Assume that users do not speak any English to create an experience that is accessible to the largest audience possible. Do not require users to click through English-language content in order to access content in their preferred language.
The use of flags can be problematic for many reasons, including the fact that languages are not limited by national boundary.
Verify your translations with multiple sources.
During the time I spent at the White House Iniatiative on AAPIs, our process looked something like this:
- Translation via third-party service (contractor)
- Verification/correction by in-house native speaker
- Verification/correction by third-party stakeholder group
- (Optional) Final verification by another native speaker
It is essential that translations are verified by multiple individuals. I've seen plenty of translations come back from contractors that appear to have been translated by machine. In a nutshell: do not accept publish translated documents without verification.
While the translations (step 1) should be created by professionals, your verifiers (step 2-4) don't need to be professional translators: colleagues and community organizations are a good place to start. The Dept of Ed. has an index of in-house employees who fluently speak and write other languages.
The above process should be seen as an MVP product. Ideally, English-language content is created in parallel with other languages from the start.
- LEP.gov
- Identifying LEP populations with data from the American Community Survey
- Census.gov Fact Finder: "Language spoken at home by ability to speak English for the population 5 years and over"
Experts in the field for further discussion;
- NY Service Initiative offers a certification program to bilingual and multilingual individuals to serve as volunteers for public services. This program is a partner of the Mayor's Office of Immigrant Affairs
- Department of Justice Civil Rights Division is responsible for helping agencies ensure compliance with limited english proficiency access laws.
- Marriott Hotels are widely considered to be best practicers of localization and language access. They have dedicated designers located in each region in which they do business. This is discussed in Jared Spool's UIE podcast.
I've found that the best place to find data on Limited English Proficient individuals is the U.S. Census American Community Survey's 3-year estimates.
Data source: ACS 2013 3-year estimate (B16001)
| Language | Population | % of total U.S. population | % of LEP population |
|---|---|---|---|
| Spanish or Spanish Creole | 16258423 | 5.532 | 64.508 |
| Chinese | 1646092 | 0.560 | 6.531 |
| Vietnamese | 846444 | 0.288 | 3.358 |
| Korean | 617636 | 0.210 | 2.451 |
| Tagalog | 524212 | 0.178 | 2.080 |
| Russian | 420281 | 0.143 | 1.668 |
| Arabic | 368744 | 0.125 | 1.463 |
| Other Indic languages | 342037 | 0.116 | 1.357 |
| French Creole | 332787 | 0.113 | 1.320 |
| African languages | 298360 | 0.102 | 1.184 |
| Other Asian languages | 296424 | 0.101 | 1.176 |
| French (incl. Patois Cajun) | 263565 | 0.090 | 1.046 |
| Portuguese or Portuguese Creole | 256254 | 0.087 | 1.017 |
| Polish | 231832 | 0.079 | 0.920 |
| Japanese | 190480 | 0.065 | 0.756 |
| Italian | 184647 | 0.063 | 0.733 |
| German | 169712 | 0.058 | 0.673 |
| Other Indo-European languages | 163448 | 0.056 | 0.649 |
| Other Pacific Island languages | 162366 | 0.055 | 0.644 |
| Persian | 151225 | 0.051 | 0.600 |
| Hindi | 138136 | 0.047 | 0.548 |
| Gujarati | 131569 | 0.045 | 0.522 |
| Other Slavic languages | 122331 | 0.042 | 0.485 |
| Urdu | 119260 | 0.041 | 0.473 |
| Mon-Khmer Cambodian | 112114 | 0.038 | 0.445 |
| Armenian | 108831 | 0.037 | 0.432 |
| Serbo-Croatian | 102000 | 0.035 | 0.405 |
| Hmong | 92827 | 0.032 | 0.368 |
| Thai | 82501 | 0.028 | 0.327 |
| Greek | 75130 | 0.026 | 0.298 |
| Other West Germanic languages | 73255 | 0.025 | 0.291 |
| Laotian | 73115 | 0.025 | 0.290 |
| Other and unspecified languages | 61909 | 0.021 | 0.246 |
| Yiddish | 53104 | 0.018 | 0.211 |
| Hebrew | 35050 | 0.012 | 0.139 |
| Navajo | 33295 | 0.011 | 0.132 |
| Other Native North American languages | 27008 | 0.009 | 0.107 |
| Hungarian | 24831 | 0.008 | 0.099 |
| Scandinavian languages | 12596 | 0.004 | 0.050 |