A total of 661 manuscripts were identified from searches of databases, Google Scholar, and reference lists, of which 218 were duplicates, leaving 443 that were screened for eligibility. Of these, 295 were removed at the title and abstract screening stages, leaving 148 for eligibility screening of full-text articles. After removal of 51 articles due to ineligibility, and inclusion of an additional 6 and 8 articles from the two updated searches in November 2017 and September 2018, respectively, 111 manuscripts (representing 112 studies) were considered eligible and included in the review (Figure 1).
Studies were excluded for a number of reasons: (1) household chaos was included within a composite measure of family functioning or household environment, (2) no child, parent, or family outcomes were reported, (3) only predictors of household chaos were assessed, (4) the construct of household chaos was not adequately measured (i.e. a different construct was measured), (5) household chaos was only included as a covariate in the analysis and not reported in the results, (6) the paper reported simulation data, and (7) the paper was retracted.
A summary of characteristics of the included studies can be found in Table 3. Overall there was generally an even mix of cross-sectional and longitudinal study designs (n=57 vs n=52), in addition to two experimental/laboratory studies and one case-control study. With respect to the longitudinal studies, samples were primarily drawn from large, nationally representative, longitudinal cohorts, which was reflected in the relatively large number of analyses conducted in sample sizes of greater than 1000 participants (24/104; 23%).
There has been an increasing trend in recent years in the number of manuscripts investigating the relationship between household chaos and child/parent/family outcomes (Figure 2), with the majority of publications from research conducted in the USA (n=74) and the UK (n=21) (Figure 3).
Methods of household chaos assessment
Studies varied somewhat in the method of household chaos assessment, although overwhelmingly the most frequently used methods were the long-form version of the Confusion, Hubbub, and Order Scale (CHAOS)(22) (n=44) and the short-form version of CHAOS(23) (n=42). The long-form version consists of 15 questions in a True-False response format, with each item reflecting household characteristics that directly represent a chaotic home environment, while at the same time specifically excluding any items that represent adequacy of the dwelling or quality of furnishings, or imply unsanitary conditions. The questionnaire was developed to be used in families regardless of the child’s age, and has been demonstrated to have satisfactory internal consistency, test-rest reliability, and adequate psychometric properties(22). The short-form version of CHAOS consists of six items rated on a five-point scale (1=definitely untrue, 5=definitely true), which has been demonstrated to have acceptable internal consistency(23). Another 10 studies used adapted long- and short-versions of CHAOS, and the remaining 16 studies relied on questionnaires developed specifically for the study and/or direct observation in the home.
Outcomes of Interest
A summary of study characteristics, including sample size, age of participants, study design, SES status of participants, measure of household chaos, and outcome/s of interest can be found in Table 4. Here we provide a brief narrative review of study findings, with outcomes of interest mapped into seven categories.
Cognitive and Academic
Sixteen manuscripts were identified that investigated the relationship between household chaos and cognitive/academic outcomes(3, 7, 11, 24-36). The studies covered the spectrum of age groups, from very early childhood to later adolescence, although the majority were conducted in young children (i.e. ≤5 years). Outcomes assessed included executive function, IQ, general cognitive ability, and a range of academic measures, including reading comprehension, academic achievement, study skills, and learning. Overall, household chaos was consistently associated with adverse cognitive and academic outcomes. There was also evidence for an effect of household chaos on outcomes after controlling for SES(3, 30, 32). However, null findings were reported in 2 of the 16 studies; one study (n=203) did not find a significant relationship between household chaos and cognitive performance(33), while another study (n=65) failed to demonstrate a significant relationship between household chaos and intelligence, academic achievement, and executive functioning in 6-16 years olds(36).
Socio-emotional and Behavioural
The majority of studies investigated the relationship between household chaos and socio-emotional and behavioural outcomes(3, 4, 6, 10, 14, 24, 25, 28, 29, 31-33, 37-83), although 18/59 of these studies only investigated household chaos as a mediator or moderator in the relationship between a predictor and a socio-emotional/behaviour outcome. Outcomes included responses to challenges, social skills/competence, emotion regulation, risky behaviours, attention, aspirations, aggression, conduct problems, and callous-unemotional traits. Overwhelmingly, household chaos was shown to be associated with adverse outcomes in both younger children and adolescents(3, 4, 6, 10, 25, 31-33, 38, 41-44, 49, 55, 57, 62-64, 67, 68, 72, 80, 82, 83); however, it was not shown to be associated with self-regulation and effortful control(53, 83), empathy(64), sexual risk or other violent behaviours in adolescents(44), or occupational aspirations in 7 year olds(47).
A total of six manuscripts investigated the link between household chaos and communication, all of which were conducted in the early childhood setting(6, 7, 23, 25, 84, 85). Outcomes assessed included non-verbal abilities, receptive and expressive language, and phonological awareness. Household chaos was consistently linked with adverse effects on communication outcomes across all six analyses. After controlling for all other measures of household chaos, lack of routines was significantly associated with lower receptive vocabulary scores in 5 year olds(6), and in an analysis that controlled for 13 covariates, including maternal education and poverty, household disorganisation was associated with significant decreases in both receptive and expressive language in 3-year-old children (n=1145)(85). Finally, when investigating the heritability of cognitive abilities as a function of the child’s early environment, household chaos, which is classified as a proximal environmental determinant, had stronger effects than distal environmental determinants (e.g. SES) on the heritability of verbal ability(84).
Parenting, and Family and Household Functioning
Twenty-one studies investigated the effects of household chaos on parenting and family functioning(4, 10, 45, 75, 85-101), of which five focused on the role of household chaos as a mediator or moderator. Outcomes assessed included parenting, parent-child interactions, discipline, sibling relationships, parental response and reactions to child behaviours, family dysfunction, and food insecurity. Household chaos was associated with increased parent-child conflict, decreased parent-child closeness, decreased supportive parenting, decreased positive parenting, and increased negative parenting(87), in addition to less favourable co-parenting, and less emotional availability at bedtime(98). A chaotic home environment was also shown to be associated with less responsive and less stimulating parenting(88), less effective parental discipline(10), greater non-supportive responses to children’s emotions and fewer supportive responses(75, 93), and greater paternal hostility(94). Greater chaos in the home was associated with increased odds of household members disturbing the efforts of adolescents to fall asleep, and decreased the odds of adolescents reporting that nothing was keeping them awake or making it difficult to sleep(96). Finally, cumulative family disorganisation, but not cumulative family instability, was indirectly associated with children’s representation of family dysfunction in drawings, through parenting behaviours (100).
Food insecure households were more likely to have greater household chaos scores compared with food secure households, even after controlling for education and marital status, with higher chaos homes reporting less planning around mealtimes(90). In another study, high and medium household chaos homes were more likely to experience low or very low food security compared with low chaos homes(95). Further, low chaos predicted greater availability of fruits and vegetables in the home and more family meals, while high chaos was a significant predictor of food insecurity risk and greater availability of salty and fatty snacks(101).
The relationship between household chaos and parent outcomes, including maternal executive function, parent sleep, parent feeding behaviours and weight status, and maternal depression, were assessed in six studies(12, 101-105), of which two studies included sample sizes of less than 50 participants(102, 104), and all of which were conducted in the early childhood setting. Household chaos was associated with poorer maternal executive function (n=153), although the modest effect of household chaos overlapped with the effects of co-varying factors, including SES and verbal ability(12). In parents of children aged 18 months to 5 years (n=44), household chaos was significantly associated with fat intake and high serum cortisol levels; however, the relationship between household chaos and fat intake appeared to be somewhat mediated by cortisol levels, although this relationship was not significant(102). In mothers of children aged 2-5 years (n=550), high chaos was associated with greater engagement in emotional and disinhibited eating, while mothers in low chaos households were more likely to be adventurous eaters(101).
Household chaos was lower in mothers with trajectories of low-stable levels of depression compared with moderate-increasing levels of depression, but not in those with remitting depression(103). In mothers of infants(104), household chaos was strongly correlated with measures of maternal depression, sleep, wake disturbances, and fatigue. Similarly, mothers and fathers of infants in high chaos homes demonstrated greater variability in sleep duration compared with low chaos families, while parental sleep fragmentation mirrored that of the child in low chaos homes, where fragmented sleep decreased for both the parents and child over the course of the first year(105).
Eight studies investigated the role of household chaos in cortisol and autonomic nervous system activity, inflammatory profiles, and gut biomarkers(60, 106-112), with sample sizes ranging from 32 to 1292 participants. For stress physiology, household chaos in early childhood was associated with (1) a blunted diurnal cortisol slope in middle school(112), (2) cortisol levels in 7 year olds who had lower levels of resting heart rate variability(106), (3) stable morning cortisol levels in 6 year olds(60), and (4) lowered morning cortisol levels in 3-4 year olds(108). In 13-16 year olds(111), household chaos was associated with increased systemic inflammation, interleukin-6 (IL-6) levels, and C-reactive protein levels, although the relationship between chaos and systematic inflammation and IL-6 levels was moderated by SES. In participants aged 18-66 years, household chaos was not related with hair cortisol levels(110) or gut-derived biomarkers associated with appetite and regulation(109).
Physical Health, Health Behaviours and Communication Disorders
Overall, 19 analyses were conducted that investigated the role of household chaos on health and health behaviours: (1) 9 papers specifically looked at physical health outcomes, disease, and communication disorder outcomes(82, 101, 108, 113-118), including glycaemic control, child health, weight status, and stutter, (2) 3 studies investigated diet and dietary behaviours(101, 119, 120), (3) 6 studies looked at sleep(13, 52, 96, 105, 121, 122), and (4) 2 studies assessed other outcomes, including TV viewing behaviours in a laboratory setting(123) and mothers’ perceptions on children’s physical activity(124). For glycaemic control, both maternal and paternal household chaos scores were positively associated with HbA1c in children and adolescents with type 1 diabetes(114) and in children aged 1-13 years with type 1 diabetes mellitus(116). Chaotic homes were also linked with low cortisol levels (hypocorticolism) in children aged 3-4 years, which in turn predicted overweight in girls, both directly and indirectly through the mediating role of satiety responsiveness, and in boys, indirectly through the mediating role of emotional overeating(108). In infants followed during the first year of life, household chaos was also found to significantly predict weight-for-height z-scores, even after controlling for possible confounders(118).
Chaos was found to be associated with maternal feeding goals(120), with lower household chaos associated with more positive maternal feeding goals, such as promotion of child autonomy around eating. Greater household chaos was also shown to be associated with greater consumption of sugar-sweetened beverages in preschoolers(101). Further, chaos was associated with mother-report of general child health, even after controlling for SES, maternal health status, and family structure(115). Similarly for older children, self-report of household chaos in 10-year-olds was associated with worse physical health two years later(82).
With respect to sleep, a more chaotic home environment was associated with mixed effects on sleep outcomes in adolescents, including sleep onset latency and sleep duration(121, 122); however, in younger children greater scores for household chaos were significantly associated with higher parent-reported scores for bedtime resistance, sleep anxiety, and total sleep problems(13). One study found that infants from highly chaotic homes demonstrated delays in sleep consolidation patterns and greater fragmentation of sleep; however, they also reported longer and more variable sleep duration compared with infants in low chaos homes(105). The authors suggested that this unexpected difference in sleep duration may have reflected a higher quality of sleep in the low chaos households, where sleep was less fragmented and bedtimes and wake times less variable. Chaos in the home environment was not found to be associated with maternal perceptions about physical activity in children(124), stutter severity in children with stutter(117),TV viewing behaviours (i.e. looking patterns) in a laboratory setting(123), or eating in the absence of hunger in low-income toddlers(119).
Mediation and Moderation Analyses
Table 5 presents results from studies where mediation and moderation analyses were conducted. Household chaos was consistently found to mediate the relationship between predictors of adverse child outcomes. Importantly, a number of studies demonstrated the mediating role of chaotic homes between SES and outcomes(14, 26, 107), suggesting that the adverse effects of low SES on child outcomes may, at least in part, be mediated by the effects of household chaos. Further, chaotic environments were also shown to moderate the relationship between several predictors of adverse children and family outcomes, including SES(50), parenting(24, 45, 58), and parental executive function(40), whereby the effects of these predictors were more pronounced in highly chaotic homes. A number of studies also demonstrated the mediating role of parenting on the relationship between household chaos and adverse child outcomes (Table 6)(6, 76, 85, 91), where adverse parenting behaviours may partially explain the relationship between chaos and child outcomes.