NYC Health Opinion Poll: Methodology

The topics covered vary by polling wave.

HOP1

The first wave (HOP1) of polling explored the following topics:

• Naloxone awareness and knowledge
• Awareness and knowledge of medical treatment options for opioid addiction
• Exposure to local naloxone, buprenorphine and methadone media campaigns
• Barriers to treatment for mental health problems
• Knowledge of mental health parity in insurance
• Stigma about mental illness
• Awareness of mental health treatment options in NYC

HOP2

The second wave (HOP2) of polling explored the following topics:

• Expectations for antibiotic prescriptions
• Knowledge of antibiotic effectiveness and antimicrobial resistance
• Emergency preparedness among people who are dependent on medical devices
• Permanent inking or microneedling procedures, removal and experiences of adverse events
• Opinions about public use of marijuana and operationalizing legalization

This survey samples New Yorkers from five opt-in online panels across NYC. Online panels are made up of volunteers who were previously recruited and who have agreed to regularly complete surveys online. Panelists receive incentives. They are recruited online, from direct marketing lists and through affiliate partners, such as retailers and non-profits, in a non-probability sampling procedure (i.e., they are not randomly selected from the entire population of NYC).

The benefits of online panels include:

• Data collection is faster, more efficient and less costly than phone or mail surveys because participants have already consented and demonstrated interest in completing surveys.


• The polls are self-administered, minimizing social desirability bias and concerns about sensitive questions.


• Previously collected panel data provide demographic data and information about the panelists, minimizing participant fatigue at each polling wave.

Each participant consents to join the panels through a double opt-in process. The panel management company then verifies the authenticity of each panelist.

Panel members are invited to complete NYC HOP surveys based on their geodemographic characteristics, so that the sample of participants closely matches that of the NYC population based on their age, gender, race/ethnicity and borough. Using data from the 2013-2017 US Census Bureau’s American Community Survey 5-Year estimates, each poll is fielded for the following minimum quotas:

Variable	Minimum Quota
Sex
Male	46%
Female	51%
Age
18 to 24 years	11%
25 to 34 years	24%
35 to 54 years	34%
55 to 64 years	14%
65 years and older	17%
Race/Ethnicity
Non-Hispanic White	32%
Non-Hispanic Black	22%
Latino/Hispanic	29%
Non-Latino Asian or Pacific Islander	13%
Other	6% (maximum)
Borough
Bronx	17%
Brooklyn	31%
Manhattan	19%
Queens	27%
Staten Island	6%




About 1,200 adults complete each NYC HOP wave.

Data were collected for the first two waves of NYC HOP concurrently, from April 16 to May 13, 2019, (HOP1) and from April 16 to May 10, 2019 (HOP2).

Polls are self-administered online by a sample of panelists recruited across a variety of demographic and socioeconomic indicators. When a poll is available, selected panelists are invited to complete it by email. Non-responders receive up to two follow-up emails to improve response rates. Participants can complete the polls on any Internet-enabled device.

Panelists receive a small incentive for their time, equivalent to about $1.50 to $2.50 per 10 minute survey, in a variety of forms depending on the panel. Incentives include cash, digital currency, points to an online store, gift cards and donations to charities.

Questionnaires are available in English and Spanish. The average length of the questionnaire was 15 minutes for HOP1 and 12 minutes for HOP2.

The survey platform uses digital fingerprinting and bot detection software to prevent duplicate respondents and fraud. After data collection is complete, the vendor evaluates individual responses for quality by searching for indicators of poor quality responses, such as speeding and straight-lining. Poor quality data are removed from the dataset and replaced.

Prior to analysis, weights are applied to each record to produce population estimates. Topline frequency distributions and cross-tabulation analyses are compiled into a public report for each wave. You can request analytic datasets with a Data Use Agreement by completing the EPI data request form.

The online panels are assembled using non-probability methods, so they may not be representative of the general population of NYC in the following ways:

• Respondents may be fundamentally different from the general population in some characteristics.
• NYC HOP does not capture people who do not have regular access to the internet.
• NYC HOP does not include people who cannot comfortably respond to a survey in English or Spanish.

Since a person’s probability of being sampled is unknown, there is no simple way to calculate and report the margin of sampling error.

In contrast to data collected using probability-based samples, studies based on opt-in panels cannot calculate response rates. Rather, participation rates — defined as the proportion of panelists who were invited to take part in the survey who ultimately completed the survey — are reported.

The sample size (completed interviews) and participation rates are provided for each polling wave in the table below.