Pilot Survey of Marshallese living in Arkansas, Guam, Hawaii, the CNMI and in the Marshalls
Preliminary Results

RMI Press Release

Embassy of the Republic of the Marshall Islands
2433 Massachusetts Avenue, N.W.
Washington, D.C. 20008
Telephone: (202) 234-5414 ~ Facsimile: (202) 232-3236


For Immediate Release
April 01, 2002


Washington, D.C. - The Marshall Islands Embassy in Washington, D.C., has received preliminary statistics from a recently conducted pilot household survey on the Marshallese migrant community in northwest Arkansas.

The pilot survey was conducted at the request of the Embassy and carried out under the auspices of the Office of Insular Affairs (OIA)/Census Bureau Statistical Enhancement Program. The preliminary data covers a sample of 541 migrants in 78 households concentrated in the town of Springdale. The pilot survey was the first of its kind for a US mainland Marshallese community and follows similar but more extensive surveys conducted on the Marshallese communities in the US island areas of Hawaii, Guam and CNMI (in 1997 and 1998) and a pre-census RMI survey conducted one year prior to the RMI Census of Population and Housing (in 1998).

The two primary objectives of the Arkansas pilot survey were to obtain basic housing and population data on the Marshallese community (now considered to be the fastest growing and perhaps largest Marshallese community outside of the RMI) and to make an estimate of the total size of the Marshallese community. While the pilot survey was relatively limited in its scope, survey coordinators were able to estimate that as of late 2001, between 2,000 and 4,000 Marshallese had migrated to and taken residence in Arkansas.

With preliminary data on the Marshallese migrants in Arkansas now available, side-by-side comparative analysis can take place in order to determine how demographic characteristics might differ between Marshallese in the RMI and those who have migrated out, and secondly, how demographic characteristics might differ between migrants in the US island areas (Hawaii/Guam/CNMI) and those in the state of Arkansas.


All three surveys (Arkansas, Hawaii/Guam/CNMI and RMI) were funded by the OIA and utilized U.S. Census Bureau methods and standards. Data collection for all three surveys was done by Marshallese enumerators, supervisors, and office staff. With regards to methodology, however, the three surveys differ in significant ways from each other. The 1998 RMI survey was collected by enumerating 5 randomly selected households in each of 77 geographically-defined enumeration areas on Majuro Atoll, 100 random households on Ebeye, and 50 households on Jaluit. In 1997, for Hawaii and Guam, and 1998 for CNMI, attempts were made to enumerate all Marshallese. While inevitably some Marshallese will be missed using the ñsnowballî method, most of the Marshallese in these three areas were enumerated. The same approach was tried in Arkansas as a pilot project, but with limited funding and logistical challenges, the scope of the project was comparatively small; hence, the data presented here are only impressionistic.

Comparative analysis of the survey data reveal the following noteworthy findings:

Housing Characteristics

Household size. RMI households tend to be more crowded than migrant households, in general. The average household in the 1998 RMI survey has almost 8 persons. These conditions were similar in Arkansas, where the average household has almost 7 members, while the Marshallese households in Hawaii, Guam, and CNMI had fewer than 5 persons, on average.

Building type. Marshallese in the RMI are much more likely to live in a single-family house separated from all other houses than those living in the receiving areas (Arkansas, Hawaii, Guam and CNMI). Almost 9 in every 10 of the households in the RMI lived in a single ñdetachedî house, compared to 3 in 8 of those in Arkansas, and only 1 in 4 in the other three areas. The Marshallese migrants to Hawaii, Guam, and CNMI were the most likely to live in apartments, migrants to Arkansas were less likely to live in apartments, and few of the Marshallese in the RMI lived in apartments.

Tenure. Typically, Marshallese in the RMI do not rent their units- only about 3 percent of the housing units in the 1998 survey were reported as being rented. Renting was the norm, however, among the migrants, with more than two-thirds of the migrantsä units being rented; average rent in Arkansas was about $400 monthly, but data on rent werenät available for the RMI or the other US areas. The US categories of type of ownership donät apply well in the RMI since most land remains communal, and it is difficult to tell whether households are living in a unit ñowned free and clearî rather than ñoccupied without cash rentî. Of those with a mortgage, however, the average monthly payments in the RMI and in Hawaii, Guam and CNMI were about the same, at about $1000, compared to less than $600 in Arkansas. Since incomes were higher in Arkansas, mortgage payments make up a much smaller portion of the total expenses, somewhat of an incentive to live there.

Structure. The average size of housing units for migrants, at 4 rooms, was about one room more than the average for housing units in the RMI. Migrant housing units had an average of two bedrooms. All 78 of the housing units surveyed in Arkansas had complete plumbing ¿ bathtub or shower, toilet, and hot and cold running water ¿ compared to 83 percent of those in Hawaii, Guam, and CNMI, and only 27 percent in the RMI survey. Similarly, all of the Arkansas units had complete kitchens ¿ stove, refrigerator, and sink ¿ compared to 86 percent of those among the rest of the migrants.

Facilities. The presence of air conditioners, telephones, televisions, and automobiles are social indicators. About 70 percent of the housing units in the RMI had no air conditioning, compared to 82 percent in Guam, Hawaii, and CNMI, but only 1 unit in Arkansas. Also, the migrants were more likely to have telephones and television. More than 60 percent of the housing units in the RMI had no telephone compared to 42 percent in Hawaii, Guam, and CNMI, but only 23 percent of those in Arkansas. And, while 14 percent of the Arkansas units had no television, 22 percent of the rest of the migrants had no television, and 61 percent of the 1998 RMI households were in this category. About 8 in every 10 housing units in the RMI had no automobile available compared to about half of the units in Hawaii, Guam, and the CNMI, and less than 1 in every 10 units in Arkansas.

Population Characteristics

Demography. Because both the Arkansas and the 1998 RMI surveys were non-representative samples, their data must be used with caution. Even the censuses of Marshallese in Hawaii, Guam, and the CNMI were probably not complete. Nonetheless, the data show that while slightly more females than males lived in the RMI, and even a larger proportion of the Hawaii, Guam, and CNMI migrants were female, the Arkansas migrants tended to skew male.

The median age among the migrants (more than 20) tended to be about two years higher than among those living in the RMI (about 18). The median age is the age which cuts the population in half (half are older and half are younger than the median). Migrants tend to delay marriage, either because marriage makes migration more difficult, or because these individuals deliberately delay marriage for travel. About one-third of the adult RMI population were never married in 1998, compared to about half of the Hawaii, Guam, and CNMI, compared to almost 6 in 10 of those in Arkansas.

The average number of children born to women of reproductive age, showed, as expected, that the average woman between 15 and 49 years old in the RMI had about 2.8 children compared to 1.7 among the Arkansas migrants, and 1.6 among the rest of the migrants. For the group of women 40 to 49, the end of the reproductive period that gives an idea of completed fertility, women in the RMI had an average of 6.0 children, compared to 5.4 among Arkansas women of these ages, and 4.4 among other migrant women. In general, Marshallese women continued to have high fertility.

Migration. As would be expected, when migrants depart the RMI, when they do marry, they begin to have children born in the United States. Almost 97 percent of RMI population were born in the RMI, compared to 87 percent of those in Hawaii, Guam, and the CNMI, and only 79 percent of those in Arkansas. It is important to note that while some of the non-Marshallese born are non-Marshallese spouses, step-children, and other relatives, children born in the US are automatically US citizens. This fact, in itself, would be incentive to have children in the US. This detail is seen in the data on citizenship, where 11 percent of the Hawaii, Guam, and CNMI migrants were US citizens compared to more than 21 percent of those living in Arkansas ¿ about twice the percentage.

Social characteristics. While religion was not collected in Hawaii, Guam, and CNMI, data from the other two areas show that Assembly of God is ñover-representedî in Arkansas, meaning that people of this religion are more likely than others to migrate to Arkansas. On the other hand, Protestants and Catholics were under-represented in Arkansas compared to in the RMI. It is important not to read too much into these fragmentary data since both surveys were fairly small, and may not be representative of the whole population.

The Marshallese who moved to Arkansas were much more likely than those moving to the other areas to continue to speak Marshallese at home. In fact, the percentage of Marshallese speakers in the RMI and in Arkansas was about the same at around 98 percent. In Hawaii, Guam, and CNMI, about 1 in every 5 Marshallese migrants spoke English at home.

Education. Because many of the Arkansas migrants went there specifically to work in factories, the school age population were less likely to be in school than the other migrants or those continuing to live in the RMI. Only about 38 percent of the 1998 RMI survey population were high school graduates compared to 47 percent of the migrants to Hawaii, Guam, and CNMI, and 65 percent of those in Arkansas. Hence, the migrants present a potential pool of skilled labor for economic development if they were to return to the RMI. Unfortunately, few Marshallese in any of the areas were college graduates ¿ only 1.8 percent in the RMI, 1.5 percent in Arkansas, and only 1.0 percent among the other migrants.

Labor force. Data from the migrant surveys showed that 77 percent of the Arkansas migrants were in the labor force compared to only 33 percent of the other migrants. Also, among those in the labor force, a much larger percentage of those in Arkansas were actually working compared to the other migrants. Only 7 percent of the Arkansas labor force were unemployed (using US standards for measuring unemployment), compared to more than 24 percent of the other migrants, and 31 percent of those in the RMI. Similarly, while 60 percent of the Arkansas migrants worked in the calendar year before the survey, this was true for only 36 percent of the Marshallese in the RMI, and 30 percent of the other migrants.

As would be expected, 94 percent of the Arkansas migrants were in manufacturing industries. The largest category, with more than one-third of all workers, for the other migrants and for the Marshallese in the RMI was ñservicesî which included persons in education and health care. A similar percentage in Hawaii and the RMI were in what would be a combined category of wholesale and retail trade. About 1 in every 8 Marshallese in the RMI were in public administration, which, of course, is not found among the migrants.

About 3 in every 10 workers in the RMI worked for the government, compared to about 1 in 10 in Hawaii, Guam, and CNMI, and almost none in Arkansas.

Income. While the three groups of Marshallese were not enumerated at exactly the same time, so inflation could affect comparisons, the data still show that out-migration ñbenefitsî the financial status of the household. The median income of households in the RMI in 1998 was about $18,000 ¿ the median income is the halfway point at which half the households earned more than $18,000 and the other half earned less than that. This median was higher than what was seen among the Hawaii, Guam, and CNMI migrants, where the median was less than $17,000. However, the Arkansas migrant households had a median of more than $46,000, considerably more than double the median in the RMI and Hawaii, Guam, and CNMI. Similarly, the per capita income in the RMI was only $2,281 compared to $3,241 among non-Arkansas migrants, and $6,691 among the Marshallese in Arkansas. Even this latter value, while 3 times the value for the Marshallese in the RMI, is still significantly below the US average.

Poverty Status. Survey results indicate that while nearly two-thirds (64 percent) of Marshallese families in Hawaii, Guam and CNMI were living below the US poverty line in 1997 and 1998, only about one-third (34 percent) of Marshallese families in Arkansas lived below the poverty line in 2001.

[Politics & Economy]

Bibliographic citation for this document

Republic of the Marshall Islands. (2002). Press Release: Preliminary Results of Pilot Survey of Marshallese living in Arkansas indicate positve outcome URL: http:/marshall.csu.edu.au/Marshalls/html/RMIPressRel/ArkansasStudy_PR2002.html

Dirk H.R. Spennemann, Institute of Land, Water and Society, Charles Sturt University, P.O.Box 789, Albury NSW 2640, Australia.
e-mail: dspennemann@csu.edu.au

© Republic of the Marshall Islands 2002