| Household Income | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| Less than $15,000 | 4.6% | 9.7% | 7.7% | 6.2% | |
| $15,000-$24,999 | 5.6% | 7.8% | 7.4% | 5.7% | |
| $25,000-$34,999 | 6.4% | 8.5% | 6.7% | 6.2% | |
| $35,000-$49,999 | 11.1% | 10.7% | 9.7% | 9.8% | |
| $50,000-$74,999 | 19.8% | 18.6% | 17.0% | 15.1% | |
| $75,000-$99,999 | 17.1% | 14.8% | 16.5% | 13.0% | |
| $100,000-$149,999 | 19.8% | 17.2% | 18.9% | 17.6% | |
| $150,000-$199,999 | 8.8% | 7.6% | 8.0% | 8.6% | |
| $200,000 or more | 6.7% | 5.2% | 8.1% | 17.9% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
Participant Demographics
This section analyzes the demographic composition of the final dataset. Unless otherwise noted, all analyses use unweighted data, for an evaluation of how closely the dataset matched the ACS data for the region.
The demographic composition of survey participants changed substantially from 2019 to 2021, in many cases reflecting four changes to study design initiated during the 2021 survey:
- A change to sample strata in order to better target hard-to-survey households;
- Higher incentive offerings in invitations sent to block groups with a greater share of low-income households;
- Higher incentive offerings at the conclusion of the signup survey, for households making less than $50,000 per year; and
- Incorporation of supplemental sampling through community-based organizations (2021) and transit rider lists (2021). Additional community-based outreach was performed in 2023, but are not included in these analyses as the data were not eligible to be weighted.
In 2021 and 2023, the survey had a higher share of lower-income (i.e., 50K or under) households participate in the survey than are found in the region, indicating that oversampling, supplemental sampling and differential incentives were effective.
Across all three waves, the survey had a higher share of one-person and two-person households participate than are found in the region.
To boost sample representation for large households, the 2021 Travel Behavior Inventory implemented incentive structure changes for those households in the final months of fielding: households with four or more members were offered higher incentives after completing the signup survey (when they provided info on household size). Overall, this offering was effective at increasing the conversion rate (thus decreasing attrition) for large households for whom it is more burdensome to complete the survey, but was not sufficient to bring the share of large households in the survey to closer equivalence with ACS percentages.
The weighting process adjusted for this discrepancy, which is commonly seen in household travel surveys. Survey burden is often provided as a reason for lower response rates in large households. Future iterations of the Travel Behavior Inventory might seek to implement substantive changes to the survey itself to reduce burden, rather than try to correct the problem through sampling and incentive structures. For example, the survey currently asks households with children to report children’s trips on one assigned day, instead of all seven days. Large households could be asked to complete only one or two days of the smartphone app survey as a whole, whether or not they have children.
| Household Size | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| 1 person | 37.8% | 40.8% | 40.8% | 26.2% | |
| 2 people | 39.1% | 36.9% | 37.2% | 32.2% | |
| 3 people | 10.1% | 10.8% | 10.2% | 13.9% | |
| 4 people | 8.6% | 7.5% | 7.5% | 12.8% | |
| 5 people | 3.0% | 2.8% | 2.6% | 5.5% | |
| 6 people | 0.9% | 0.9% | 1.1% | 2.0% | |
| 7 or more people | 0.6% | 0.4% | 0.7% | 7.5% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
In 2021 and 2023, the survey had a higher share of zero-vehicle households in the survey than are found in the region. Although zero-vehicle households were not specifically recruited with differential incentives or special address-based sampling, the strategies to reach lower-income households likely played a role in recruiting a greater number of these households over time.
| Household Vehicles | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| 0 (no vehicles in household) | 6.9% | 11.3% | 10.6% | 6.5% | |
| 1 vehicle | 39.1% | 43.3% | 42.8% | 28.4% | |
| 2 vehicles | 38.4% | 34.9% | 35.2% | 38.7% | |
| 3 vehicles | 10.7% | 7.8% | 8.3% | 14.4% | |
| 4 vehicles | 3.3% | 1.9% | 2.0% | 4.4% | |
| 5 or more vehicles | 1.7% | 0.7% | 1.0% | 7.7% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
The oversample of zero-vehicle households is apparent even when broken down by household size. For example, in 2023, 23.5% of one-person households in the survey had no vehicle available, compared to 17.4% of one-person households in the region. Low-vehicle households were similarly oversampled: for example, in 2021 and 2023, the share of four-or-more person households in the survey with only one vehicle available was nearly twice that expected from census estimates.
| Vehicle Sufficiency | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| Zero Vehicles | 6.9% | 11.3% | 10.6% | 6.9% | |
| Insufficient | 11.2% | 14.9% | 14.2% | 33.5% | |
| Sufficient | 81.9% | 73.8% | 75.2% | 59.6% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
The biggest discrepancy between the ACS and the survey data was that 65-74 year-olds were over-represented, while children 5-17 were underrepresented each year. The same factors that led to over-representation among small households likely contributed to this trend. Additionally, over-representation of retired and older adults in surveys is a common phenomenon outside of household travel surveys.
| Age | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| Under 5 | 5.8% | 5.1% | 5.2% | 5.5% | |
| 5 to 17 | 13.2% | 11.2% | 11.8% | 29.4% | |
| 18 to 24 | 3.8% | 6.0% | 5.4% | 7.3% | |
| 25 to 34 | 13.2% | 17.5% | 16.0% | 12.3% | |
| 35 to 44 | 14.3% | 16.1% | 15.7% | 11.3% | |
| 45 to 54 | 11.5% | 10.6% | 12.2% | 11.3% | |
| 55 to 64 | 16.9% | 14.6% | 13.0% | 11.1% | |
| 65 to 74 | 14.4% | 13.6% | 14.7% | 6.9% | |
| 75 or older | 7.0% | 5.2% | 6.0% | 4.8% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
The percentage of Asian participants went up each survey year, moving closer to the ACS percentage.
| Race | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| American Indian or Alaska Native | 0.4% | 0.7% | 0.7% | 0.5% | |
| Asian | 3.2% | 4.8% | 5.1% | 6.2% | |
| Black or African American | 2.5% | 5.1% | 4.2% | 8.1% | |
| Native Hawaiian or other Pacific Islander | <0.1% | <0.1% | 0.2% | <0.1% | |
| Two or more races | 2.6% | 3.1% | 3.1% | 3.2% | |
| White | 88.6% | 83.1% | 84.9% | 77.7% | |
| Other race, ethnicity, or origin | 1.2% | 1.2% | 1.8% | 4.2% | |
| Don't know | 1.4% | 2.0% | NA | NA | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
The percentage of Hispanic or Latino participants increased each survey year, moving toward the ACS percentage.
| Ethnicity | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| Hispanic or Latino | 2.5% | 3.2% | 5.0% | 19.9% | |
| Not Hispanic or Latino | 97.5% | 96.8% | 95.0% | 80.1% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
The distribution of genders was fairly consistent between survey years. Women were slightly over-represented in the survey, consistent with a female bias in many types of surveys.
| Gender* | 2019 | 2021 | 2023** | 2015-2019 ACS | |
|---|---|---|---|---|---|
| Female | 52.8% | 51.9% | 51.9% | 50.4% | |
| Male | 46.6% | 46.6% | 46.5% | 49.6% | |
| Transgender/Non-binary/Other/prefer to self-describe | 0.6% | 1.5% | 1.6% | NA | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |
| *Respondents who identified as a gender other than Male or Female are excluded from this table. | |||||
| **In 2023, the option for 'Female' was changed to 'Female/Woman/Trans woman/Girl', and 'Male' was updated to 'Male/Man/Trans man/Boy' | |||||
The data also show an increase in the percentage of respondents who decline to identify their gender, or identify as something other than Male
or Female.
In 2023, 1.2% of respondents identified as A gender other than singularly male or female
, compared to 0.9% and 0.3% of respondents who identified as Non-binary/third gender
in 2021 and 2019, respectively.
| Gender | 2019 | 2021 | 2023 | |
|---|---|---|---|---|
| Female* | 52.8% | 51.9% | 51.9% | |
| Male* | 46.6% | 46.6% | 46.5% | |
| A gender other than singularly male or female (e.g., non-binary, genderfluid, agender, culturally specific gender) | — | — | 1.3% | |
| Non-binary/third gender | 0.3% | 0.9% | — | |
| Transgender | 0.2% | 0.4% | — | |
| Other/prefer to self-describe | 0.1% | 0.2% | 0.3% | |
| Total | — | 100.0% | 100.0% | 100.0% |
| *In 2023, the option for 'Female' was changed to 'Female/Woman/Trans woman/Girl', and 'Male' was updated to 'Male/Man/Trans man/Boy' | ||||
The distribution of employment was fairly consistent between survey years. Unemployed persons are slightly over-represented in the survey relative to ACS percentages. This is consistent with the over-representation of older adults in the survey, who are more likely to be retired than younger adults.
| Employment | 2019 | 2021 | 2023 | 2015-2019 ACS | |
|---|---|---|---|---|---|
| Employed | 65.8% | 67.0% | 64.6% | 91.8% | |
| Unemployed | 34.2% | 33.0% | 35.4% | 8.2% | |
| Total | — | 100.0% | 100.0% | 100.0% | 100.0% |