The findings for this project are two-fold. Primarily, the findings include data on the gender makeup of elected officials in New Jersey, as well as the racial makeup of officeholders at the congressional, state legislative, and county levels. Unfortunately, and this speaks directly to the second finding in this report, data collection limitations resulting from inadequate support for this effort forced us to exceed the mandate of the legislature in order to produce any meaningful results.
Secondly, the process of creating these datasets yielded important findings about the efforts to conduct this type of data collection and analysis about the population of elected officials in a state. These findings are useful for researchers and government administrators exploring the concept of demographic data collection on government officials.
Women are underrepresented at every level of office in New Jersey. Across all levels of office, women comprise 29.5% of officeholders in New Jersey. This mirrors similar proportions nationally, with women’s representation across most levels of office hovering under a third of officeholders. The New Jersey congressional delegation is at the bottom of national rankings for women’s representation with women comprising only 14.3% of the delegation (two of 14 members.) A total of seven women have ever served as members of Congress representing the state of New Jersey, of those women, only one is a woman of color (Bonnie Watson Coleman). The New Jersey legislature had the highest proportion of women’s representation of all levels, with women comprising 34.2% of its membership. Following closely were county offices at 33.8% and municipal offices at 29.1%. Among mayors of cities with populations over 30,000, the gender disparity grows even wider, with women holding only 15.4% of these positions.
White men are by far the most overrepresented group at every level of office in New Jersey. Among other groups, only Black men and Black women are represented in some elective offices near or barely above their representation in the population. Asian American/Pacific Islander and Latino women and men in New Jersey face the greatest disparities between representation in government and representation in the population.
Overall, white men account for 52.6% of New Jersey officeholders at the congressional, state legislative, and county levels. According to 2021 U.S. Census population estimates, white men account for 27.0% of New Jersey’s population, signaling a vast overrepresentation (25.6%) in officeholding.
White men are the only demographic that are significantly overrepresented in the state’s 14-member congressional delegation. White women are 27.8% of New Jersey’s population while only 7.1% of the delegation. The representation of Black men and Black women is near or slightly exceeds their population in the state. Latinas, Asian American/Pacific Islander women, and Native American/Alaska Native/Native Hawaiian men and women are completely absent at this level of office, leaving nearly a fifth of New Jersey’s population without representation in our nation's capital.
At the state legislative level, once again, only white men are significantly overrepresented. At the municipal level, race data is only available for mayors of cities with populations over 30,000. This level has the highest overrepresentation of white men at 66.7% versus the state population of 27%. It should be noted that 49% of New Jersey’s citizens live in cities over 30,000.
There are also significant partisan differences by race and gender. Across levels for which we have race data, white men and women combined make up well over half (59%) of all Democratic officeholders and over 90% of Republican officeholders. At the state legislative level, 96% of Republican legislators are white. In comparison, more than a quarter (27%) of Democratic state legislators identify as Black, 12.9% identify as Latino/a, and 8.6% identify as Asian American/Pacific Islander. Across all levels of office, 39% of Democratic officeholders and 27% of Republican officeholders are women. In the state legislature, women comprise 40% of the Democratic caucus and 28% of the Republican caucus.
Coding for race in this project follows CAWP’s current method for race data collection. Officeholders who identify as more than one race/ethnicity are included in each group with which they identify. As a result, percentages may not add up to 100% across levels. While these numbers do include mayors of cities with populations over 30,000, they do not include other municipal officeholders in these cities, as expert coding was not possible at this level and the response rate was too low to report findings.
As the project developed, it became clear that this study would become as much about how the data is collected as it is about the data itself. For this reason, we are including this section on process findings.
One key obstacle in creating this dataset was a low survey response rate, and a major contributor to the low response was that officeholders were not legally compelled to participate. In addition, the original version of the legislation passed by the legislature required the secretary of state’s office to sign the request to complete the survey, thus giving the request the weight and authority of the state government. This portion of the legislation was deleted by the governor’s office in a conditional veto. While this provision did not compel officeholders to participate, it did provide an endorsement of the survey that would have likely improved the response rate.
Without the full-throated endorsement of the data collection by the state or, ideally, compulsory participation by elected officials, a self-report approach to collecting demographic information on officeholders will yield insufficient data to fully analyze the diversity of elected officials.
While we are grateful to the many elected officials who replied to our survey and voluntarily provided their information, the participant response rate for officeholders was just 16.1%. By level, response rates varied, with the lowest being at the congressional level (7.1%). The highest response rate by far was among state legislators at 53.3%.2 There are several factors that can account for these differences; for example, responses from state legislators could have been higher because the legislature was the body of government responsible for the legislation that made the survey possible.
Response rates also varied by gender. Of the women surveyed, 21.9% filled out the questionnaire with their demographic information. Conversely, only 13.5% of men did the same, accounting for an 8.5% gender gap in completed surveys.
Limitations of Surveys and Need for Proxy Coding
Another limitation of the existing legislation is its reliance on a survey alone. Surveys are an excellent tool utilized to gather information about various populations, but they also have some practical disadvantages. As noted, survey response rates are typically low, which can create flawed results when used on a finite population such as elected officials. Because participation was voluntary, response rates were low, and CAWP researchers had to employ expert proxy coding – collecting data through public statements, direct contact, and other means – to determine gender codes at all levels and race codes at most levels. Demographic information determined via survey accounts for 13.5% of the total dataset. Expert coding became necessary when it was determined that the response rate for each level would have resulted in an overly weighted sample. For example, if only survey responses were utilized in these findings, then New Jersey’s congressional delegation would be 100% white women. A requirement for and/or guidance on this type of expert coding was not included in the legislative language.
Proxy coding allowed us to report far more extensively on the demographics of New Jersey’s elected officials. While proxy coding is feasible, it is both time-consuming and labor intensive. In addition, this type of coding for personal identity data increases the risk of misidentification. Self-reporting of gender and racial identity is preferred.
In the absence of requiring self-reporting of demographic data, the state can increase the feasibility and sustainability of this data collection by maintaining up-to-date and accurate contact information for all elected officials.
A major challenge in conducting a study like this is the dearth of readily available officeholder contact information in list form. The legislation directed the secretary of state’s office to provide election results. While the secretary of state’s office was helpful and shared what was available, centralized lists of those who currently serve in elected office do not exist at a governmental level in New Jersey. To compile a complete list of all elected officials and their contact information, researchers must source several individual lists, utilizing both government sources and data services companies, and combine the lists, as was done in this project. This is easy at the federal and state legislative levels but becomes much more difficult at county and municipal levels due to the lack of readily available lists and the sheer number of elected officials. The data service company utilized for this study only had information for municipalities with populations of 10,000 and above. Information for smaller municipalities needed to be hand collected by going to each municipality’s website or contacting their town hall.
Additionally, individual contact information for public officials is spotty at best. Existing lists from data service companies and other sources often do not have complete personalized contact information for each officeholder, requiring individual data collection to fill in gaps in the information, assuming it is available. For example, at the municipal and county level, contact information is routinely the same for every member on the council/board of commissioners (such as a council- or commission-wide shared email or one phone number for the entire governing body), or the only electronic contact avenue is a webform, which is incompatible with survey software and makes the dissemination of personalized survey codes impossible. Moreover, individual contact information for municipal officeholders was difficult to collect and, in some cases, not possible. Of the municipal officeholders, 17% had unusable contact information. This was due to a lack of publicly available individual email addresses or phone numbers. For long-term, ongoing sustainability of this project, a centralized database of elected official contact information along with personalized email addresses and phone numbers is necessary.
2 Response rate in this section is calculated using the number of respondents that provided demographic information in their survey response and does not include partial or incomplete surveys. These calculations also do not include the 15.9% of officeholders who did not have sufficient contact information and therefore never received a survey.