response category curves for the nine items of the EDS-Solastalgia scale.

and binary gender identities. This key finding indicates that systematic differences in how solastalgia is measured that
how participants view solastalgia does not vary between would artificially influence mean solastalgia scores.
bushfire (Samples 1 and 3) and more general environ- Experiences of solastalgia have, to date, predominantly
mental change (Sample 2). It also suggests that demo- been understood through interviews (Galway et al., 2019).
graphic groups view solastalgia in the same way, and older Such qualitative work is informative as it provides a rich
people perceive solastalgia the same way as younger people. understanding of the lived experience of environmental
Importantly, this invariance enables valid comparisons of distress. However, quantitative approaches are needed to
mean differences across groups and samples, as there are no learn about the prevalence of solastalgia and how intensely it
B. K. Christensen et al.

Figure 2. A Information curves for each of the 9 Solastalgia items. B Item probability functions for each of the 9 Solastalgia items. Note. A
higher resolution version will be attached upon publication.

Table 4. Correlations between the Solastalgia scale and external validity variables.

Domain Measure Brief Solastalgia scale

Demographics Age - .04
Gender+ .03
Income .02
Education .02
Subjective social status - .07**
Wellbeing Life satisfaction - .10***
Subjective health - .03
Environmental Nature relatedness .42***
Eco-anxiety .32***
Eco-depression .31***
Eco-anger .31***
Identification as an environmentalist .29***
+ Local environmental change - .26***
(1 = getting much worse, 2 = getting a bit worse)

Results are from Sample 2. Correlations with binary variables are point-biserial correlations.
*
p < .05, ** p < .01, *** p < .001, +Binary variable given sample size.

is experienced across groups and in relation to other envi- who are more connected to nature, and those who experience
ronmental attitudes. It is also useful for researchers and more intense emotions when thinking about climate change.
community organisations, such as those focussed mental Although Albrecht and colleagues (2005, 2007) have sug-
health and environmental advocacy, to have at their disposal gested that solastalgia may be a ‘psychoterratic illness’ cap-
a very brief and economical indicator of solastalgia. Con- able of manifesting in physical illness, we did not find an
tributing to these gaps, the BSS reveals associations compa- association with self-rated health. This was consistent when
rable to the full 9-item scale with related constructs and solastalgia was operationalised using the EDS-Solastalgia and
supports the idea that solastalgia is felt more intensely by BSS and, thus, reflects that solastalgia can be experienced
those who perceive greater environmental change, by those independent of general physical wellbeing.
 The Brief Solastalgia Scale

Expressions of concern and grief about changes to one’s CONCLUSIONS

landscape are key characteristics of solastalgia (Albrecht,
2007). Supporting the face validity of the BSS, the items we Our work builds on Higginbotham and colleagues’ (2006)
retained captured worry and sadness about environmental substantive scale development efforts by producing a valid
change and loss and how this threatens one’s lifestyle, sug- and reliable 5-item solastalgia scale. Importantly, and sup-
gesting these are the core components of solastalgia. How- ported by measurement invariance, we present the BSS as a
ever, people experiencing environmental change also express standard set of items. While the original authors recommend
feelings of powerlessness and a disrupted sense of belonging altering the items to the population and environmental
(Albrecht, 2007; Phillips & Murphy, 2021). Interestingly, degradation of interest, we recommend retaining the current
items tapping these experiences were not retained in the BSS item wording to reduce item heterogeneity and increase the
because they were less informative than items capturing validity of cross-study and cross-sample comparisons. This
lifestyle change—i.e. a sense of loss and a sense of sadness due means that solastalgia scores can be compared more generally
to environmental change. This may suggest that feelings of to understand the distress evoked by different environmental
belonging and powerlessness are context specific, or related events, and monitor how solastalgia changes across time,
to (but separate from) solastalgia itself. circumstances and communities, which is particularly rele-
While the BSS demonstrates many desirable psycho- vant as the climate crisis continues.
metric characteristics, several limitations of scale and the
current study should be acknowledged. Standardised error
increases rapidly outside approximately 95% of the middle FUNDING
of the latent trait (- 2 to 2 logit range), despite specifically
selecting items to capture the higher ends of solastalgia Open Access funding enabled and organized by CAUL and
through IRT. As a result, this BSS is ideal for measuring the its Member Institutions. This research was supported by a
general response (population) to solastalgia. To measure grant from the Medical Research Future Fund (2020/
the high ends, new item/s specifically designed for this MRF1201732).
population would be useful (there were no candidate items
in the original 9 items that met this criterion). Potentially,
DECLARATIONS
those heavily affected may have unique and traumatic CONFLICT OF INTEREST The authors have no con-
stories that are more appropriate for interview or short flicts of interest to declare, financial or otherwise
answer responses rather than brief self-report instruments.
Additionally, whether the BSS is equally valid in measuring
solastalgia across varied environmental alterations (e.g.
floods, bushfires, mining, deforestation) cannot be adju- OPEN ACCESS
dicated by the current data and requires additional research
to address this issue. Similarly, the current results do not This article is licensed under a Creative Commons
yet support the use of the BSS as a clinical or individual Attribution 4.0 International License, which permits use,
measure of solastalgia since normative data were not col- sharing, adaptation, distribution and reproduction in any
lected across a representative population. Instead, the medium or format, as long as you give appropriate credit to
samples here represent individuals at-risk for solastalgia the original author(s) and the source, provide a link to the
given their recent experience of the 2019/20 Australian Creative Commons licence, and indicate if changes were
bushfires and/or their experience of local environmental made. The images or other third party material in this
decline. To estimate quantitative departures from normal article are included in the article’s Creative Commons
levels of solastalgia, an appropriate normative sample licence, unless indicated otherwise in a credit line to the
would have to be ascertained and could be a constructive material. If material is not included in the article’s Creative
focus for future research. In the meantime, the level of Commons licence and your intended use is not permitted
solastalgia in the current sample can be estimated from the by statutory regulation or exceeds the permitted use, you
aggregated mean, and standard deviation, across the three will need to obtain permission directly from the copyright
samples for the BSS, which is 3.19 (0.87). holder. To view a copy of this licence, visit http://creativec
ommons.org/licenses/by/4.0/.
B. K. Christensen et al.

APPENDIX 1: THE BRIEF SOLASTALGIA SCALE

Instructions
Please rate (circle the number that best describes) the extent you agree or disagree with the following statements relating to
change in your local environment.

Scoring Instructions
Calculate the mean score based on responses to all survey items. Please note, there are no reversed-scored items.

Statements
1. I am worried that aspects of this area that I value are being lost.

1 2 3 4 5 N/A
Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Does not apply

2. I am upset at the way this area looks now.

1 2 3 4 5 N/A
Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Does not apply

3. My lifestyle is being threatened by change in my local

area.

1 2 3 4 5 N/A
Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Does not apply

4. Unique aspects of nature that made this place special

are being lost forever.

1 2 3 4 5 N/A
Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Does not apply

5. I am saddened by unwelcome change I see in my

landscape.

1 2 3 4 5 N/A
Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Does not apply
 The Brief Solastalgia Scale

REFERENCES Hu LT, Bentler PM (1999) Cutoff criteria for fit indexes in

covariance structure analysis: conventional criteria versus new
alternatives. Structural Equation Modelling: A Multidisciplinary
Adler NE, Epel E, Castellazzo G, Ickovics J (2000) Relationship of Journal 6(1):1–55. https://doi.org/10.1080/10705519909540118
subjective and objective social status with psychological and
physiological functioning: preliminary data in healthy white Khan NY, Ghafoor N, Iftikhar R, Malik M (2012) Urban annoy-
women. Health Psychology 19:586–592. https://doi.org/10.1037/ ances and mental health in the city of Lahore Pakistan. Journal
0278-6133.19.6.586 of Urban Affairs 34(3):297–315. https://doi.org/10.1111/j.1467-
9906.2011.00585.x
Albrecht G, Sartore GM, Connor L, Higginbotham N, Freeman S,
Kelly B, Pollard G (2007) Solastalgia: the distress caused by Luce, C. (2021), Grief, loss and climate change: validation of a
environmental change. Australasian Psychiatry 15(sup1):S95– solastalgia scale. [Doctoral dissertation thesis, Virginia Com-
S98. https://doi.org/10.1080/10398560701701288 monwealth University].
Albrecht G (2005) ‘‘Solastalgia’’. A new concept in health and McNamara KE, Westoby R (2011) Solastalgia and the gendered
identity. PAN Philosophy Activism Nature 3:41 nature of climate change: an example from Erub Island. Torres
Strait. Ecohealth 8(2):233–236. https://doi.org/10.1007/s10393-
Chen FF (2007) Sensitivity of goodness of fit indexes to lack of 011-0698-6
measurement invariance. Structural Equation Modelling 14:464–
504. https://doi.org/10.1080/10705510701301834 Nisbet EK, Zelenski JM (2013) The NR-6: A new brief measure of
nature relatedness. Frontiers in Psychology 4:813. https://doi.org/
Cheung GW, Rensvold RB (2002) Evaluating goodness-of-fit in- 10.3389/fpsyg.2013.00813
dexes for testing measurement invariance. Structural Equa-
tion Modelling: A Multidisciplinary Journal 9:233–255. https:// Phillips C, Murphy C (2021) Solastalgia, place attachment and
doi.org/10.1207/S15328007SEM0902_5 disruption: insights from a coastal community on the front line.
Regional Environmental Change 21(2):1–14. https://doi.org/
Clayton S, Karazsia BT (2020) Development and validation of a 10.1007/s10113-021-01778-y
measure of climate change anxiety. Journal of Environmental
Psychology 69:101434. https://doi.org/10.1016/ Rodney RM, Swaminathan A, Calear AL, Christensen BK, Lal A,
j.jenvp.2020.101434 Leviston Z, Reynolds J, Trevenar S, Vardoulakis S, Walker I
(2021) Physical and mental health effectos fo bushfire and
Connor L, Albrecht G, Higginbotham N, Freeman S, Smith W smoke in the Australian capital territory 2019–20. Frontiers of
(2004) Environmental change and human health in upper Public Health 9:1–13. https://doi.org/10.3389/
hunter communities of new south Wales Australia. EcoHealth. . fpubh.2021.682402
https://doi.org/10.1007/s10393-004-0053-2
Samejima F (1969) Estimation of latent ability using a response
Cunsolo Willox A, Harper SL, Ford JD, Landman K, Houle K, pattern of graded scores. Psychometrika Monograph Supplement
Edge VL (2012) ‘‘From this place and of this place:’’ climate 34(Supp 1):1–97. https://doi.org/10.1007/BF03372160
change, sense of place, and health in Nunatsiavut Canada. Social
Science & Medicine 75(3):538–547. https://doi.org/10.1016/ Sartore GM, Kelly B, Stain H, Albrecht G, Higginbotham N (2008)
j.socscimed.2012.03.043 Control, uncertainty, and expectations for the future: a quali-
tative study of the impact of drought on a rural Australian
Eisenman D, McCaffrey S, Donatello I, Marshal G (2015) An community. Rural and Remote Health 8(3):1–14
ecosystems and vulnerable populations perspective on solastal-
gia and psychological distress after a wildfire. EcoHealth Stanley SK, Hogg TL, Leviston Z, Walker I (2021) From anger to
12(4):602–610. https://doi.org/10.1007/s10393-015-1052-1 action: differential impacts of eco-anxiety, eco-depression, and
eco-anger on climate action and wellbeing. The Journal of Cli-
Elser H, Goldman-Mellor S, Morello-Frosch R, Deziel NC, Ran- mate Change and Health 1:100003. https://doi.org/10.1016/j.jo-
jbar K, Casey JA (2020) Petro-riskscapes and environmental clim.2021.100003
distress in West Texas: community perceptions of environ-
mental degradation, threats, and loss. Energy Research & Social van der Eijk C, Rose J (2015) Risky business: Factor analysis of
Science 70:101798. https://doi.org/10.1016/j.erss.2020.101798 survey data—assessing the probability of incorrect dimension-
alisation. PLoS ONE 10(3):e0118900. https://doi.org/10.1371/
Galway LP, Beery T, Jones-Casey K, Tasala K (2019) Mapping the journal.pone.0118900
solastalgia literature: a scoping review study. International
Journal of Environmental Research and Public Health Warsini S, Buettner P, Mills J, West C, Usher K (2014) Transla-
16(15):2662. https://doi.org/10.3390/ijerph16152662 tion, cultural adaptation, and psychometric testing of the
environmental distress scale withIndonesian survivors of a
Higginbotham N, Connor L, Albrecht G, Freeman S, Agho K volcanic eruption. Disaster Medicine and Public Health Pre-
(2006) Validation of an environmental distress scale. EcoHealth paredness 8(3):229–238. https://doi.org/10.1017/dmp.2014.45
3(4):245–254
Yen WM (1984) Effects of local item dependence on the fit and
Hogg TL, Stanley SK, O’Brien LV, Wilson MS, Watsford CR equating performance of the three-parameter logistic model.
(2021) The hogg eco-anxiety scale: development and validation Applied Psychological Measurement 8(2):125–145. https://
of a multidimensional scale. Global Environmental Change 71:1– doi.org/10.1177/014662168400800201
10. https://doi.org/10.1016/j.gloenvcha.2021.102391