Home and the extended-self: Exploring associations between clutter and wellbeing

Research into home and place attachment comes from a variety of disciplines, typically taking a multi-dimensional approach (Anton & Lawrence, 2004; Mallett, 2004; Scannell & Gifford, 2010). Within this scholarship, home is often seen as an aspect or an extension of one's self-identity, incorporating elements of thinking, feeling and doing at home (Mallett, 2004; Sigmon, Whitecomb, & Snyder, 2002). Accordingly, identity is seen as a predominant component that mediates the relationship between home and wellbeing (Leith, 2006). Within the home setting, identity (how we see and value ourselves both currently and ideally) (Rogers, 1959), is expressed in a variety of ways, such as the deliberate showcasing of possessions (Cristoforetti, Gennai, & Rodeschini, 2011; Jacobs & Malpas, 2013), the home modifications that people implement (Omar, Endut, & Saruwono, 2012) and the establishment of individualised home-cultures and ways of being (Cristoforetti et al., 2011; Leith, 2006; Mallett, 2004; Shenk, Kuwahara, & Zablotsky, 2004).

These manifestations of people's identities are evident to others. Observers entering other peoples' rooms or seeing of photographs of them can correctly infer the occupants' personality traits, as well as socio-demographic and socio-economic information (Gosling, 2009; Gosling, Ko, Mannarelli, & Morris, 2002; Pérez-López, Aragónés, & Amérigo, 2013).

Interestingly, within the literature lies an unspoken undertone or metaphor of home as a parent. Seamon (1979), for example, explains "at-homeness" (p. 69) as rootedness (grounding), appropriation (possessing), warmth (nurturing), at-easeness (unconditionally accepting) and regeneration (restoring). In pursuit of home, Ginsberg (1999) talks of womb, nest, and cradle (p. 83).

House is where we live, but home is how we live (Ginsberg, 1999). The concept of psychological home (Sigmon et al., 2002) encapsulates the connection between home and self-identity. Sigmon et al. (2002) conceptualised psychological home as a dynamic process (rather than a state) within which a person modifies, structures or preserves a home environment that reflects and communicates their sense of self. Accordingly, the concept includes: 1. a cognitive component: attributions about oneself in relation to the environment, 2. an affective component: emotions linked to the creation and preservation of psychological home (such as security, stability, warmth), and 3. behavioural components: the actions or investment of time and effort that a person may expend in order to express their identity, modify or maintain their home environment to meet their ideals of home. The concept therefore reflects the implicit understanding that investment of thinking, feeling and doing in the home environment creates a sense of psychological attachment to the home and contributes to wellbeing (Sigmon et al., 2002). These components are not mutually exclusive and indeed influence each other.

People who score highly on the psychological home scale invest in their homes and create a stronger sense of home as an extension of their self-identity. These actions have been found to positively correlate with wellbeing and positive psychological functioning (Cicognani, 2011; Roster et al., 2016; Sigmon et al., 2002).

Definitions of clutter are judgemental and often scornful. Roster et al. (2016) defined clutter as an "overabundance of material possessions that collectively create disorderly and chaotic home environments" (p. 32) – an objective definition that does not encompass individual, subjective experiences in distinguishing between desired possessions and clutter. It also does not encompass digital clutter (Belk, 2013). Belk, Seo, and Li (2007) used a metaphor of house as body and clutter as faeces, implying that clutter is (exclusively) waste. These judgemental overtones

assigning a negative valence to clutter are understandable as the word 'clutter' derives from associations with disarray and muddle (McCarthy, 2016). However, homes and relationships with home are as varied as those who occupy them. Perhaps one person's clutter is another's "blessing of abundance" (Löfgren, 2017, p. 2).

Clutter accumulation in the general populations is differentiated from hoarding which is a clinical condition (APA, 2013). Compared to 'normative collecting', which is structured, considered, and pleasurable, hoarding is unstructured, unpremeditated, and distressing (Nordstetten, de la Cruz, Billotti, & Mataix-Cols, 2013; Stumpf et al., 2018). The acts of thinking, feeling and doing at home are obstructed and confined at the most basic level, as homes become full to the point of debilitating daily practices like sleeping, cooking and hygiene, and can become a safety risk (Frost & Steketee, 2011).

Assessing the tipping point that differentiates normative clutter from hoarding disorder is complex (Potts, 2015). One indicator is the Clutter Image Rating Scale (Frost, Steketee, Tolin, & Renaud, 2008). This pictorial tool showing series of the same rooms with increasing levels of clutter, allocates scores of 1 (uncluttered) to 9 (full) for each room. A mean score of 4 or higher is a recommended cut-off indicating "clutter requiring clinical attention" (Frost et al., 2008, p. 201).

Within the hoarding literature the impact of clutter on mental health, quality of life and other domains of life is evident. Comorbidities of hoarding include anxiety and depression (An et al., 2009), risk to personal and psychological safety, self-neglect, increased financial burden, and not sustaining employment (Holmes, Whomsley, & Kellett, 2015; Tolin, Frost, Steketee, Gray, & Fitch, 2008). Wider burden or cost falls on families, neighbours, local services and communities (Roane, Landers, Sherratt, & Wilson, 2017). Identified factors within hoarding disorder are degrees of disinhibition when acquiring possessions, increased attachment to and anthropomorphising of possessions, procrastination, perfectionism and indecision (Dozier & Ayers, 2017; Kings, Moulding, & Knight, 2017).

Clutter research in non-clinical populations is limited, generally focussing on its detriment to quality of life and wellbeing. This body of research suggests that the impact of clutter on non-clinical populations has parallel upshots to those with hoarding disorder, though it's less severe. Evidence of the prevalence of clutter in non-clinical populations includes Schor's (1993) study showing that Americans own twice as many possessions compared to 25 years previous, and Arnold and Lang's (2007) study finding Los Angeles' residents to be commonly storing seldom-used possessions in garages, leaving no room for cars, alongside exacerbated stress and debt.

In samples mainly consisting of students combined with small community samples procrastination is proposed as a causal, predictive factor of clutter accumulation and vice versa (Ferrari & Roster, 2018; Ferrari, Roster, Crum, & Pardo, 2018). Indecision inhibits possession disposal and is connected with lower psychological home and wellbeing (Ferrari, Crum, & Pardo, 2018). In another study the researchers found that undergraduates make less healthy eating choices in 'chaotic' kitchens compared to 'standard' kitchens (Vartanian, Kernan, & Wansink, 2017). Diaries, home-tour recordings and cortisol measures reveal more depressed language and mood when homes are experienced as cluttered (Saxbe & Repetti, 2010).

Children from comparatively environmentally chaotic homes perform cognitive tasks less well and demonstrate behavioural challenges (Dumas et al., 2005; Goux & Maurin, 2005). Lower reading accomplishment was also found in adolescents from homes gauged as physically problematic (which includes, but is not limited to clutter), and this was supported by evidence of thinner prefrontal cortices in brain scans (Uy, Goldenberg, Tashjian, Do, & Galvan, 2019).

Roster et al. (2016) challenge the 'home is where the heart is' literature and reported that clutter negatively impacts psychological home and wellbeing. The researchers however recognised that their 1394 respondents, were subscribers to the Institute for Challenging Disorganisation, and therefore identified as individuals with "clutter issues"

(Roster et al., 2016, p. 39) leaving readers questioning what associations might exist in a general population. Further research comparing Roster et al.'s (2016) data with younger samples (not subscribers of clutter help websites), found that clutter affects the wellbeing of older populations more than younger participants (Ferrari & Roster, 2018). It is unclear whether this relationship is due to age or sampling bias. Further analysis of the same dataset revealed cultural differences in clutter assessments and impact (Crum & Ferrari, 2019), suggesting that further scrutiny into the subjective-objective clutter experience is warranted. Roster et al. (2016) recommend the use of objective clutter measures to advance scholars understanding of the complex relationships between clutter and wellbeing, which can inform strategies for clutter reduction.

In non-clinical populations, every aspect of Season's (1979) "at-homeness" (p. 69) becomes compromised when the habit of accumulating possessions expressing self-identity results in the overwhelm described above. Do those who choose and make effort to live with a subjectively and/or objectively manageable clutter experience enhanced wellbeing? If so, what aspects of wellbeing are affected? One way to inform clutter reduction strategies is to examine its association with wellbeing in a more nuanced way.

Cherrier and Belk (2015) describe decluttering as the process of creating orderly environments by considering, organising and discarding possessions. In hoarding, decluttering is an intervention delivered alongside clinical help (Holmes et al., 2015). In non-clinical populations decluttering can be done single-handedly or with assistance of a professional organiser (Belk et al., 2007; Roster, 2015). Decluttering is a skill and can be learned. In one study participants at risk of hoarding were trained to 'reduce, organise, review and maintain' (Aso, Yamaoka, Nemoto, Nagamura, & Saito, 2017, p. 5). The follow up measurement seven months after the training revealed that participants continued to declutter and manifested higher self-esteem than before the trial (Aso, 2018; Aso et al., 2017). Psychological home and wellbeing were not measured but Aso (2018) concludes that decluttering training can "improve quality of life" (p. 7).

Some evidence of decluttering's positive effect emerged from Thacher, Onywer and Tuthill's (2017) paper in which the result of hierarchical regression revealed that decluttering predicted improved sleep in subscribers to a house-keeping website. Additional evidence emerged from a phenomenological study (Lee, 2017) where participants described their decluttering process as "uplifting and refreshing", reporting "a happier, less stressful life" (p. 455–456). Further qualitative research within non-clinical populations provided insight that individual relationships with clutter vary considerably, as do conceptions of what is acceptable, unacceptable or organised (Dion et al., 2014; Swan, Taylor, & Harper, 2008). These variations are informed by how the construct of possession attachment and possession meaning are conceptualised (Belk, 1988; Kleine, Kleine, & Allen, 1995).

Since the existing research has come from the position of questioning clutter's negative effect, measures of wellbeing in studies to date have typically measured wellbeing as the lack of or low symptoms of depression or anxiety (ie., An, et al., 2009), or made use of Diener, Emmons, Larsen and Griffin's (1985) Satisfaction With Life Scale as a means of approximating wellbeing (ie., Crum & Ferrari, 2019; Ferrari & Roster, 2018; Roster et al., 2016).

Positive Psychology as the science of wellbeing offers a growing body of research which claims that wellbeing is to some degree in people's control, and that taking specific actions and developing constructive habits can promote wellbeing (Sin & Lyubomirsky, 2009). While discussions continue around whether wellbeing is a state or a set of strategies, and what are its key components, positive psychology offers a

variety of wellbeing theoretical models and measures which incorporate hedonic (feeling good) and eudemonic (functioning well) elements. Importantly, these measures indicate that wellbeing is more than the absence of disorder and therefore aim to directly measure the components of wellbeing. Seligman's (2011) PERMA model for example categorises wellbeing as a construct that has 5 routes to wellbeing: Positive emotions (joy, happiness), Engagement (interest, involvement and flow), Relationships (positive connections with others, communities and societies), Meaning in life and Accomplishment (having meaningful goals and working to achieve them).

The present study aimed to use a cross-sectional correlational design to explore the associations between home self-extension factors (subjective clutter, objective clutter, psychological home, and decluttering habits) and wellbeing in a general population sample. The hypothesis was that home clutter, decluttering habits and psychological home will be strong predictors of participants' wellbeing.

Convenience and snowball sampling were used to invite participants. A short YouTube film, placed on social media platforms (LinkedIn, Twitter and Facebook), welcomed participation. The only participant inclusion criterion was a requirement to be 18 or over. The first author shared the YouTube film with all her networks (i.e., neighbourhood, community groups and clubs, university and academic circles). Members of the UK Association of Professional Declutters and Organisers, of which the first author is a member, were also invited to participate and share the YouTube film with their personal networks. The only reward offered was the opportunity to contribute to knowledge.

Of 1411 responses, 246 did not provide consent or missed excessive data and were eliminated. Eighteen cases were assessed as extreme outliers (see below) and removed. To report confidence intervals and perform regressions, 36 cases who did not respond to all four independent variables were also removed. Of the final 1111 respondents most were female (90.6%, $n\; =\; 1007$) and educated beyond high school (88.8% $n\; =\; 980$). The majority were UK based (74.1% $n\; =\; 822$). The rest were based in the US ($n\; =\; 201$), other parts of Europe ($n\; =\; 39$), Australasia ($n\; =\; 28$), Africa ($n\; =\; 11$) and Asia ($n\; =\; 8$). Age ranged from 19 to 79, averaging at 48.

A cross-sectional correlational design survey was used and made available online through the Qualtrics platform in summer 2019. Ethical approval was granted by the authors' university ethics board. Pilot testing generated feedback, resulting in edits augmenting clarity and ease of completion and indicated 7–10 min completion time. The general tone of the survey was both friendly and thankful. The only forced answers were consent form and age eligibility.

The PERMA-profiler (Butler & Kern, 2016) is a 23-item scale, ranging on a 0–10 Likert scale (where 0 is not at all and 10 is completely). The scale has 8 subscales: the 5 PERMA components: positive emotion, engagement, relationships, meaning and accomplishment, and in addition: a health subscale, negative emotions, and loneliness. The PERMA-profiler provides a multi-dimensional measure of wellbeing and the developers Butler and Kern (2016) who tested it with over 30,000 individuals worldwide report a strong convergent and divergent validity and good reliability. In the current study PERMA overall wellbeing had a Cronbach's alpha of 0.94. Sample items include: In general, how often do you feel joyful? (Positive Emotion). How often do you become absorbed in what you are doing? (Engagement). To what extent do you

feel loved? (Relationships). In general, to what extent do you lead a purposeful and meaningful life? (Meaning). How much of the time do you feel you are making progress towards accomplishing your goals? (Accomplishment). In general, how often do you feel anxious? (Negative Emotion). To view the entire scale see Butler and Kern (2016).

The Clutter Image Rating Scale (CIRS) (Frost et al., 2008) measures objective clutter levels. Three sets of 9 images show identical rooms (a kitchen, bedroom and bathroom) becoming progressively fuller. Scores, calculated as a mean, range from 1 for the first images showing little or no clutter to 9 for completely full rooms. The scale's pictorial property eliminates subjective experiences of clutter; participants choose which pictures most represent their home. Low scores indicated less objective clutter. A cut off point of 4.0 and above indicates significant clutter tendencies that require clinical attention. In the sections that follow we drew on this cut-off point to differentiate between 'clinical population' (people with clutter levels that require clinical support) and 'non-clinical population'. CIRS has demonstrated consistency and reliability with Cronbach alphas above 0.8 across age groups (Dozier & Ayers, 2015). This study returned a Cronbach's alpha of 0.81. To view the images on The Clutter Image Rating Scale see the CIRS app (Boston University, 2015).

The Clutter Quality of Life Scale (CQLS) (ICD, n.d.); an 18-item scored on a 1–7 Likert scale (where 1 is strongly disagree and 7 is strongly agree). It measures the subjective experience of clutter at home, regardless of the objective level. Scores are calculated as a total of the Likert values; lower scores (18–53) indicated a positive subjective experience of clutter, medium scores (54–89) indicate some impact of clutter on quality of life, and higher scores (90 and above) indicated negative subjective experience. The scale is undergoing validation by the authors. This study returned a Cronbach's alpha of 0.96. Sample items include: I'm concerned about what others might think of me if they knew about the clutter in my home; I have to move things in order to accomplish tasks in my home; I often buy things I already have because I don't know where things are in my home; The clutter in my home upsets me; and I avoid having people come to my home because of the clutter (see https://www.challengingdisorganization.org/clutter-quality-of-life-scale-cqls-for-pos for details of all scale items).

The Psychological Home Scale (PH) (Sigmon et al., 2002); an 8-item scored on a Likert scale between 1 and 7 (where 1 is strongly disagree and 7 is strongly agree). It measures psychological home. Scores were calculated as a total of the Likert values with high scores indicating high psychological home. This study returned a Cronbach's alpha of 0.84. Sample items include: I have grown attached to many of the places I have lived; I put a lot of time and effort into making a place my own; I feel more relaxed when I'm at home; I surround myself with things that highlight my personality; and I get a sense of security from having a place of my own (see Sigmon et al., 2002 for details of all scale items).

Demographic questions, including a question on decluttering habit were presented at the end of the questionnaire. This item was coded according to frequency of decluttering (0- never – 6- daily). High scores indicated more frequent decluttering.

The similarity of the Likert ratings within individual questionnaires eased cognitive load, creating consistency. Order effects were avoided by setting the scales to appear in random order. To prevent potential response tendency bias (Barnette, 2000; Schriesheim & Hill, 1981) the scales differed in appearance with some requiring participants to tick a box, and others requiring selecting an image. Additionally, reversed items are included in PERMA, and the home self-extension variables varied in the meaning of the high Likert values: in the psychological home scale ascribing high Likert values reflects a positive state, while in the subjective clutter scale high Likert values reflect a negative state.

In homage to Belk's (1988) seminal paper on possessions being 'extensions of the self', we used the term home self-extension to encapsulate The Clutter Image Rating Scale (CIRS), hereafter referred to as objective clutter, The Clutter Quality of Life Scale (CQLS), hereafter referred to as subjective clutter, The Psychological Home Scale (PH), hereafter referred

to as psychological home) and Decluttering Habit (DH), representing the frequency of decluttering).

To encourage a general population over those with clutter problems, the study was publicised by welcoming participants with 'no clutter, loads of clutter, or anywhere in between'. A briefing page described the study and potential involvement. Confidentiality and anonymity were assured. Procedure for data withdrawal was explained as well as treatment of data at closure.

Those opting to participate confirmed age eligibility and signed a consent form. Having submitted survey responses participants were thanked, provided a debrief, signposting to further information on clutter, hoarding and help. The researchers received no withdrawal requests.

The data was analysed using IBM SPSS Statistics 26. Cases missing excessive data were deleted ($n\; =\; 249$). No mean substitutions were made for missing data as this has been found to elevate correlational results (Parent, 2013).

Since the intention of the study was to examine a non-clinical sample, all extreme outliers were removed. For this purpose, boxplots were produced which indicated 18 extreme outliers (15 Objective clutter >5.5, 1 psychological home <1.3, 2 PERMA <1.1). These cases were deleted. The deletions also left us more confident of remaining participants representing a general population (Osborne, 2010). Reliability levels were recalculated indicating that all Cronbach alpha levels remain above 0.69, indicating good reliability for further analysis (Field, 2018).

Normality tests following the deletions of outliers indicated that all variables were normally distributed, which justified parametric testing (Field, 2018).

First stage of analysis was Pearson's correlations testing levels of shared variance between home self-extension variables. Next, nine multiple linear regressions were conducted to ascertain the degree to which the independent home self-extension variables accounted for variability in the PERMA wellbeing measure and its subscales. Analyses were conducted to ensure satisfaction of the assumptions of normality, linearity and homoscedasticity. The possibility of multicollinearity was also assessed, and since all tolerance values were greater than 0.1, all values for VIF were lower than 10, it was considered that multicollinearity was not likely to be a problem.

Note: $N\; =\; 1111$. DH = Decluttering Habit reflects the frequency of decluttering: 0 = never to 6 = daily.

The model with the four home self-extension predictors (objective clutter, subjective clutter, psychological home, decluttering habit) reached significance, accounting for 22.7% of the variance in positive emotion ($R^2\; =\; 0.227$, $F(4,1106)\; =\; 82.09$, ). Table 3 shows significant contributions at from subjective clutter ($\backslash beta\; =\; -0.37$) and psychological home ($\backslash beta\; =\; 0.20$) respectively indicating that positive emotion was predicted by subjective clutter and psychological home. Neither objective clutter ($\backslash beta\; =\; 0.02$, $p\; =\; .684$) nor decluttering habit ($\backslash beta\; =\; 0.02$, $p\; =\; .450$) made a significant contribution.

The model reached significance, accounting for 12.9% of the variance in engagement ($R^2\; =\; 0.129$, $F(4,1106)\; =\; 42.22$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; -0.25$), psychological home ($\backslash beta\; =\; 0.25$) and objective clutter at ($\backslash beta\; =\; 0.08$) suggest engagement was predicted by subjective clutter and psychological home, with objective clutter making a relatively smaller, unexpected contribution. Decluttering habit ($\backslash beta\; =\; -0.03$, $p\; =\; .311$) made no significant contribution.

The model reached significance, accounting for 18.6% of the variance in relationships ($R^2\; =\; 0.186$, $F(4,1106)\; =\; 63.24$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; -0.30$) and psychological home ($\backslash beta\; =\; 0.22$) indicate relationships was predicted by subjective clutter and psychological home respectively. Neither objective clutter ($\backslash beta\; =\; -0.01$, $p\; =\; .796$) nor decluttering habit ($\backslash beta\; =\; 0.03$, $p\; =\; .246$) made a significant contribution.

The model reached significance, accounting for 19.9% of the

Summary of enter method multiple regression analyses for home self-extension predictors of PERMA elements and overall wellbeing.

Note: N = 1111. B=Unstandardized coefficients. SE=Standard error. $\backslash beta$=Standardized coefficients. *, **, ***. CI = 95% Confidence interval for B, LB = Lower bound, UB=Upper bound.

variance in meaning ($R^2\; =\; 0.199$, $F(4,1106)\; =\; 68.59$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; -0.31$) and psychological home ($\backslash beta\; =\; 0.25$) suggest meaning was predicted by subjective clutter and psychological home respectively. Neither objective clutter ($\backslash beta\; =\; 0.04$, $p\; =\; .232$) nor decluttering habit ($\backslash beta\; =\; 0.02$, $p\; =\; .548$) made a significant contribution.

The model reached significance, accounting for 29.3% of the variance in accomplishment ($R^2\; =\; 0.293$, $F(4,1106)\; =\; 114.61$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; -0.41$), psychological home ($\backslash beta\; =\; 0.20$) and decluttering habit at ($\backslash beta\; =\; 0.06$) suggest accomplishment was predicted by subjective clutter and

psychological home respectively with a smaller contribution from decluttering habit. Objective clutter ($\backslash beta\; =\; -0.01$, $p\; =\; .751$) made no significant contribution.

The model reached significance, accounting for 14.4% of the variance in negative emotion ($R^2\; =\; 0.144$, $F(4,1106)\; =\; 46.67$, ). The only significant contribution was at from subjective clutter ($\backslash beta\; =\; 0.39$), the positive $\backslash beta$ in this instance indicating higher subjective clutter scores predict negative emotion. Psychological home ($\backslash beta\; =\; -0.02$, $p\; =\; .553$), objective clutter ($\backslash beta\; =\; -0.04$, $p\; =\; .327$) and decluttering habit ($\backslash beta\; =\; -0.02$, $p\; =\; .561$) made no significant contribution.

The model reached significance, accounting for 14.7% of the variance in health ($R^2\; =\; 0.147$, $F(4,1106)\; =\; 47.64$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; -0.32$) and at from psychological home ($\backslash beta\; =\; 0.05$) indicate health was predicted by subjective clutter and psychological home respectively. Neither objective clutter ($\backslash beta\; =\; -0.06$, $p\; =\; .096$) nor decluttering habit ($\backslash beta\; =\; -0.01$, $p\; =\; .661$) made a significant contribution.

The model reached significance, accounting for 10.2% of the variance in loneliness ($R^2\; =\; 0.102$, $F(4,1106)\; =\; 31.29$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; 0.33$) and at from psychological home ($\backslash beta\; =\; -0.07$) suggest loneliness was predicted by subjective clutter and psychological home respectively. Neither objective clutter ($\backslash beta\; =\; -0.06$, $p\; =\; .160$) nor decluttering habit ($\backslash beta\; =\; 0.04$, $p\; =\; .156$) made a significant contribution.

The model reached significance, accounting for 22.9% of the variance in overall wellbeing ($R^2\; =\; 0.229$, $F(4,1106)\; =\; 82.09$, ). Significant contributions at from subjective clutter ($\backslash beta\; =\; -0.37$) and psychological home ($\backslash beta\; =\; 0.20$) indicate overall wellbeing was predicted by subjective clutter and psychological home respectively. Neither objective clutter ($\backslash beta\; =\; 0.02$, $p\; =\; .684$) nor decluttering habit ($\backslash beta\; =\; 0.02$, $p\; =\; .450$) made a significant contribution.

To summarise, the results show that home self-extension variables explained a substantial amount of the variance in all 8 elements of the PERMA profiler as well as overall wellbeing. Subjective clutter is the strongest contender: lower scores in subjective clutter (positive subjective experience of clutter) predicted higher scores in positive emotion, engagement, relationships, meaning, accomplishment, overall wellbeing and health. Subjective clutter positively predicted negative emotion and loneliness, but less strongly than the other elements. In contrast, objective clutter only made a significant (and unexpectedly positive) contribution to the engagement subscale. Psychological home significantly accounted for variance to all the elements except negative emotions. Decluttering habit only significantly accounted for variance to accomplishment.

The hypothesis that combined home self-extension variables predict wellbeing was largely supported by the results. Combined variance accounted for by home self-extension predictors is strongest for accomplishment, followed respectively by positive emotion, meaning, relationships, health, negative emotion, engagement, and loneliness.

The study findings contribute to theories of home and psychological home, bolstering Belk's (1988) possession self-extension theory. The key purpose of psychological home is to promote a greater sense of wellbeing (Sigmon et al., 2002) and our findings show that subjective clutter and psychological home make a considerable contribution to wellbeing. Drawing on these findings, we propose that the connection between psychological home and possession self-extension may not be wholly due to the sinister valence typically attributed to clutter, but rather to the self-extension of possessions explained by Belk (1988).

Possessions housed in homes have meanings and biographies that deny and decry their material nature and value. However, they also have a temporal nature (Roster, 2015) whereby the subjective relationships with possessions alters and transforms over time. We suggest that the meaning assigned to possessions and clutter and people's capacity to express themselves through their home environment, impinges on wellbeing through the ability to strike a balance between subjective and objective clutter, and to notice the point at which possessions convert from a treasured items to clutter, and act accordingly. These repeated actions continually curate home as an expression of self-identity; "home acquires its meaning through mundane, daily practices" (Dion et al., 2014, p. 585).

We therefore gently challenge extant theories that clutter is maladaptive, drawing attention to the subjective experience of items being in or out of place at home, aligning (or not) with self-identity; a factor overlooked in clutter definitions to date. We propose the following refined definition of clutter: "Clutter is a subjective experience of possessions (material or other) that inhibit the curation of self-identity at home."

Our findings have practical use for professional organisers, and for individuals wanting to tackle clutter that is detracting from their quality of life. Given the emphasis on subjectivity, this study highlights the

importance of non-judgemental, person-centred approaches from professional organisers. Our results underline the importance of cultivating understanding of client values, and familiarising themselves with household culture and norms, before embarking on clutter reduction strategies (see Roster, 2015 for an excellent summary of person-possession disposal strategies).

Given that the wellbeing accomplishment factor was most strongly predicted by home self-extension dimensions, this indicates that accomplishment interventions could be useful within clutter reduction strategies. Establishing clear, achievable person-specific goals to create home self-identity can galvanise those struggling with clutter to tackle it.

Recognising the important part positive emotions play in home self-extension can help organisers motivate clients to undertake what might feel like a mammoth task. A first step with an overwhelmed client might be to suggest interventions that promote positive emotion (see Sin & Lyubomirsky, 2009) improving resilience to confront the task.

Our sample, although sufficiently large to return statistically significant results, was predominantly female, restricting opportunity to explore gender differences. This research stands on the shoulders of others with mostly or all female participants (Dion et al., 2014; Lee, 2017; Roster et al., 2016), leading us to join their call for future research into clutter, home and gender. Additionally, the correlations and regressions reported in this study may be sample specific; our convenience and snowball sampling restrict the ability to generalise from this sample to larger populations (Field, 2018), and since no race or ethnic origin data was collected, comparisons cannot be made. This oversight is disappointing due to indications that wellbeing and clutter are experienced differently in different cultures (Birch et al., 2019; Crum & Ferrari, 2019).

Our 'decluttering habit' variable was a single item, untested for reliability or validity. We also noted that some participants may have found the response values confusing, meaning that inferences from this item could be unreliable. Therefore, its small role in our findings – low to moderate associations with PERMA elements must be interpreted cautiously.

The study's intention to prevent potential response tendency bias could be improved by adding scales that assess rating bias (i.e., social desirability and/or acquiescent responding). If controlling for these biases had no significant impact on statistical analysis, it would have reinforced the study's findings (Kreitchmann, Abad, Ponsoda, Nieto, & Morillo, 2019).

A fruitful area for further research might be to focus on decluttering interventions: to define it as an action orientated intervention and test its impact on wellbeing in an experimental research design. We suggest that within decluttering coaching or training, those overwhelmed by clutter may benefit from understanding the extended-self theory. For instance, Belk (1988) talks of 'contagion' – possessions permeating essence of their owners – which is why rituals to cleanse and neutralise can be beneficial in dispossession. Understanding the normalcy of seemingly illogical feelings about possessions could open the mind to the temporal nature of possession self-extension. Articulating whether a possession is 'still me' can be challenging, but the triumph and relief in doing so is liberating (Lastovicka & Fernandez, 2005; Lee, 2017).