Individual Random Effects Research Articles

Individual random effects model for differences in trait distribution among respondents

The homogeneity hypothesis is a common assumption in classic measurement. However, the item response theory model assumes that different respondents with same ability have the same option probabilities, which may not hold. The aim of this study is to propose a new individual random effect model that accounts for the differences in option probabilities among respondents with same latent traits by using within-person variance. The performance of the new model is evaluated through simulation studies and real data using the PRESUPP scale of PISA. The model parameters are estimated by the MCMC method. The results show that the individual random effect model can provide more accurate parameter estimates and obtain a scale parameter to describe the distribution of respondents’ abilities, under different within-person variances. The new model has lower RMSE and better model fit than the classic IRT model.

Telemedicine and Disparities in Mental Health Service Use Among Community-Dwelling Older Adults With Alzheimer Disease and Related Dementias

ObjectiveTo examine disparities in mental health (MH) service utilization, via in-person and telemedicine (ie, tele-MH), by individuals' race, ethnicity, and community socioeconomic status, among community-dwelling older adults with Alzheimer disease and related dementias (ADRD) before and after the expansion of the Centers for Medicare and Medicaid Services’ (CMS’s) telemedicine policy. DesignObservational study. Setting and ParticipantsA total of 3,003,571 community-dwelling Medicare beneficiaries with ADRD between 2019 and 2021 were included in the study. MethodsMultiple national data were linked. The unit of analysis was individual-quarter. Three outcomes were defined: any MH visits (in-person or tele-MH), in-person MH visits, and tele-MH visits per quarter. Key independent variables included individual race and ethnicity, the socioeconomic status of the community, and an indicator for the implementation of the telemedicine policy. Regression analyses with individual random effects were used. ResultsIn general, Black and Hispanic older adults with ADRD and those in socioeconomically deprived communities were less likely to have MH visits than white adults and those from less-deprived communities. In-person and tele-MH visits varied throughout the pandemic and across subpopulations. For instance, at the beginning of the pandemic, white, Black, and Hispanic older adults experienced 5.05, 3.03, and 2.87 percentage point reductions in in-person MH visits, and 3.53, 1.26, and 0.32 percentage point increases in tele-MH visits (with P < .01 for racial/ethnic differences), respectively. During the pandemic, the increasing trend in in-person MH visits and the decreasing trend in tele-MH visits varied across different subgroups. Overall, racial and ethnic differences in any MH visits were reduced, but the gap in any MH visits between deprived and less-deprived communities doubled during the pandemic (P < .01). Conclusions and ImplicationsTelemedicine may have provided an opportunity to improve access to MH services among underserved populations. However, although some disparities in MH care were reduced, others widened, underscoring the importance of equitable health care access strategies to address the unique needs of different populations.

Determinants of injuries and accidents in Ethiopia: A Multilevel Analysis

Background: Injury is a major health issue worldwide, especially in developingcountries such as Ethiopia, where no comprehensive national injury data exists.There is a need to better define the epidemiology of injury and determinants asa basis for the formulation of violence and injury prevention strategies.Objective: This study aimed to determine individual and community-levelfactors associated with injuries and accidents in Ethiopia.Methods: Using the latest nationally representative data obtained fromEthiopia Demographic and Health Survey (EDHS) 2016, the survey collectedinformation about the occurrence of injuries in the past 12 months among 12,841 members of 16,650 households. Sample from 645 clusters of 12372 un-weighted households and 12841 weighted households were selected from nine regions and two city administrations of Ethiopia using a stratified clustersampling procedure. Multilevel logistic regressions were used to identify thedeterminants of injuries. Four models were built to estimate both fixed effectsof the individual and community-level factors and random effects betweencluster variation on injuries and/or accidents.Results: The current study found that 3.4 % of injuries and accidents occurredin the reference 12-month period. At the final best-fitted model (model IV), asecondary educational level of household members (AOR= 0.54; 95 % CI (0.297,0.987), household members from middle-level wealth index (AOR = 1.72; 95 %CI (1.048, 2.807), and history of smoking within the last 24 hours (AOR = 0.28;95 % CI (0.092, 0.863) were significantly associated with injury and accidents.Conclusion: The study's analysis of injuries that occurred during the study'sreference 12-month period revealed that Ethiopia has low injury prevalence,considering Ethiopian demographic survey data limitations on major injuries.Additionally, having a secondary education, falling into the middle of thewealth index, and having a history of smoking were all significantly associatedwith injuries and accidents.

The multidimensional Mundlak estimator

Mundlak (1978) shows that the fixed effects estimator is equivalent to the random effects estimator in the one-way error component model once the random individual effects are modeled as a linear function of all the averaged regressors over time. In the spirit of Mundlak, this paper shows that this result also holds for the multidimensional error component model. This is a generalization of Baltagi (2023) from the two-way Mundlak model to higher order multidimensional error components model, see Balazsi et al. (2024a) for the multidimensional fixed effects model and Balazsi et al. (2024b) for the multidimensional random effects model. The F test suggested by Mundlak (1978) to test for this correlation between the random effects and the regressors generate Hausman (1978) type tests that are easily generalizable to the multi-dimensional Mundlak regression.

Factors Influencing Alcohol Consumption in Russia

The main purpose of the study is to determine the differences in factors influencing alcohol consumption in Russia (among urban and rural residents).The first part of the article provides the author's assessment of studies in which place of residence is considered as a specific factor influencing alcohol consumption. Attention is paid to works that focus on differences in factors affecting alcohol consumption depending on the place of residence.The main section of the article contains a description of the empirical testing of the hypotheses based on statistical data for 2012–2021 from the Russian Longitudinal Monitoring Survey (RLMS) NRU HSE. The author used panel logistic regression with random individual effects for the probability of alcohol consumption and a panel Tobin model for the volume of alcohol consumption. The study revealed differences in the degree of influence of the level of education and religiosity, marital status and the presence of children on the level of alcohol consumption among urban and rural residents. Thus, a higher level of religiosity reduces the likelihood of alcohol consumption by rural residents; the volume of alcohol consumption decreases with a rise in the level of religiosity in the city. It is also shown that the presence of children in families reduces per capita alcohol consumption only in villages.According to the author, the findings help make more effective and targeted recommendations for implementing state alcohol policy.

On the Nexus between Economic growth and Environmental Degradation in 28 Countries Classified by Income Level: A Panel Data with an Error-components Model

This paper aims to examine the relationship between economic growth and environmental degradation in a sample of 28 countries. A panel data model is used dividing the sample into four groups classified by income level where fixed and random individual effects are estimated without the time component using an error-components model (ECM). Annual data from the World Bank for carbon dioxide emissions per capita and gross domestic product per capita are used for the period 1970-2016. The empirical results and their graphical analysis, using a panel data approach with an ECM, suggest an absence of an Environmental Kuznets Curve (EKC) in the whole sample. These results considering panel data with an ECM, 28 countries and 46 years differ from many of the studies that support the existence of an EKC in similar samples.

Modeling Longitudinal Outcomes in a Small Matched-Pair Sample Motivated by Cardiovascular Data: A Simulation Study

Introduction: In cardiovascular research, matched pairs with longitudinal outcomes are used to assess time and treatment effects on relevant cardiovascular parameters of interest. Generalized estimating equations (GEE) with independent working correlation are commonly employed for unbiased estimation, but their consistency depends on large sample properties. This study investigates the validity of independent GEE for small sample matched pair data under various working correlations, comparing the results with quasi-least squares (QLS) estimates through simulation. Methods: We simulate a hospital cohort with longitudinal outcomes for two exposure groups and individual random effects. After obtaining the propensity score-matched sample, we performed different simulation scenarios across various mortality rates and the random effect. We compare results from GEE and QLS methods under independence, exchangeable, and AR1 working correlation structures. Results: The independence structure often yields a wider range of relative biases and higher standard errors when there is considerable drop-out. Conversely, the exchangeable structure appears as the true correlation structure, providing more accurate and reliable estimates. No significant discrepancies are observed between the results generated by GEE and QLS methods in this study, and the performance of both approaches is significantly influenced by mortality rates and the standard deviation of the random intercept. Conclusion: Our findings suggest that GEE with an independent working correlation structure is misspecified due to incorrect correlation assumptions and is less efficient. Therefore, correctly specifying the working correlation structure is important for small sample data.

Assessing the Use of GEE Methods for Analyzing Continuous Outcomes from Family Studies: Strong Heart Family Study

Background: Because of its convenience and robustness, the generalized estimating equations (GEE) method has been commonly used to fit marginal models of continuous outcomes in family studies. However, unbalanced family sizes and complex pedigree structures within each family may challenge the GEE method, which treats families as clusters with the same correlation structure. The appropriateness of using the GEE method to analyze continuous outcomes in family studies remains unclear. In this paper, we performed simulation studies to evaluate the performance of GEE in the analysis of family study data.&#x0D; Methods: In simulation studies, we generated data from a linear mixed effects model with individual random effects. The random effects covariance matrix is specified as twice that of the pedigree matrix from the Strong Heart Family Study (SHFS) and other hypothetical pedigree structures. A Bayesian approach that utilizes the pedigree matrix was also conducted as a benchmark to compare with GEE methods with either independent or exchangeable correlation structures. Finally, analysis with a real data example was included.&#x0D; Results: Our simulation results showed that GEE with independent correlation structure worked well for family data with continuous outcomes. Real data analysis revealed that all GEE and Bayesian approaches produced similar results.&#x0D; Conclusion: GEE model performs well on continuous outcome in family studies, and it yields estimated coefficients similar to a Bayesian model, which takes genetic relationship into account. Overall, GEE is robust to misspecification of genetic relationships among family members.

Optimal designs for prediction of random effects in two-groups models with multivariate response

In this work an analytical solution is proposed for optimal designs for the prediction of individual random effects and the group difference in two-groups models with multivariate response. The solution is given by optimality conditions for approximate designs. In particular two-groups models with the same regression function for both groups, Bayesian optimal designs are optimal for the prediction of the group difference. The results are illustrated by examples of linear and bi-linear regression.

Professional and Social Statuses as Predictors of Alcohol Consumption by Working Russians

The paper analyses alcohol consumption patterns among working residents of Russia, depending on their professional and social statuses, using statistical methods. The study is based on data from the Russian Longitudinal Monitoring Survey of the National Research University Higher School of Economics for the period of 2013–2019. Sampling included working women aged 16 to 55 and men aged 16 to 60.The introductory part of the article covers the main trends in alcohol consumption among working Russians with different professional and social statuses. The main part of the article presents the results of an empirical study of the statistical relationship between the characteristics of an employee's professional and social statuses and indicators characterizing the probability and volume of consumption of specific alcoholic beverages (beer, wine, vodka, etc.) and alcohol in general. An econometric analysis of the relationships was carried out based on a random effect panel logistic model and an individual random effect Tobin model. Statistical estimates were made separately for men and women. The regression analysis helped the authors find an increase in the probability of consuming alcohol in general, as well as wine and spirits (cognac, whiskey, liqueurs) for workers with the highest professional status (legislators; high-ranking officials; senior and mid-level executives), as well as women with mid-level qualifications compared to unskilled workers. At the same time, the authors recognized a downward trend in the probability of vodka consumption for the respondents of both gender groups with the growth of their professional status (most pronounced for women). In addition, with an increase in the professional status of an individual, there was a trend towards a decrease in the volume of alcohol consumption in general and by specific beverages except for wine. The volume of wine consumption by men from the categories of "clerks and workers in trade and services" and "mid-level professionals; officials" exceeded the value of the corresponding indicator for unskilled workers. An analysis of the statistical relationship between the social status of working Russians and their predisposition to consume alcoholic beverages revealed an increase in the probability of consumption of selected alcoholic beverages among representatives of both gender groups and the probability of alcohol consumption in general (among men) with an increase in the social status of an individual with more moderate consumption compared to workers with lower social status (for men).According to the authors, the results of this study can be relevant to the elaboration of specific alcohol policy measures aimed at reducing alcohol consumption among the employed population of our country.

Clinical trajectories of individuals with severe mental illness continuing and discontinuing long-acting antipsychotics: a one-year mirror-image analysis from the STAR Network Depot study

Evidence on long-acting antipsychotics (LAIs) in unselected populations with severe mental illness is scant. In this mirror-image study, we compared multiple clinical outcomes 1 year before and after a first LAI prescription in adults with severe mental illness, describing clinical trajectories of LAI continuers and discontinuers. We compared LAI continuers and discontinuers through Mann–Whitney U test, Kaplan–Meier survival curves, regression for interval-censored data, and a maximum-likelihood mixed-model with individual random-effect and time as predictor. Of the 261 participants analyzed, 71.3% had schizophrenia-spectrum disorders, and 29.5% discontinued the LAI before 1 year. At baseline, LAI discontinuers had a shorter illness duration, lower attitude and adherence scores. The mirror-image analysis showed reduced hospital admissions only for LAI continuers. Over time, continuers spent less days hospitalized, but had more adverse events and more antipsychotics prescribed, with higher overall doses. In conclusion, this study shows that LAIs might be beneficial in unselected patient populations, provided that adherence is maintained. LAI continuers spent less time hospitalized, but received more antipsychotics and suffered from more cumulative adverse events over time. Therefore, the choice of initiating and maintaining a LAI should be carefully weighed on a case-by-case basis.

Optimal designs for prediction in random coefficient regression with one observation per individual

The subject of this work is random coefficient regression models with only one observation per observational unit (individual). An analytical solution in form of optimality conditions is proposed for optimal designs for the prediction of individual random effect for a group of selected individuals. The behavior of optimal designs is illustrated by the example of linear regression models.

Gender Differences in the Relationship Between Labour Market Outcomes and Well‐being*

The COVID‐19 pandemic has had widespread impacts on the labour market and on individuals' well‐being. This paper examines gender differences in the relationship between labour market outcomes and well‐being over the pandemic. We draw on longitudinal data from the ANUpoll survey that has been conducted before COVID‐19 and at several points throughout the pandemic. We find higher levels of well‐being among those who are employed relative to most other labour market outcomes, controlling for individual random effects. These gaps in well‐being are greatly reduced or disappear after controlling for financial circumstances and health. We also observe gender differences in the relationship between well‐being and labour market outcomes, in particular: (i) life satisfaction is more closely related to labour market outcomes for females relative to males, and (ii) there is a stronger association between being unemployed and not actively looking and serious mental distress for males relative to females. Well‐being tends to be lower during lockdowns and during the Omicron wave in January 2022, particularly among females. Our findings have implications for ongoing and future pandemic policies. Where restrictions that reduce employment are necessary, we show that during a pandemic financial support has the potential to reduce or in some cases potentially close the gap in well‐being between those who are employed and those in other employment activities.

The two-way Mundlak estimator

Mundlak shows that the fixed effects estimator is equivalent to the random effects estimator in the one-way error component model once the random individual effects are modeled as a linear function of all the averaged regressors over time. In the spirit of Mundlak, this paper shows that this result also holds for the two-way error component model once the individual and time effects are modeled as linear functions of all the averaged regressors across time and across individuals. Wooldridge also shows that the two-way fixed effects estimator can be obtained as a pooled OLS with the regressors augmented by the time and individual averages and calls it the two-way Mundlak estimator. While Mundlak used GLS rather than OLS on this augmented regression, we show that both estimators are equivalent for this augmented regression. This extends Baltagi’s results from the one-way to the two-way error component model. The F test suggested by Mundlak to test for this correlation between the random effects and the regressors generate a Hausman type test that is easily generalizable to the two-way Mundlak regression. In fact, the resulting F-tests for the two-way error component regression are related to the Hausman type tests proposed by Kang for the two-way error component model.

Hierarchical random‐effects model for the insurance pricing of vehicles belonging to a fleet

SummaryWe propose a count‐data model with hierarchical random effects for the posterior insurance ratemaking of vehicles belonging to a fleet, by allowing random effects for the fleet, the vehicles, and time. We derive a simple closed‐form ratemaking formula based on a hierarchical random‐effects specification. We estimate the corresponding econometric model and compute insurance premiums according to the past experience of both the vehicle and the fleet. Our model can be used in other count‐data applications with random individual and common effects on events involving many agents having activities with a principal in a hierarchical principal–agent environment, such as in education, health care management, finance, and business firms.

Using Vocal Characteristics To Classify Psychological Distress in Adult Helpline Callers: Retrospective Observational Study

Elevated psychological distress has demonstrated impacts on individuals' health. Reliable and efficient ways to detect distress are key to early intervention. Artificial intelligence has the potential to detect states of emotional distress in an accurate, efficient, and timely manner. The aim of this study was to automatically classify short segments of speech obtained from callers to national suicide prevention helpline services according to high versus low psychological distress and using a range of vocal characteristics in combination with machine learning approaches. A total of 120 telephone call recordings were initially converted to 16-bit pulse code modulation format. Short variable-length segments of each call were rated on psychological distress using the distress thermometer by the responding counselor and a second team of psychologists (n=6) blinded to the initial ratings. Following this, 24 vocal characteristics were initially extracted from 40-ms speech frames nested within segments within calls. After highly correlated variables were eliminated, 19 remained. Of 19 vocal characteristics, 7 were identified and validated as predictors of psychological distress using a penalized generalized additive mixed effects regression model, accounting for nonlinearity, autocorrelation, and moderation by sex. Speech frames were then grouped using k-means clustering based on the selected vocal characteristics. Finally, component-wise gradient boosting incorporating these clusters was used to classify each speech frame according to high versus low psychological distress. Classification accuracy was confirmed via leave-one-caller-out cross-validation, ensuring that speech segments from individual callers were not used in both the training and test data. The sample comprised 87 female and 33 male callers. From an initial pool of 19 characteristics, 7 vocal characteristics were identified. After grouping speech frames into 2 separate clusters (correlation with sex of caller, Cramer's V =0.02), the component-wise gradient boosting algorithm successfully classified psychological distress to a high level of accuracy, with an area under the receiver operating characteristic curve of 97.39% (95% CI 96.20-98.45) and an area under the precision-recall curve of 97.52 (95% CI 95.71-99.12). Thus, 39,282 of 41,883 (93.39%) speech frames nested within 728 of 754 segments (96.6%) were classified as exhibiting low psychological distress, and 71455 of 75503 (94.64%) speech frames nested within 382 of 423 (90.3%) segments were classified as exhibiting high psychological distress. As the probability of high psychological distress increases, male callers spoke louder, with greater vowel articulation but with greater roughness (subharmonic depth). In contrast, female callers exhibited decreased vocal clarity (entropy), greater proportion of signal noise, higher frequencies, increased breathiness (spectral slope), and increased roughness of speech with increasing psychological distress. Individual caller random effects contributed 68% to risk reduction in the classification algorithm, followed by cluster configuration (23.4%), spectral slope (4.4%), and the 50th percentile frequency (4.2%). The high level of accuracy achieved suggests possibilities for real-time detection of psychological distress in helpline settings and has potential uses in pre-emptive triage and evaluations of counseling outcomes. ANZCTR ACTRN12622000486729; https://www.anzctr.org.au/ACTRN12622000486729.aspx.

How sensitive is the compensating income variation to severity and adaptation?

Knowing the monetary value individuals place on health is essential in tackling resource allocation between health and other uses. However, health conditions vary greatly, not only with respect to main characteristics but also by severity and duration. We apply the compensating income variation (CIV) method to data from the Swiss Household Panel, years 2004–2019, to explore the sensitivity of CIV estimates to severity and adaptation across five different health conditions: headaches, back problems, sleep problems, fatigue, and chronic illness. Accounting for income endogeneity in the life-satisfaction equations and adjusting for individual random effects, we found the CIV estimates to range from $3184 for moderate headaches (women) to $100,066 for severe fatigue (men). Individuals with severe conditions needed to be compensated about three times more than those suffering less. Across the five conditions with two severity levels explored for adaptation, individuals only adapted to moderate headaches and severe sleep problems. In conclusion, not only does the value of health conditions vary greatly, but within each health condition its value on average triples when severe condition is reported as opposed to moderate. Adaptation plays a minor role in CIV estimates for the five health conditions explored.

Order does matter: recency effects in tourist satisfaction after multi-destination trips

ABSTRACT Whereas the drivers of tourist satisfaction with a destination have been largely studied, we know little about whether the serial order of the destination in multi-destination trips plays a role in explaining satisfaction. Based on a large longitudinal data set and using ordered probit models, we show that tourists are systematically highly satisfied with those destinations visited later within a multi-destination trip. This finding is conditional on a large set of controls and individual random effects. Interestingly, length of stay at each destination, age and travel party size are found to be significant mediators of the relationship between the serial order of the trip and satisfaction, partially counterbalancing recall bias. The results provide evidence of recency effects in tourist satisfaction recall and have important implications for the interpretation of tourist satisfaction studies.

Histomorphological and biochemical alterations on the liver of Wistar rats following co-administration of NSAIDs (Piroxicam, Diclofenac and Ibuprofen)

Non-steroidal anti-inflammatory drugs (NSAIDs) are an important therapeutic class of drugs widely used to suppress acute or chronic pains and inflammatory diseases such as in rheumatoid arthritis (RA), osteoarthritis (OS) etc. The aim of this study was to investigate the random combination and individual effect of COX-2 inhibitors (NSAIDs) on the liver. This study was conducted using a total of 40 adults Wistar rats. The rats were divided into eight groups of 5 rats each. Group one was the control group, group 2 was given piroxicam (0.29mg/kg) daily, group 3 was given ibuprofen (5.71mg/kg) daily, and group 4 received diclofenac (1.42mg/kg) daily. Group 5 received piroxicam (0.29mg/kg bodyweight) plus ibuprofen (0.58mg/kg), group 6 received piroxicam (0.29mg/kg) plus diclofenac (1.42mg/kg) and group 7 received ibuprofen (0.58mg/kg) plus diclofenac (1.42mg/kg) respectively daily. Group 8 was given Piroxicam (0.29mg/kg bodyweight) plus Ibuprofen (5.71mg/kg) plus diclofenac (1.42mg/kg) daily. Intervention was over a period of 3days. Animals were sacrificed after 24 hours and the liver tissues were excised, some were fixed in 10% neutral buffered formalin for histological studies while others were placed in normal saline for biochemical analysis. The results revealed that there was no significant statistical difference in total bilirubin, ALT and AST when compared to the control. This suggests that NSAIDs may have no significant effects on the liver for this regimen. Histological changes observed were congested portal vein, haemorrhagic bile ducts, distortion of portal triad and sinusoids, and degeneration of hepatocytes. In conclusion, NSAIDs may have harmful effect on the cytoarchitecture of the liver which can lead to liver damage especially when given in high doses and in combination, but however it had no effect on liver function.

Did the COVID-19 Pandemic Affect the Use of Antipsychotics Among Nursing Home Residents With ADRD?

ObjectiveTo examine whether and how the COVID-19 pandemic affected the use of antipsychotics among residents with Alzheimer's disease and related dementias in nursing homes.DesignObservational study based on the Minimum Data Set and Medicare claims.SettingMedicare- and/or Medicaid-certified nursing homes.ParticipantsNursing home residents diagnosed with Alzheimer's disease and related dementias between 2017 and 2020.MeasurementsThe main outcome variable was any antipsychotic use during a quarter. The secondary outcome was certified nursing assistants’ staffing hours per bed per day in a quarter. We categorized nursing homes into quartiles based on the distribution of nursing home racial and/or ethnic composition. To explore the relationship between the COVID-19 pandemic and the frequency of antipsychotic use, we estimated a linear probability model with robust standard errors, individual and facility random effects. We used a similar model for certified nursing assistant hours.ResultsAbout 23.7% of residents with ADRD had antipsychotic uses during the study period. The frequency of antipsychotic use declined from 23.7%–23.1% between the first quarter of 2017 (2017Q1) and the first quarter of 2020 (2020Q1) but increased to 24.8% by the last quarter of 2020 (2020Q4). Residents in all four racial and/or ethnic groups experienced an increase in antipsychotic use during the COVID-19 pandemic, but the extent of the increase varied by race and/or ethnicity. For example, while residents in the very-high minority nursing homes experienced a greater increase in antipsychotic use than did the residents of other nursing homes at the beginning of the pandemic, the increasing trend during the pandemic was smaller in the very-high minority nursing homes compared to the low-minority nursing homes (0.2 percentage points less, p<0.001, based on heteroskedasticity-robust t statistics, t = 3.67, df = 8,155,219). On average, the certified nursing assistant hours decreased from 1.8–1.7 hours per bed per day between 2017Q1 and 2020Q1, and further decreased to 1.5 hours per bed per day by 2020Q4. There was also a decreasing trend in staffing hours across all racial and/or ethnic groups during the pandemic.Conclusions and RelevanceThe COVID-19 pandemic was associated with an increase in the use of antipsychotics among nursing home residents with Alzheimer's disease and related dementias and decreased staffing of certified nursing assistants, especially among nursing homes with a high minority penetration. Future research is needed to explore means for reducing antipsychotic use, particularly in homes with a high penetration of minority residents.