Four in Five Had Clean Kitchens: What the Sanitation Evidence Actually Shows

1. Introduction: the first line of every report

Improve sanitation. It is the opening recommendation of virtually every integrated pest management programme, every inspection report and every extension factsheet, and this journal has repeated it without examining the evidence.

1.1 Both are from the same paper

The first is quoted in every press account of it. The second is in the discussion section.¹³

1.2 What this article argues

That the association is genuine, that it explains a minority of infestations, and that two causal-direction problems make even the association weaker than it looks. Sections 9, 12 and 14 are the case.

2. The study

The design, which is good.

Interviews with residents of 388 randomly selected apartments in seven low-income senior citizen buildings, with apartment conditions also assessed and glue traps placed to detect the presence of the cockroaches.³

2.1 The trapping protocol

Four sticky traps per apartment, over a 2-wk period.¹

Which is a stated effort, a stated duration and a stated number of devices, and our article on what a trap catch means spent five thousand words complaining that those three things are usually missing.

2.2 Which is the design's real strength

A study relying on residents reporting whether they have cockroaches would have missed more than a third of the infestations, per §17. The traps are what make the awareness finding possible at all.

2.3 Three data streams

What residents say, what an inspector sees, and what the traps catch. The study can therefore compare reported infestation against measured infestation, which is where §17 comes from.

2.4 And the population is specific

Low-income senior and disabled residents in two New Jersey municipalities, which is a narrow population and §27 says so.⁴

3. What it found

The headline results.

Among 344 apartments where trap count data were obtained, 30% had German cockroaches.⁶ Press accounts give 29 per cent.⁴

The odds of having cockroaches in apartments with a 'poor' sanitation rating in kitchens and bathrooms was 2.7 times greater than that in apartments with better sanitation conditions.⁶

And: residents' tolerance to cockroaches is significantly associated with presence of cockroaches and cockroach population size.⁶

3.1 A thirty per cent prevalence is high

One account calls it a high rate compared with the rest of the society.⁴

A comparison study of 258 apartments occupied by families revealed 28% of homes had German cockroaches, so the figure is consistent across two surveys of comparable housing.²

3.2 And a third survey, elsewhere

A study in a southwestern Iranian city monitored 573 sampling units over five weeks, with questionnaires to occupants of 348 residential units, and found almost 39% of sampling units were infested, recording five species from two families.⁶

Different country, different housing, similar order of magnitude. Which suggests the prevalence is a property of multi-unit housing rather than of any particular jurisdiction, and that reading is ours.

4. The association is real

Stated plainly before anything else, because this article is not a debunking.

An odds ratio of 2.7 with a sample of several hundred apartments is a substantial and credible finding.⁶

4.1 And it is biologically sensible

German cockroaches need food, water and harbourage. Accumulated food residue supplies the first, and our articles on the species describe how little of it a population requires.

4.2 So nothing below disputes that sanitation matters

The argument concerns how much of the total problem it accounts for, which is a different question from whether it is a real factor.

5. And it is not new

The authors place it in a literature.

Their findings are consistent with previous literature, that the presence of cockroach infestations is associated with apartment sanitation conditions, citing three earlier studies spanning nearly thirty years.¹

5.1 And the authors are candid about the state of it

Noting that prevalence is affected by both environmental conditions and building occupant behavior, but their relationships are not well studied.⁶

Which is an unusual admission to find attached to a recommendation the whole industry makes daily.

5.2 Which makes this a replication

And replications are worth more than novel findings, which this journal has said often enough that we should say it when one appears.

6. With one dissent

Recorded by a later summary of the same literature.

That although Lee and Lee (2000) did not find a relationship between sanitation and cockroach population levels, most published studies have found one.⁶

6.1 And one source frames the balance

That sanitation conditions in the home are strongly correlated with the presence of cockroaches in apartments, with the dissenting study named as the exception.⁶

Which is the honest way to report a literature with one contrary result in it.

6.2 Note what the dissent is about

Population levels rather than presence. Those are different outcomes, and it is entirely coherent for sanitation to affect whether cockroaches establish without affecting how numerous they become once established.

That distinction is ours and we have not seen it drawn in these sources.

7. The number nobody quotes

From the discussion of the same paper.

Only 10 and 7% of the apartments had poor sanitation conditions in kitchens and bathrooms, respectively.¹

7.1 It appears in none of the press coverage

Five separate accounts of this study report the odds ratio and none of them reports the exposure prevalence.³⁴⁵⁷

7.2 And it is not hidden

It sits in the discussion section of an open paper, in the same paragraph as the sentence about estimating the benefit of good housekeeping.¹ Nobody concealed it and nobody carried it forward either.

7.3 Which is the usual pattern

A relative measure travels and a base rate does not, and our article on the brown recluse found the same asymmetry in a clinical context.

8. Why that changes the picture

The logic, before the arithmetic.

Thirty per cent of apartments had cockroaches. Ten per cent had poor kitchen sanitation. Even if every single poorly-kept apartment were infested, it could account for at most a third of the infested ones.

8.1 And not every one was

An odds ratio of 2.7 is not certainty. It raises the probability substantially and leaves plenty of poorly-kept apartments uninfested.⁶

8.2 So the maximum is a ceiling nobody reaches

Which means the true share is lower than a third, and §9 estimates how much lower.

9. The arithmetic

Our working, shown so it can be checked.

Take the reported figures: overall prevalence 30 per cent, exposure prevalence 10 per cent for poor kitchen sanitation, odds ratio 2.7.⁶¹

9.1 Solving for the two group rates

An infestation rate of roughly 28 per cent among the well-kept apartments and roughly 51 per cent among the poorly-kept ones satisfies both the odds ratio and the overall prevalence to within rounding.

9.2 Which distributes the infested apartments

Of 344 apartments, about 34 were poorly kept and about 310 were not. Roughly 17 infested in the first group and roughly 87 in the second, totalling about 104, against the 103 implied by a thirty per cent prevalence.

9.3 So about four in five infested apartments were not poorly kept

Eighty-three per cent of them, on these figures.

9.4 A sanity check

If the poorly-kept group had been infested at a hundred per cent, it would contribute 34 of the 103 infested apartments, or a third. Our estimate of 17 is half that, which is what a 2.7-fold odds difference rather than a certainty should produce.

So the estimate sits sensibly between the two bounds the reported figures impose.

9.5 The assumptions

That the odds ratio applies to the kitchen rating alone, that the 10 per cent exposure figure applies to the 344 apartments with trap data rather than the 388 surveyed, and that the reported figures are as published. None is certain and the estimate is approximate rather than exact, which the chart note says.

10. What it means

Stated carefully.

Sanitation is a real risk factor and a minor explanation. Most of the infested apartments in this study were kept to a standard the inspectors rated acceptable.

10.1 The reservoir argument deserves separating

A small group with a high rate can matter out of proportion to its size if it seeds the rest, and the authors say exactly that: those apartments may suffer chronic infestations and become sources of new infestations in multi-unit dwellings.¹

Which is a claim about dispersal rather than about attribution, and our cockroach articles describe the movement between units that would make it work. It is compatible with §9 and it is a different argument.

10.2 Which is not what the headlines said

One press account is titled around clean apartments being less likely to house cockroaches, which is true and carries the implication that the dirty ones are where the cockroaches are.⁷

10.3 And it is not what the authors said either

They wrote that the exposure figure can be used for estimating the benefit of education and good housekeeping in cockroach management, and that investing extra effort in these apartments will likely maximize the control effect since they may suffer chronic infestations and become sources of new infestations in multi-unit dwellings.¹

10.4 Which is a targeting argument, and it is good

The poorly-kept apartments are a small group with a high rate, which makes them efficient to target and makes them plausible reservoirs. Nothing in §9 contradicts that.

What §9 contradicts is the inference that sanitation explains the infestations.

11. Which is a standard result

And worth naming, because it recurs.

A strong relative risk on an uncommon exposure accounts for a small share of total cases. Epidemiology calls the share the population attributable fraction and treats the distinction as elementary.

11.1 Why the distinction gets lost

Because a relative risk sounds like an explanation and an attributable fraction sounds like bookkeeping. One makes a headline and the other makes a policy.

Our article on allergen thresholds found action levels quoted without their derivation and our article on trap catch found counts quoted without their effort. This is the same failure applied to a risk factor.

11.2 The intuition

Relative risk tells you what the exposure does to an individual. Attributable fraction tells you what removing it would do to the total. They answer different questions and only the first is ever quoted.

11.3 And the practical form

Perfect sanitation in every apartment in this study would, on these figures, have prevented something like a sixth of the infestations. The rest would have happened anyway.

That estimate carries all the caveats in §9.4.

12. The first causal problem

Which makes even the 2.7 an upper bound.

The sanitation rating was produced by an inspector assessing apartment conditions.³

12.1 This is a general problem with observational ratings

Whenever the exposure is assessed by looking and the outcome leaves visible traces, the two measurements share a source. Epidemiology has a name for it and the remedy is blinding the assessor.

12.2 An infested kitchen looks unsanitary

Cockroach droppings, smear marks, shed skins, egg cases and dead insects are among the most conspicuous things in an infested kitchen, and they are exactly what an observer rating cleanliness would notice.

12.3 So the exposure is measured partly by the outcome

A kitchen rated poor may be rated poor because it has cockroaches in it, which would produce an association between the rating and the infestation with no causal contribution from housekeeping at all.

That is our objection and we found no source raising it.

13. How large that effect might be

We cannot say, and neither can anybody from this study.

It would depend entirely on the rating instrument: whether it enumerated specific conditions, whether pest evidence was scored separately, and whether the assessors were blinded to the trap results.

13.1 None of which our sources describe

The press accounts say conditions were assessed and the abstract says a rating was assigned.³⁶ The instrument itself is in the paper, which we have not read in full.

13.2 So this is a question rather than a criticism

If pest evidence was excluded from the rating and the assessors were blind, the objection disappears. The published summaries do not let us tell.

14. The second causal problem

And this one we think is clearer.

Residents who were more tolerant of cockroaches also had higher rates of cockroach infestation and higher cockroach numbers.⁴

14.1 Two associations, not one

Tolerance was associated with both whether cockroaches were present and how many there were.⁶ The second is the stronger signal, because a count is a continuous measure and less easily explained by a single threshold effect.

14.2 The paper's framing treats tolerance as an input

Recommending assessing and reducing cockroach tolerance thresholds as a management measure.¹

15. Reading the tolerance finding

Two ways.

Tolerance permits a population to grow, because a resident who is not bothered does not report, does not clean and does not act.

Or: living with cockroaches for years produces tolerance, because habituation to a persistent unpleasant stimulus is among the most reliable findings in human psychology.

15.1 The study is cross-sectional

Everything was measured at one time, so it cannot distinguish them. That is a limitation of design rather than of care.

15.2 We think the second is likelier

Because a resident newly infested is not tolerant, and a resident infested for a decade in housing where complaints have not produced results has learned something about the value of being bothered.

That reading is ours and it changes what the finding recommends.

15.3 Which matters for what you do about it

If tolerance is a cause, the intervention is education. If it is a consequence, the intervention is fixing the infestation, and the tolerance resolves on its own.

16. The threshold buried in it

A number worth extracting.

The median cockroach count when residents were bothered by cockroaches was ≥3, on four sticky traps over two weeks.¹

16.1 Three cockroaches is the complaint threshold

Which is a measured human action level for a structural pest, and our article on injury levels noted that this trade has almost none.

16.2 It is a threshold of annoyance, not of harm

Which makes it a measure of the residents rather than of the cockroaches, and it will move with tolerance, per §15. A population that has been present for years is measured against a higher bar.

16.3 Compare the health threshold

Our allergen article recorded action levels of 2 units per gram for sensitisation and 8 for asthma morbidity, established in a national study.

Whether a trap count of three corresponds to either is unknown to us, and the relationship between what people notice and what harms them is the question our threshold articles keep asking.

17. The awareness finding

The most striking number in the study.

Among interviewed residents whose apartments had existing cockroach infestations, 36% were unaware of the presence of cockroaches.⁶

Press accounts give 35 per cent and describe the finding as surprising.⁴

17.1 More than a third

Of people living inside a trap-confirmed infestation did not know it was there.⁶

17.2 And it is measured against a defined standard

Not against an inspector's impression but against four traps over two weeks, which means the unawareness is failure to notice a trap-confirmed population rather than disagreement about what counts as infested.¹

17.3 Which is a detection finding about humans

Our articles on inspection and on what a zero means concern professional detection failure. This is the same failure by the person with unlimited access and unlimited time.

18. Which undermines the behavioural model

An argument we think follows and that nobody makes.

If more than a third of infested residents do not know they are infested, then their housekeeping cannot be a response to the infestation, and the causal story in which resident behaviour drives infestation has to operate without their knowledge.

18.1 And it reverses a common assumption

That an infested household knows it is infested and has made a choice about it. More than a third had made no choice, because there was nothing in front of them to choose about.⁶

18.2 Which is possible

Habitual cleaning practices exist whether or not somebody knows about cockroaches, so the mechanism survives.

18.3 But the intervention does not

Education aimed at residents who do not know they have a problem is addressing the wrong stage. The first task is telling them, and §19 is what the authors say about education generally.

19. What the authors themselves recommend

And it is not what the press coverage implies.

Previous studies on the effect of education on sanitation had mixed results, citing three.¹

And then: since knowledge is insufficient for behavioral change, and this is a low resource population, providing regular cleaning services to apartments with ongoing sanitation issues is an option that also should be considered.¹

19.1 Knowledge is insufficient for behavioural change

Which is a sentence about every education-based intervention this journal has examined, including the certification training our applicator article found thinner evidence for than expected, and the client advice in §22.

People do not generally fail to act because they are uninformed. They fail because acting is costly, and the study population's costs are named in the next clause.

20. Which is a remarkable thing to have written

In an entomology journal.

The authors report that the standard intervention has mixed evidence, state that knowledge is insufficient for behavioural change, observe that the population lacks resources, and propose paying for cleaning instead.¹

20.1 It also concedes the sequence is backwards

The standard programme asks the resident to improve sanitation so that treatment will work. The authors' version supplies the sanitation as a service, which makes it part of the treatment rather than a precondition for it.¹

20.2 That is a resource allocation argument

Rather than a behavioural one, and it treats the sanitation deficit as a capacity problem instead of a compliance problem.

20.3 Which fits §9

If poor sanitation is concentrated in a tenth of apartments and education does not reliably change it, then supplying the cleaning directly is the intervention that actually addresses the exposure.

21. And it matches an earlier finding of ours

Which we had not connected.

Our article on allergen thresholds described an abatement trial that achieved a 96 per cent reduction on kitchen floors. Its interventions were education, bait, and professional cleaning.

21.1 Professional cleaning was in the successful protocol

Not education alone, and not chemical control alone. The trial that worked paid somebody to clean the apartments.

Two independent lines pointing at the same conclusion, and the connection is ours.

22. What this means for the client conversation

The practical payload.

Sanitation advice is worth giving. The association is real, replicated and biologically sensible.¹

It is not an explanation of why this client has cockroaches. On these figures roughly four in five infested homes were not rated poorly.

So leading with it is usually wrong. The client has almost always heard it already, frequently from somebody implying it was their fault.

Lead with the building instead. In multi-unit housing the neighbouring unit, the shared voids, the service risers and the common areas are where our cockroach articles locate the population.

22.1 And that is not evasion

It is where the attributable fraction actually is. If sanitation accounts for something like a sixth of infestations, then five sixths are explained by something else, and in multi-unit housing the something else is largely the building and its neighbours.

A contractor who leads with the building is leading with the larger cause, which is both the kinder conversation and the more accurate one.

22.2 The single-family case differs

Where there is no neighbouring unit, the introduction route matters more, and our articles on goods brought into a home describe it. The study population was multi-unit and the advice above is scoped to that.

23. Why blame is the wrong instrument

Three reasons, and only the first is about kindness.

It is usually factually wrong. Section 9.

It suppresses reporting. A tenant who expects to be blamed does not call, and our articles on reporting and on the social distribution of infestation describe what that does to a building.

And it misdirects the work. A contractor who has located the cause in the occupant's housekeeping has stopped looking for the harbourage, the entry route and the source unit.

23.1 And one the sources support

The authors note that lack of awareness contributes to the failure of future efforts to control infestations,³ which is the reporting problem stated from the other end: what the resident does not report, the programme does not treat.

23.2 The third is the professional failure

Because a satisfying explanation ends an investigation, and our articles on diagnosis repeatedly find that the comfortable answer displaces the correct one.

24. What sanitation advice is actually for

Reframing it usefully.

It improves the odds. A 2.7-fold difference is worth having even when it explains a minority of cases.⁶

It makes control work better. Our bait articles describe competing food sources reducing bait uptake, which is a mechanism specific to treatment rather than to infestation.

And it removes allergen. Which our threshold article established is a separate endpoint from killing the insects, addressed by cleaning directly.

24.1 And there is a fourth, rarely stated

Sanitation makes an infestation visible. A kitchen kept clear shows droppings and smears against a clean surface, and a cluttered one conceals them.

Which means better sanitation improves detection as well as control, and would on its own produce part of the association in §4 without changing a single cockroach. That is a third mechanism behind the same correlation and it is ours.

24.2 That third one is the strongest version

Cleaning is not primarily a control measure. It is the only intervention that addresses the protein the cockroaches leave behind, and it should be recommended on that basis rather than as a prerequisite for treatment working.

That framing is ours and it is the most useful thing in this article.

25. Our own position

The disclosure.

We write sanitation recommendations into reports. Section 22 says we should stop leading with them and §24 says we should justify them differently, and both are changes to our own practice.

25.1 And we have written it as a precondition

Reports that say treatment cannot be expected to succeed without sanitation improvements. Section 9 says most infested homes did not have a sanitation problem, which makes that framing wrong in the majority of cases we would write it about.

25.2 The commercial dimension

A sanitation recommendation is also a liability instrument. A report stating that the client was advised to improve housekeeping protects the contractor when the treatment does not work.

We should be honest that this is one reason the recommendation appears so reliably, and it is not a reason a client benefits from.

26. The Manitoba position

Two notes.

The housing type transfers. The study population is low-income multi-unit housing, of which this city has a great deal, and our cockroach articles describe the same building dynamics.

And the season does not. New Jersey and Winnipeg differ in almost everything except that German cockroaches live entirely indoors, which is the one pest for which our climate argument does not apply.

26.1 And the winter reporting effect

A cold climate concentrates complaints in the heating season, when residents are indoors and the population is at its most active. Our articles on seasonal service demand describe the pattern, and it means a survey done in one season would give a different unawareness figure than one done in another.

That is our speculation and the New Jersey study does not state its season.

26.2 What we could not find

Any Manitoba or Canadian prevalence survey for German cockroaches in multi-unit housing, any local work on sanitation ratings, and any assessment of how infestation is distributed across the housing stock here.

27. Limitations and open questions

One study, one population. Low-income senior and disabled residents in two New Jersey municipalities, in seven buildings.⁴

The paper was read as abstract, press summaries and a discussion extract. We do not have the full methods, the rating instrument or the confidence intervals, and §13.1 says why that matters.¹⁶

The arithmetic in §9 is ours and approximate. Its assumptions are listed in §9.4 and the true distribution could differ.

Sources disagree on the prevalence figures. Twenty-nine or thirty per cent infested, thirty-five or thirty-six per cent unaware.⁴⁶

Five of our seven sources are coverage of one paper. Press releases and aggregators reporting the same study, which is why the reference list is shorter than it looks.³⁵⁷

And we did not find the dissenting paper. The study reported as finding no relationship is known to us only as a citation inside a summary of the literature.⁶

Sections 1.2, 4.2, 6.1, 8, 9, 10.3, 11, 12, 13, 15.2, 15.3, 16.2, 18, 20.1, 21, 22, 23, 24 and 25.1 are our reasoning. The attributable fraction argument and its arithmetic, the measurement-contamination objection, the reading of the tolerance finding, the argument from the awareness result, the case against leading with blame and the reframing of what sanitation advice is for are ours rather than sourced positions.

28. Conclusion

The evidence for sanitation is real. Apartments rated poor in kitchens and bathrooms had 2.7 times the odds of German cockroaches, a finding the authors place in a literature spanning three decades.⁶¹ What the press coverage left out is in the discussion: only ten per cent of the apartments had poor kitchen sanitation, and thirty per cent had cockroaches.¹ Working those figures together gives roughly twenty-eight per cent infestation among the well-kept apartments and roughly fifty-one per cent among the poorly-kept ones, which puts about four in five infested apartments in homes an inspector rated acceptable. That is the ordinary epidemiological result that a strong relative risk on an uncommon exposure explains a small share of the total, and it is the reason a real risk factor can be a poor explanation.

Two things weaken even the association. A sanitation rating produced by an inspector looking at an infested kitchen is partly a measurement of the infestation, since droppings, smears and shed skins are both. And residents' tolerance, reported as associated with higher counts and framed as something to reduce, is at least as plausibly the product of years of living with an unresolved problem as its cause. Thirty-six per cent of residents whose apartments were infested did not know it,⁶ which means their housekeeping was not a response to anything and that education aimed at them is addressing the wrong stage.

The authors' own recommendation is the part worth carrying. They report that studies of education have had mixed results, state that knowledge is insufficient for behavioural change, note the population's lack of resources, and propose providing regular cleaning services.¹ That is a capacity argument rather than a compliance one, and it matches the abatement trial our allergen article described, whose successful protocol combined bait with professional cleaning. For a contractor the consequence is that sanitation advice is worth giving and is not an explanation, that leading with it is usually both wrong and unkind, and that its strongest justification is not control at all: cleaning is the only intervention that removes the protein the cockroaches leave behind.

References

1. Residents' attitudes and home sanitation predict presence of German cockroaches in apartments for low-income senior residents. Journal article in an economic entomology title, publisher page, read as abstract and discussion extracts. Used for the species background that the German cockroach is the most common and troublesome cockroach species in apartment buildings in that country, contaminates food, produces allergens leading to sensitivity, increases asthma morbidity and causes economic loss through control costs and contaminated food; for the observation that despite well-documented health and economic impacts some residents tolerate low, moderate or even high levels of cockroach activity; for the trapping protocol of four sticky traps per apartment over a two-week period; for the finding that the median cockroach count when residents were bothered by cockroaches was three or more; for the recommendation that assessing and reducing cockroach tolerance thresholds and improving housekeeping through resident education and assistance from community and housing management should be incorporated into management; for the statement that the findings are consistent with previous literature associating cockroach infestations with apartment sanitation conditions, citing three earlier studies; for the reported exposure figures that only 10 and 7 per cent of apartments had poor sanitation conditions in kitchens and bathrooms respectively, and the authors' note that this can be used for estimating the benefit of education and good housekeeping, with extra effort in those apartments likely to maximise control effect since they may suffer chronic infestations and become sources of new infestations in multi-unit dwellings; for the statement that previous studies on the effect of education on sanitation had mixed results, citing three; and for the conclusion that since knowledge is insufficient for behavioural change and this is a low resource population, providing regular cleaning services to apartments with ongoing sanitation issues is an option that should also be considered. https://academic.oup.com/jee/article/112/1/284/5128815
2. Author's copy of the above paper, hosted by a university entomology department. Used for the keyword list covering the species, tolerance, sanitation and low-income housing; for the comparison figure that a survey of 258 apartments occupied by families revealed 28 per cent of homes had German cockroaches; and for the same discussion passages on exposure prevalence and on the mixed results of education studies. https://entomology.rutgers.edu/personnel/changlu-wang/docs/Wang2018ResidentsAttitudesGermanCockroaches.pdf
3. University newsroom account of the above study. Institutional press material, cited as attributed material. Used for the study description of interviews with residents of 388 randomly selected apartments in seven low-income senior citizen buildings in one state, with apartment conditions also assessed and glue traps placed to detect the cockroaches; for the finding that apartments with a poor sanitation rating in kitchens and bathrooms were 2.7 times more likely to have cockroaches than apartments with better conditions; for the finding that residents more tolerant of cockroaches had higher rates of infestation and higher cockroach numbers; for the concern that residents' lack of awareness allows cockroaches to continue reproducing, contaminating food, spreading to neighbours and leaving allergens, and contributes to the failure of future control efforts; for the researchers' conclusion that improving control in these communities requires educating residents on the dangers of infestation and helping them improve housekeeping through education and assistance from community and housing management; and for the funding attribution to a federal integrated pest management centre. https://sebsnjaesnews.rutgers.edu/2018/10/rutgers-study-reveals-residents-attitudes-and-home-sanitation-predict-presence-of-german-cockroaches/
4. University news account of the same study. Institutional press material, cited as attributed material. Used for the location of the study in two named municipalities; for the description of the participants as senior citizen and disabled residents in 388 apartments in seven apartment buildings; for the reported finding that 29 per cent of surveyed apartments had German cockroaches, described as a high rate compared with the rest of society; and for the reported finding that 35 per cent of residents in apartments with cockroaches were unaware of their presence, described as surprising. https://www.rutgers.edu/news/home-cleanliness-residents-tolerance-predict-where-cockroaches-take-residence
5. Science news aggregator account of the same study. Aggregated press material, cited as attributed material. Used for the same reported findings on the 2.7-fold odds difference, the association between resident tolerance and both infestation rates and cockroach numbers, and the researchers' conclusion regarding education and assistance with housekeeping; and for the full author list of the underlying paper. https://www.sciencedaily.com/releases/2018/11/181107082459.htm
6. Research repository record reproducing abstracts of the above paper and of related work on sanitation in cockroach integrated pest management. Repository listing rather than the primary publications, flagged accordingly. Used for the abstract figures that among the 344 apartments where trap count data were obtained 30 per cent had German cockroaches, that among interviewed residents whose apartments had existing infestations 36 per cent were unaware of the presence of cockroaches, that the odds of having cockroaches in apartments with a poor sanitation rating in kitchens and bathrooms was 2.7 times greater than in apartments with better conditions, and that residents' tolerance is significantly associated with both presence and population size; for the statement that sanitation conditions in the home are strongly correlated with the presence of cockroaches in apartments, and that although one named 2000 study did not find a relationship between sanitation and cockroach population levels, most published studies have; for the framing that prevalence of infestations is affected by both environmental conditions and building occupant behaviour but their relationships are not well studied; and for a separate survey in a southwestern Iranian city monitoring 573 sampling units over a five-week trapping period with questionnaires to occupants of 348 residential units, in which almost 39 per cent of sampling units were infested and five species from two families were recorded. https://www.researchgate.net/publication/270956415_Evaluation_of_sanitation_in_an_IPM_program_for_cockroach_infestation_in_housing
7. Research news site account of the same study. Aggregated press material, cited as attributed material. Used for its framing that clean apartments are less likely to house cockroaches; and for its restatement of the reported prevalence, the 2.7-fold odds difference, the tolerance association and the researchers' recommendations. https://www.futurity.org/cockroaches-sanitation-1906542/