Estimating hoarding disorder prevalence using the Network Scale-Up Method

class: center, middle, inverse, title-slide

.title[
# Estimating hoarding disorder prevalence using the Network Scale-Up Method
]
.author[
### Prof. Thomas Pollet, Northumbria University (<a href="mailto:thomas.pollet@northumbria.ac.uk" class="email">thomas.pollet@northumbria.ac.uk</a>)
]
.date[
### 2026-04-27 | <a href="http://tvpollet.github.io/disclaimer">disclaimer</a>
]

---

<div>
<style type="text/css">.xaringan-extra-logo {
width: 110px;
height: 128px;
z-index: 0;
background-image: url(NU2.png);
background-size: contain;
background-repeat: no-repeat;
position: absolute;
top:1em;right:1em;
}
</style>
<script>(function () {
  let tries = 0
  function addLogo () {
    if (typeof slideshow === 'undefined') {
      tries += 1
      if (tries < 10) {
        setTimeout(addLogo, 100)
      }
    } else {
      document.querySelectorAll('.remark-slide-content:not(.title-slide):not(.inverse):not(.hide_logo)')
        .forEach(function (slide) {
          const logo = document.createElement('div')
          logo.classList = 'xaringan-extra-logo'
          logo.href = null
          slide.appendChild(logo)
        })
    }
  }
  document.addEventListener('DOMContentLoaded', addLogo)
})()</script>
</div>

## Outline

* The problem: How prevalent is hoarding disorder in the UK?

* Why existing approaches fall short

* The Network Scale-Up Method (NSUM)

* Why hoarding is a good NSUM candidate

* Proposed study design

* Funding and collaboration

---
## Hoarding disorder: A brief overview

* Recognised as a standalone diagnosis in DSM-5 <a name=cite-APA2013></a>([American Psychiatric Association, 2013](#bib-APA2013)), separated from OCD <a name=cite-Mataix-Cols2010></a><a name=cite-Pertusa2010></a>([Mataix-Cols, Frost, Pertusa, Clark, Saxena, Leckman, Stein, Matsunaga, and Wilhelm, 2010](https://doi.org/10.1002/da.20693); [Pertusa, Frost, Fullana, Samuels, Steketee, Tolin, Saxena, Leckman, and Mataix-Cols, 2010](https://doi.org/10.1016/j.cpr.2010.01.007))

* Persistent difficulty discarding possessions, resulting in clutter that compromises the intended use of living spaces <a name=cite-Frost1996></a><a name=cite-Steketee2003></a>([Frost and Hartl, 1996](#bib-Frost1996); [Steketee and Frost, 2003](https://doi.org/10.1016/j.cpr.2003.08.002))

* Associated with significant "problems". For example: fire risk, falls, pest infestations, social isolation, self-neglect <a name=cite-Tolin2008></a><a name=cite-Frost2004></a>([Tolin, Frost, Steketee, Gray, and Fitch, 2008](https://doi.org/10.1016/j.psychres.2007.08.008); [Frost, Steketee, and Williams, 2000](https://doi.org/10.1046/j.1365-2524.2000.00245.x)).

* Tends to increase with age, with onset typically in adolescence -- but clinical severity peaking in older adulthood <a name=cite-Cath2017></a><a name=cite-Samuels2008></a>([Cath, Nizar, Boomsma, and Mathews, 2017](https://doi.org/10.1016/j.jagp.2016.11.006); [Samuels, Bienvenu, Grados, Cullen, Riddle, Liang, Eaton, and Nestadt, 2008](https://doi.org/10.1016/j.brat.2008.04.004))

---
## The prevalence problem

* Existing estimates range from **2-6%** of the adult population <a name=cite-Postlethwaite2019></a>([Postlethwaite, Kellett, and Mataix-Cols, 2019](https://doi.org/10.1016/j.jad.2019.06.004))

* Most estimates derive from self-report screening instruments: the Savings Inventory-Revised, the Clutter Image Rating <a name=cite-Frost2008SIR></a>([Frost, Steketee, Tolin, and Renaud, 2008](https://doi.org/10.1007/s10862-007-9068-7))

* But: people with hoarding problems often have **limited insight** into the severity of their situation <a name=cite-Frost2011></a>([Steketee and Frost, 2003](https://doi.org/10.1016/j.cpr.2003.08.002); [Frost, Steketee, and Tolin, 2011](https://doi.org/10.1002/da.20861)) → underreporting in surveys!

* **No reliable UK-specific prevalence estimate exists.** The most-cited UK figure comes from a twin study <a name=cite-Iervolino2009></a>([Iervolino, Perroud, Fullana, Guipponi, Cherkas, Collier, and Mataix-Cols, 2009](https://doi.org/10.1176/appi.ajp.2009.08121789)) using a non-clinical screening measure.

???
* Meta-analytic estimate of around 2%
* Opportunity sample -- 2% prevalence - found greater prevalence in men than women.

---
## Existing data records...

* **No register of people who hoard.** No administrative dataset captures this population.

* **GP records**: Patients rarely present hoarding as their primary complaint.

* **Mental health services**: Only see the tip of the iceberg? -- people referred (usually by others) after crisis.

* **Fire services / Environmental health**: Capture only severe cases that generate fire risk assessments, complaints, or actual fires?

* **Housing associations**: Miss owner-occupiers and private renters.

→ We are relying on self-report from a population with limited insight, or counting only the most severe cases that come to professional attention. → problem!

<img src="https://media2.giphy.com/media/v1.Y2lkPTc5MGI3NjExYmRsOXI0dGFsNjYxemJyaWN2YjI5Mmh2Z3ZoNnN3dTJ3MTU3dnA1NyZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/3o6Mb3suz7pLr4zaAo/giphy.gif" alt="" width="200px" style="display: block; margin: auto;" />
---
## What is the Network Scale-Up Method?

A survey-based technique for estimating the size of hard-to-count populations <a name=cite-Killworth1998a></a><a name=cite-McCarty2001></a>([Killworth, Johnsen, McCarty, Shelley, and Bernard, 1998](#bib-Killworth1998a); [McCarty, Killworth, Bernard, Johnsen, and Shelley, 2001](https://doi.org/10.17730/humo.60.1.efx5t9gjtgmga73y)): e.g., HIV+ population, Intravenous drug users, etc.

**Core logic**: Ask a representative sample of the general population _"How many people do you know who [have characteristic X]?"_

**Step 1: Estimate personal network size**: Use responses to questions about populations of known size ("How many Nicks do you know?", "How many nurses do you know?",... .)

`$$\hat{d}_i = \frac{\sum_{k=1}^{K} y_{ik}}{\sum_{k=1}^{K} N_k} \times N$$`

where `$y_{ik}$` = respondent `$i$`'s count for known subpopulation `$k$` (e.g., Davids, nurses), `$N_k$` = true size of that subpopulation, `$K$` = number of calibration items, and `$N$` = total population.

???

---
## Network scale up method (continued... )

**Step 2: Estimate hidden population size**: Scale up responses about the target population using these estimated network sizes.

`$$\hat{N}_h = \frac{\sum_{i=1}^{n} y_{ih}}{\sum_{i=1}^{n} \hat{d}_i} \times N$$`

where `$y_{ih}$` = respondent `$i$`'s count of people in hidden population `$h$` (e.g., people who hoard).

---
## NSUM: Key developments

* **Overdispersed models** accounting for non-random mixing between respondents and subpopulations <a name=cite-Zheng2006></a>([Zheng, Salganik, and Gelman, 2006](https://doi.org/10.1198/016214505000001168))

* **Bayesian estimation** with barrier and transmission effects <a name=cite-Maltiel2015></a><a name=cite-Josephs2024></a>([Maltiel, Raftery, McCormick, and Baraff, 2015](https://doi.org/10.1214/15-AOAS827); [Josephs, Feehan, and Crawford, 2024](https://doi.org/10.1177/00491241221122576))

* **Generalised NSUM estimator** correcting for visibility (not everyone in the hidden population is visible to their contacts) <a name=cite-Feehan2016></a>([Feehan and Salganik, 2016](https://doi.org/10.1177/0081175016665425))

* **Correlated models** capturing dependencies between subpopulations <a name=cite-Laga2021></a>([Laga, Bao, and Niu, 2023](https://doi.org/10.1080/01621459.2023.2165929))

* Applied extensively in the US, Iran, Brazil, China, Rwanda, and recently the Netherlands <a name=cite-Volker2025></a>([Völker, Hofstra, Corten, and van
Tubergen, 2025](https://doi.org/10.1016/j.socnet.2025.05.007)) and Spain <a name=cite-Lubbers2019></a><a name=cite-Lubbers2025></a>([Lubbers, Molina, and Valenzuela-García, 2019](https://doi.org/10.1016/j.socnet.2018.08.004); [Lubbers, Bojanowski, Targarona Rifa, and Ciordia, 2025](https://doi.org/10.1177/00811750251340398))

* **Never applied in the UK.** (At least that is what my googling tells me... .)

---
## Why hoarding is a good NSUM candidate

**1. The behaviour is physically observable** <a name=cite-Snowdon2012></a>([Snowdon, Pertusa, and Mataix-Cols, 2012](https://doi.org/10.1002/da.21943))

Hoarding manifests in the physical environment. Acquaintances who have visited someone's home can _see_ it. (even better than other hidden populations, e.g., intravenous drug use).

→ Even the _avoidance_ of home visits (never being invited in) is a social signal.

**2. We can have a definition which captures the problem.**

e.g.: _"How many people do you know whose home is so full of clutter and possessions that it seriously interferes with their ability to use their rooms normally — for example, they can't cook in their kitchen, sleep in their bedroom, or have visitors because of the amount of stuff?"_

→ Maps onto DSM-5 functional impairment criterion ([American Psychiatric Association, 2013](#bib-APA2013)) without requiring respondents to make a clinical diagnosis.

---
## Why hoarding is a good NSUM candidate (cont.)

**3. Genuinely hidden from administrative systems**

No register, underdiagnosed, fragmented across fire, health, housing, and social care services. Existing self-report measures are undermined by the insight problem.

**4. Asking _others_ sidesteps the insight problem**

NSUM asks acquaintances who have observed the living situation — not the person themselves. This bypasses the core limitation of self-report prevalence surveys for hoarding.

**5. Age concentration may aid visibility**

Hoarding is concentrated in older adults ([Cath, Nizar, Boomsma et al., 2017](https://doi.org/10.1016/j.jagp.2016.11.006)). who tend to have stable, long-standing networks — meaning their contacts have had more opportunity to observe their living situation over time. This may improve the visibility factor compared to hidden populations concentrated among more mobile demographics. **But**: long-standing avoidance of home visits works against this.

---
## Proposed design: Overview

A **two-sample design**:

**Sample 1: General population survey** (n ≈ 1,000--2,000)

* Nationally representative online panel (YouGov or equivalent)
* ~25 NSUM calibration items (names + group memberships) to estimate personal network size
* Target question(s) on hoarding
* Demographics

**Sample 2: Supplementary clinical/community sample** (n ≈ 100--150)

* People with known hoarding problems recruited via fire services, environmental health, NHS community mental health teams, hoarding support groups, social workers, housing associations,... ?
* Purpose: estimate the **visibility factor**: what proportion of a hoarder's contacts are aware of the problem? ([Feehan and Salganik, 2016](https://doi.org/10.1177/0081175016665425))

---
## Sample 1: Calibration items

**~12-15 name-based items** calibrated from ONS birth registration data + life tables to estimate living population counts <a name=cite-ClayWarner2022></a><a name=cite-Fenoy2024></a>([Clay-Warner, Kawashima, and Edgemon, 2022](https://doi.org/10.2105/AJPH.2022.306731); [Fenoy, Bojanowski, and Lubbers, 2024](https://doi.org/10.1177/1525822X241243115))

**~10-12 group-membership items** using UK administrative registers (e.g., registered nurses via NMC, police officers via Home Office, postal workers via Royal Mail, people who gave birth in the last 12 months)

**UK-specific calibration challenge**: No living-population-by-first-name database equivalent to the US Social Security Administration data. Requires demographic reconstruction from ONS records.

→ Plus the **name abbreviation problem** (Elizabeth → Liz, William → Bill/Will): a UK-specific issue essentially unaddressed in the existing literature.

---
## Sample 2: The visibility factor

**Two-part question** for the supplementary sample:

_"Of the people who know you well enough to contact you, roughly how many have been inside your home in the past two years?"_

_"Of those who have been inside your home, how many do you think are aware that the clutter is a significant problem?"_

**Access rate** × **Recognition rate** = **Visibility factor**

This captures the _avoidance_ pathway too: if most contacts have not been inside the home, this is informative about access restriction --> even if those contacts suspect something is wrong.

???
the multiplicative formulation assumes conditional independence between access and recognition — i.e., that whether someone recognises hoarding given they've been inside doesn't depend on the overall access rate. This may not hold: contacts who visit more frequently may also be closer to the person and more likely to notice the problem, and severity simultaneously restricts access while making the problem more obvious to those who do get in. An alternative is to estimate visibility as a single direct question, avoiding the decomposition entirely but losing the informative separation of structural (access) and perceptual (recognition) components.

---
## Validation strategy

**Primary: Internal validation (back-estimation)** <a name=cite-Habecker2015></a>([Habecker, Dombrowski, and Khan, 2015](https://doi.org/10.1371/journal.pone.0143406))

Hold out each known subpopulation in turn, treat as unknown, estimate its size from the remaining calibration items, and compare to the known truth. This is the _standard_ NSUM validation and directly tests whether the calibration is working.

**Uncertainty quantification: Bootstrap confidence intervals**

Resample respondents with replacement (10,000 times) to get CIs around the prevalence estimate. Standard in the NSUM literature ([Feehan and Salganik, 2016](https://doi.org/10.1177/0081175016665425); [Maltiel, Raftery, McCormick et al., 2015](https://doi.org/10.1214/15-AOAS827)).

**Robustness check: Split-half comparison**

Estimate prevalence in two random halves of the sample independently. A sanity check rather than a core test. With a moderately rare outcome (~3-4%), half-sample estimates will be noisier, so some divergence is expected.

**Sensitivity analyses**: Vary the visibility correction, compare basic vs. generalised estimator, examine robustness to item selection.

---
## What would the estimate tell us?

**Policy-relevant outputs:**

* First network-based UK prevalence estimate for hoarding disorder - bypassing the insight problem that limits self-report data.

* **Demographic signal**: Can examine whether _respondent_ demographics (age, region, etc.) predict reporting more hoarders in one's network → _indirect evidence_ about where hoarding concentrates.

* **Service planning**: Fire services can predict caseloads. Housing associations can justify support workers. Adult social care can plan self-neglect interventions, etc.

* **Comparison with self-report estimates**: If the NSUM estimate differs substantially from the 2-6% self-report range, that itself is informative about the insight gap.

???
Note: NSUM does not directly decompose prevalence by the _hoarders'_ demographics without additional design features (e.g., follow-up questions about contacts' characteristics).

---
## Methodological contributions

This study would deliver:

* **First UK NSUM calibration**: the item set, demographic reconstruction procedure, and back-estimation validation are publishable independently as a methods paper.

* **First application of NSUM to hoarding disorder**: novel target population for the method, internationally.

* **Documentation of UK-specific challenges**: the name abbreviation problem, the absence of a living-population-by-first-name database, and the calibration workaround.

* **Template for further UK applications?** → e.g., hidden homelessness.

---
## Potential challenges

**Threshold subjectivity**: When does clutter become "problematic"? Concrete examples help but a grey zone between "messy" and "pathologically hoarded" remains. Images from validated scale?

**Home access post-COVID**: People visit each other's homes less than they used to. Reduced home access reduces the effective visibility factor.

**Supplementary sample may skew severe**: Recruits from population at the severe end. Their visibility may be higher than moderate hoarders → potential downward bias on prevalence estimate.

**Stigma may suppress reporting**: Respondents may be reluctant to classify contacts as having a hoarding problem → e.g., loyalty or uncertainty about severity.

→ These are real but manageable and no worse than issues facing most NSUM applications (e.g., intravenous drug use).

---
## Summary

* Hoarding disorder prevalence in the UK is poorly estimated. Existing approaches are limited by the insight problem (self-report) or capture only the most severe cases (administrative data).

* NSUM offers a genuinely different approach: asking _other people_ what they have observed.

* Hoarding is a strong NSUM candidate: physically observable, definitionally concrete, genuinely hidden from administrative systems.

* The study would deliver the first UK NSUM calibration, the first network-based hoarding prevalence estimate, and a template for further applications on hidden populations?

---
## Acknowledgements

**AI declaration**:  The author(s) have made use of AI tools ([Claude Opus 4.6](https://www.anthropic.com/news/claude-opus-4-6)) to help develop and improve the presentation. In all cases, the author(s) made the final decision and take(s) full responsibility.

Slides were built with [Xaringan](https://github.com/yihui/xaringan) and [XaringanExtra](https://pkg.garrickadenbuie.com/xaringanExtra/). Many thanks to those building packages and supporting the (open source) [R](https://www.r-project.org/) eco-system.

You for listening!

---
## References and further reading (errors = blame RefManageR)

<a name=bib-APA2013></a>[American Psychiatric
Association](#cite-APA2013) (2013). "Diagnostic and Statistical Manual
of Mental Disorders".

<a name=bib-Cath2017></a>[Cath, D. C., K. Nizar, D. Boomsma, et
al.](#cite-Cath2017) (2017). "Age-Specific Prevalence of Hoarding and
Obsessive Compulsive Disorder: A Population-Based Study". In: _The
American Journal of Geriatric Psychiatry_ 25.3, pp. 245-255. DOI:
[10.1016/j.jagp.2016.11.006](https://doi.org/10.1016%2Fj.jagp.2016.11.006).

<a name=bib-ClayWarner2022></a>[Clay-Warner, J., T. Kawashima, and T.
G. Edgemon](#cite-ClayWarner2022) (2022). "Measure of Personal Network
Size Using the Known Population Method: A Methodological Guide". In:
_American Journal of Public Health_ 112.5, pp. 747-753. DOI:
[10.2105/AJPH.2022.306731](https://doi.org/10.2105%2FAJPH.2022.306731).

<a name=bib-Feehan2016></a>[Feehan, D. M. and M. J.
Salganik](#cite-Feehan2016) (2016). "Generalizing the Network Scale-Up
Method: A New Estimator for the Size of Hidden Populations". In:
_Sociological Methodology_ 46.1, pp. 153-186. DOI:
[10.1177/0081175016665425](https://doi.org/10.1177%2F0081175016665425).

<a name=bib-Fenoy2024></a>[Fenoy, A., M. Bojanowski, and M. J.
Lubbers](#cite-Fenoy2024) (2024). "Automated Name Selection for the
Network Scale-Up Method". In: _Field Methods_ 36.3, pp. 249-265. DOI:
[10.1177/1525822X241243115](https://doi.org/10.1177%2F1525822X241243115).

---
## More references

<a name=bib-Frost1996></a>[Frost, R. O. and T. L.
Hartl](#cite-Frost1996) (1996). "A Cognitive-Behavioral Model of
Compulsive Hoarding". In: _Behaviour Research and Therapy_ 34.4, pp.
341-350.

<a name=bib-Frost2011></a>[Frost, R. O., G. Steketee, and D. F.
Tolin](#cite-Frost2011) (2011). "Comorbidity in Hoarding Disorder". In:
_Depression and Anxiety_ 28.10, pp. 876-884. DOI:
[10.1002/da.20861](https://doi.org/10.1002%2Fda.20861).

<a name=bib-Frost2008SIR></a>[Frost, R. O., G. Steketee, D. F. Tolin,
et al.](#cite-Frost2008SIR) (2008). "Development and Validation of the
Clutter Image Rating". In: _Journal of Psychopathology and Behavioral
Assessment_ 30.3, pp. 193-203. DOI:
[10.1007/s10862-007-9068-7](https://doi.org/10.1007%2Fs10862-007-9068-7).

<a name=bib-Frost2004></a>[Frost, R. O., G. Steketee, and L.
Williams](#cite-Frost2004) (2000). "Hoarding: A Community Health
Problem". In: _Health & Social Care in the Community_ 8.4, pp. 229-234.
DOI:
[10.1046/j.1365-2524.2000.00245.x](https://doi.org/10.1046%2Fj.1365-2524.2000.00245.x).

<a name=bib-Habecker2015></a>[Habecker, P., K. Dombrowski, and B.
Khan](#cite-Habecker2015) (2015). "Improving the Network Scale-Up
Estimator: Incorporating Means of Sums, Recursive Back Estimation, and
Sampling Weights". In: _PLOS ONE_ 10.12, p. e0143406. DOI:
[10.1371/journal.pone.0143406](https://doi.org/10.1371%2Fjournal.pone.0143406).

---
## More references 2

<a name=bib-Iervolino2009></a>[Iervolino, A. C., N. Perroud, M. A.
Fullana, et al.](#cite-Iervolino2009) (2009). "Prevalence and
Heritability of Compulsive Hoarding: A Twin Study". In: _American
Journal of Psychiatry_ 166.10, pp. 1156-1161. DOI:
[10.1176/appi.ajp.2009.08121789](https://doi.org/10.1176%2Fappi.ajp.2009.08121789).

<a name=bib-Josephs2024></a>[Josephs, N., D. M. Feehan, and F. W.
Crawford](#cite-Josephs2024) (2024). "A Sample Size Formula for Network
Scale-Up Studies". In: _Sociological Methods & Research_ 53.3, pp.
1252-1289. DOI:
[10.1177/00491241221122576](https://doi.org/10.1177%2F00491241221122576).

<a name=bib-Killworth1998a></a>[Killworth, P. D., E. C. Johnsen, C.
McCarty, et al.](#cite-Killworth1998a) (1998). "A Social Network
Approach to Estimating Seroprevalence in the United States". In:
_Social Networks_ 20.1, pp. 23-50.

<a name=bib-Laga2021></a>[Laga, I., L. Bao, and X. Niu](#cite-Laga2021)
(2023). "A Correlated Network Scale-Up Model: Finding the Connection
Between Subpopulations". In: _Journal of the American Statistical
Association_ 118.543, pp. 1515-1524. DOI:
[10.1080/01621459.2023.2165929](https://doi.org/10.1080%2F01621459.2023.2165929).

<a name=bib-Lubbers2025></a>[Lubbers, M. J., M. Bojanowski, N.
Targarona Rifa, et al.](#cite-Lubbers2025) (2025). "The Measurement
Properties of Aggregated Relational Data and NSUM-Estimated Network
Size". In: _Sociological Methodology_ 55.2, pp. 300-328. DOI:
[10.1177/00811750251340398](https://doi.org/10.1177%2F00811750251340398).

---
## More references 3

<a name=bib-Lubbers2019></a>[Lubbers, M. J., J. L. Molina, and H.
Valenzuela-García](#cite-Lubbers2019) (2019). "When Networks Speak
Volumes: Variation in the Size of Broader Acquaintanceship Networks".
In: _Social Networks_ 59, pp. 55-69. DOI:
[10.1016/j.socnet.2018.08.004](https://doi.org/10.1016%2Fj.socnet.2018.08.004).

<a name=bib-Maltiel2015></a>[Maltiel, R., A. E. Raftery, T. H.
McCormick, et al.](#cite-Maltiel2015) (2015). "Estimating Population
Size Using the Network Scale Up Method". In: _The Annals of Applied
Statistics_ 9.3, pp. 1247-1277. DOI:
[10.1214/15-AOAS827](https://doi.org/10.1214%2F15-AOAS827).

<a name=bib-Mataix-Cols2010></a>[Mataix-Cols, D., R. O. Frost, A.
Pertusa, et al.](#cite-Mataix-Cols2010) (2010). "Hoarding Disorder: A
New Diagnosis for DSM-V?" In: _Depression and Anxiety_ 27.6, pp.
556-572. DOI: [10.1002/da.20693](https://doi.org/10.1002%2Fda.20693).

<a name=bib-McCarty2001></a>[McCarty, C., P. D. Killworth, H. R.
Bernard, et al.](#cite-McCarty2001) (2001). "Comparing Two Methods for
Estimating Network Size". In: _Human Organization_ 60.1, pp. 28-39.
DOI:
[10.17730/humo.60.1.efx5t9gjtgmga73y](https://doi.org/10.17730%2Fhumo.60.1.efx5t9gjtgmga73y).

<a name=bib-Pertusa2010></a>[Pertusa, A., R. O. Frost, M. A. Fullana,
et al.](#cite-Pertusa2010) (2010). "Refining the Diagnostic Boundaries
of Compulsive Hoarding: A Critical Review". In: _Clinical Psychology
Review_ 30.4, pp. 371-386. DOI:
[10.1016/j.cpr.2010.01.007](https://doi.org/10.1016%2Fj.cpr.2010.01.007).

---
## More references 4

<a name=bib-Postlethwaite2019></a>[Postlethwaite, A., S. Kellett, and
D. Mataix-Cols](#cite-Postlethwaite2019) (2019). "Prevalence of
Hoarding Disorder: A Systematic Review and Meta-Analysis". In: _Journal
of Affective Disorders_ 256, pp. 309-316. DOI:
[10.1016/j.jad.2019.06.004](https://doi.org/10.1016%2Fj.jad.2019.06.004).

<a name=bib-Samuels2008></a>[Samuels, J. F., O. J. Bienvenu, M. A.
Grados, et al.](#cite-Samuels2008) (2008). "Prevalence and Correlates
of Hoarding Behavior in a Community-Based Sample". In: _Behaviour
Research and Therapy_ 46.7, pp. 836-844. DOI:
[10.1016/j.brat.2008.04.004](https://doi.org/10.1016%2Fj.brat.2008.04.004).

<a name=bib-Snowdon2012></a>[Snowdon, J., A. Pertusa, and D.
Mataix-Cols](#cite-Snowdon2012) (2012). "On Hoarding and Squalor: A Few
Considerations for DSM-5". In: _Depression and Anxiety_ 29.5, pp.
417-424. DOI: [10.1002/da.21943](https://doi.org/10.1002%2Fda.21943).

<a name=bib-Steketee2003></a>[Steketee, G. and R. O.
Frost](#cite-Steketee2003) (2003). "Compulsive Hoarding: Current Status
of the Research". In: _Clinical Psychology Review_ 23.7, pp. 905-927.
DOI:
[10.1016/j.cpr.2003.08.002](https://doi.org/10.1016%2Fj.cpr.2003.08.002).

<a name=bib-Tolin2008></a>[Tolin, D. F., R. O. Frost, G. Steketee, et
al.](#cite-Tolin2008) (2008). "The Economic and Social Burden of
Compulsive Hoarding". In: _Psychiatry Research_ 160.2, pp. 200-211.
DOI:
[10.1016/j.psychres.2007.08.008](https://doi.org/10.1016%2Fj.psychres.2007.08.008).

---
## More references 5

<a name=bib-Volker2025></a>[Völker, B., B. Hofstra, R. Corten, et
al.](#cite-Volker2025) (2025). "Who's in Your Extended Network?
Analysing the Size and Homogeneity of Acquaintanceship Networks in the
Netherlands". In: _Social Networks_ 83, pp. 173-185. DOI:
[10.1016/j.socnet.2025.05.007](https://doi.org/10.1016%2Fj.socnet.2025.05.007).

<a name=bib-Zheng2006></a>[Zheng, T., M. J. Salganik, and A.
Gelman](#cite-Zheng2006) (2006). "How Many People Do You Know in
Prison? Using Overdispersion in Count Data to Estimate Social Structure
in Hidden Populations". In: _Journal of the American Statistical
Association_ 101.474, pp. 409-423. DOI:
[10.1198/016214505000001168](https://doi.org/10.1198%2F016214505000001168).