Limitations
Through use of the shift-share technique it is assumed that the demographic distribution of Latino citizens at the state level is equivalent to the one that exists at the congressional district level for all of the Latino groups studied. While the author is confident about the use of the state data as a good proxy, it does raise concerns. For example, how would the data change in light of the misrepresentation problem, using Dominicans as an example?
The data as calculated is not corrected for misrepresentation and it is difficult to say one way or the other how this error should be applied nationwide. However, there is still a need to understand how correcting for the 25 percent misrepresentation error could affect the findings of this report in terms of population and citizenship data. If the error were equally distributed throughout the country this would have a minimal or negligible effect on the data. The same goes for all the other Latino groups affected by the misrepresentation problem. In the instance that the demographic distribution is more regionally skewed, say, there exists a strong leaning toward the Northeast or toward certain cities, then the relevancy of the data would be affected. In regional cases, there would be significant underestimates generally in congressional districts throughout the northeastern part of the country. Within the same state/city this problem could be reflected as an underestimated number of citizens in some congressional districts with an overestimate in other congressional districts.
Another factor that limits the interpretation of this data is the lack of control for
socioeconomic factors. Again, an equal distribution of income, education, religion,
homeownership, etc. is assumed nationwide. Studies have shown (refer to Section 2)
that when these indicators are regressed with other minority racial or ethnic groups,
the observed clustering of higher social-educational-economical households,
reflective vis-a-vis their community investment
|
through owning a home, possessing a
college degree, and earning higher incomes worthy of a middle-class status, are more
correlated with an individuals’ likelihood to be a registered voter.
Thus, if these indicators were to be taken into account this may affect the data to some extent.
More specifically, the data may be underestimated in congressional districts where Latinos have
a higher income bracket, and be more robust in poorer, less educated Latino districts.
The estimated population projections also suffer from limitations in terms of the degree of approximation it can provide. The logarithmic formula is most accurate for residential populations and typically short time periods, assuming current population trends continue. This is easily observed in our data as well. The difference among the estimates is minimal, across both the base populations and growth rates used, for approximations within the next 10 to 15 years. The difference in estimates then starts to widen with the increasing projections towards 2020. The logarithmic formula 1) fails to include net international migration, 2) is not ethnic specific, and 3) as a population estimation model, starts to break down for long time periods. Nevertheless, this model provides a good first line estimate until better data is available to determine more tailored projections based on ethnic-specific fertility, mortality and mobility rates.
Opportunities
A few of the goals at the beginning of this research effort were to be able to estimate the number of Dominican registered voters, number of Dominican voter turnouts, and Dominican population projections at the congressional district level. Information on the Dominican community is critically needed at the city, state, and national levels. The lack of information hinders the community’s potential for empowerment. This obstacle should be perceived as welcomed challenges and serve as opportunities for Dominicans to grow in terms of their knowledge and influence. These opportunities are discussed below.
|
Estimating Registered Voters
There is no methodology and information source available to obtain a reliable and reasonable estimate of ethnic/Dominican-specific registered voters and ethnic/Dominic-specific turnout rates. Most of the data available are for state, city or county level, and listed under the generic category of Hispanic only, and a few cases with party affiliation breakdowns. Thus, it was not possible to produce nationwide estimates of registered voter population among Dominicans or other Latinos. The safest solution is to consider only the number of adult citizens as a goal to register and mobilize for communities to vote.
Nevertheless, despite not having available congressional district data (with an Latino ethnic breakdown),
the information that is currently available was used to gauge where Dominican-Americans stand in
terms of registered voters and turnout rates. Some information at the congressional district level
does exists for the State of New Jersey.71
The New Jersey Division of Elections is part of the Department of Law and Public Safety
and it has a rich amount of publicly available information. Their website
(http://www.state.nj.us/lps/
elections/vote_doe.html) provides a breakdown of the number
of registered voters within each of the congressional districts for the June 4, 2002
Primary Election and the November 5, 2002 General Election. It also provides the official
election turnout by county.
Using a back-of-the-envelope analysis, an attempt is made to overlap the congressional
district data with the county turnout data. Based on previous findings, a conservative
percentage estimate is used for the number of Dominicans registered voters and their
turnout during the Primary and General election of June 2002 and November 2002, respectively.
A focus was made on the top New Jersey Congressional Districts found on the Top 25 list of
Dominican populated states (Table 5). These districts are congressional district 8, 9, 10 and 13
represented by Rep. William Pascrell, Rep. Steven Rothman, Rep. Donald Payne and Rep. Robert Menendez,
respectively.
|
