What does linguistically isolated mean

Limited English Speaking Households Linguistic isolation is measured by the percentage of households in which no one over the age of 14 speaks only English at home or English "very well" as a second language. Californians speak over 40 languages besides English at home, with Spanish as the most common.

Households with no English speakers often face barriers to resource access as they may have trouble communicating with social service and health care providers. Children in linguistically isolated households often underperform in math and reading when compared to their native-English speaking peers. Neighborhoods with high rates of linguistic isolation often communicate and function fully without use of the English language, providing children with few opportunities to practice and learn English.

Click to open in a new window. Linguistic Isolation What is linguistic isolation? Why is this indicator included in CalEnviroScreen? Californians speak more than 40 different languages at home.

Source: U. Census Figure 2. Percentage of black or African American adults aged 18 years and older, United States. Figure 3. There has also been a corresponding growth in the percentage of U. According to the census, Linguistic isolation is defined by the U. In , approximately 4. Table 1 shows the major languages included in each group. Again, the types of languages spoken within linguistically isolated households across the United States vary by region Figures 4—6. Because most health surveys are typically conducted in English only, linguistic isolation can be expected to significantly increase the level of nonresponse among people who do not speak English.

Figure 4. Percentage of linguistically isolated Spanish-language households, United States. Figure 5. Percentage of linguistically isolated Asian-language or Pacific Island-language households, United States. Figure 6. Percentage of linguistically isolated Indo-European—language households, United States.

Yet over the past decade, BRFSS participation rates, like those of most other surveys, have declined sharply As part of an effort to reverse this trend and ensure the reliability and validity of BRFSS data, we assessed the impact of race, ethnicity, and linguistic isolation on measures of survey participation. This report includes the results of that assessment as well as a discussion of potential means of improving survey participation rates among these groups, thus making population-based health surveys like the BRFSS surveys more representative of the entire population.

However, the three territories are not included in the analysis presented here. To examine aspects of survey participation, we calculated the following six dependent measures of survey participation at the county level based on final case disposition for telephone numbers called between January 1, , and December 31, Resolution rate: the percentage of all sampled telephone numbers for which household status with a working telephone number has been determined.

Screening rate: the percentage of all known households in which the presence or absence of an eligible respondent has been determined. Cooperation rate: the percentage of known, eligible households in which a completed or partially completed interview has been obtained.

Response rate: the percentage of all confirmed and potentially eligible sample members for whom an interview has been completed, which we calculated using Response Rate 4 recommended by the American Association for Public Opinion Research Language-barrier rate: the percentage of all sampled households given a final disposition of language barrier. Interviewers could not communicate with household members because of the language spoken in the home which was presumably not English or Spanish, the two languages in which the BRFSS survey is conducted.

We conducted the analysis at the county level because of a lack of available information about survey nonrespondents at the individual or household level. Counties were included in the analysis if they had 30 or more observations in the denominator of each of the six participation measures.

Our use of these criteria ensured greater stability in the measures calculated and helped us compare the impact of independent variables across the six models estimated; however, it also limited the analysis to of the counties in the United States.

County-level predictor variables for race, ethnicity, and linguistic isolation were derived from U. On average, the included counties had somewhat higher percentages of blacks than did the nonincluded counties 9. We also developed several county-level control variables to account for some of the other factors that are thought to affect participation rates within certain geographic areas.

Socioeconomic status, often measured through a combination of income and education levels, is an important mediator of racial and ethnic health disparities and an important predictor of survey participation 2, Likewise, living in an urban area, being away from home frequently, and screening calls with answering machines, caller-identification devices, or similar devices have been shown to reduce respondent contactability and participation rates 21,38, We used U.

We used BRFSS data to calculate a fifth control variable measuring the percentage of all calls made within a county that resulted in contact with an answering machine, privacy manager, or some other identifiable type of call-screening device.

Because all variables in the analysis were expressed as percentages, we used ordinary least squares OLS regression modeling to assess the impact of race, ethnicity, and linguistic isolation on survey participation. We used separate models to determine which variables race, ethnicity, or linguistic isolation were better predictors, but we found the differences between them to be marginal. Although linguistic isolation is a definite barrier to survey participation, race and ethnicity may or may not be factors; thus, the Asian-language—only and Spanish-language—only variables were retained in the final models, but the variables Asian race and Hispanic ethnicity were not retained.

Because of the strong correlation between race, ethnicity, and language-isolation variables, however, we had difficulty determining the proportional impact of each.

We also examined the possible effects of multicollinearity in our analysis. The final OLS models were estimated for each of the six participation measures rates of resolution, screening, cooperation, response, language barriers, and refusal.

The dependent variables used were the county-level estimates for percentage of black adults, percentage of Spanish-language—only households, percentage of other Indo-European-language—only households, and percentage of Asian-language—only households. Model selection was based on forced entry of all variables into the models rather than stepwise selection. The models were estimated using SPSS Finally, we used the OLS coefficients from the final models and the maximum county-level population parameters to calculate the maximum impact of race, ethnicity, and linguistic isolation on the six measures of survey participation.

In general, minority race and ethnicity and linguistic isolation had significant negative correlations with survey participation rates Table 2. For example, for every percentage-point increase in the black population of a county, the county-level response rate declined by 0.

Counties with higher percentages of black residents tended to have significantly lower rates of participation and higher refusal rates. The percentage of black residents in a county did not have a significant effect on the rate of nonparticipation attributed to a language barrier. Linguistic isolation also had a negative effect on participation rates, although the magnitude of this effect differed across the three language types. The impact of Spanish-language isolation on response rates was more than four times the impact of the percentage of black adults in a county.

Rates of Spanish-language isolation did not, however, significantly affect the percentage of nonparticipation attributed to refusals.

Counties with higher rates of Indo-European-language—only households also had higher language-barrier and refusal rates. In contrast, Asian-language—isolated households had less effect on survey participation rates. Because the impact of race, ethnicity, and linguistic-isolation variables depended on the size of a county subpopulation, we calculated the maximum impact of these variables among the subset of counties examined here.

Table 3 shows the amount of change we might expect in the percentage of each rate in counties with the highest concentrations of black residents and language-isolated households. We calculated this expected change by multiplying the high range value for each population characteristic by its corresponding OLS coefficient from Table 2. Our study revealed that survey participation rates were significantly lower in areas with higher concentrations of racial and ethnic minorities and linguistically isolated households.

These important findings indicate the need to ensure adequate representation of these populations in large-scale health surveys such as the BRFSS. As we examine ways of increasing BRFSS participation rates, these findings will help us to design and implement more effective means of involving these hard-to-reach populations.

One particularly disturbing finding was the significant impact of Spanish-language isolation on participation rates, given that BRFSS surveys are offered in both Spanish and English. Education is an important mediating factor in survey participation among Hispanic individuals because lower levels of literacy and health literacy have been related to a greater reluctance by Hispanic individuals to participate in health surveys 26, Our study shows, however, that even after controls are added for education, areas with higher concentrations of Spanish-only—speaking households are less likely to participate in health surveys.

This may be because of ineffective procedures for contacting and eliciting participation from predominantly Spanish-speaking households, lack of bilingual or Spanish-speaking interviewers, or inadequate training of Spanish-speaking interviewers. It is also likely that current Spanish-language survey translations do not adequately address the different Spanish dialects spoken in the United States, such as those spoken by individuals or families originating from Mexico, Puerto Rico, or Cuba Moreover, it may also reflect the impact of ethnic and cultural issues.

Therefore, we may have to assume that concepts and interpretation are culturally dependent 42, We were unable, however, to disentangle the influence of language and culture.

Our findings also indicate that more needs to be done to improve participation among other minorities, such as African Americans, Asians who are isolated by language, and other linguistically isolated groups. To this end, researchers are investigating ways to address disparities in participation rates by postsurvey adjustments, culturally appropriate data-collection procedures, and multiple language use.

Standard techniques are widely used to compensate for demographic differences between a survey sample and the general population it represents Postsurvey adjustments such as weighting and stratification represent standard practices in most major health surveys. However, these techniques are often limited to a few key demographic variables for which population estimates are available.

Moreover, they may produce larger standard errors that decrease the precision of estimates. Researchers need to develop survey designs that better address the increasingly complex racial, ethnic, and linguistic mix of the U. The report further recommended that relevant cultural factors and language requirements be incorporated into survey designs when feasible. Researchers need to be cognizant of the customs, values, and beliefs of individuals in minority communities, particularly because they relate to the sharing of personal information, including health care practices and health conditions Focus groups and cognitive interviews of people from various backgrounds can help determine whether respondents will interpret and respond to survey requests and questions as intended 45,