Recent Latinx immigrants to Miami/Dade County (FL): Travel patterns before, during, and one year after the COVID-19 pandemic lockdown

Abstract

Introduction:

Understanding the transportation needs of immigrants is crucial for the design and promotion of safe, equitable, and sustainable living environments. This study examines the transportation patterns from a sample of Recent Latinx Immigrants (RLIs) upon arrival to Miami/Dade Co (MDC), Florida.

Methods:

Collected between 2018 and 2021, data came from a longitudinal study examining drinking and driving trajectories among 540 RLIs to MDC. Retrospective pre-immigration data (T0) were obtained simultaneously with the first-year post-immigration data (T1). Follow up surveys were conducted one year later, before (T2-BC) or during a pandemic lockout (T2-DC), and two years later (T3). Descriptive and repeated measures mixed-model regression were used to examine the data.

Results:

Driving declined from T0 to T1, although remained higher than previously reported for other locations. Not having a valid driver’s license was the main reason for the decline. The initial reduction in driving was paralleled by an increase in the use of transit, riding as passengers in private vehicles, and walking. A year later (T2), as RLIs’ income and access to a driver’s license grew, driving rates increased (even during the pandemic lockdown), while the use of other transportation modes decreased. A year after the pandemic lockdown (T3), driving as well as the use of other transportation modes receded. Reasons for this decline are unclear.

Conclusions:

RLIs reported elevated driving rates upon their arrival to MDC. The COVID-19 pandemic seems to have altered the RLIs’ transportation patterns, provoking an overall decline in mobility that lasted even after the pandemic lockdown ceased.

Practical applications:

Transportation planners working on developing safe and equitable transportation systems in MDC should: (1) identify and address barriers to the use of transportation modes other than driving by RLIs; and (2) understand reasons for the broad decline in transportation modes after the pandemic lockdown.

1. Introduction

Understanding the transportation needs and behaviors of immigrants is crucial for the design and promotion of efficient, equitable, and sustainable living environments. Research has shown that immigrants are more likely to use transit, walk, ride a bicycle, and carpool than their U. S.-born counterparts, although such disparity erodes as time in the United States goes by (Burbidge, 2012; Handy et al., 2008; Crane & Crepeau, 1998; Bagley & Mokhtarian, 2002; Blumenberg, 2008; Tal & Handy, 2010). As time in the United States goes by, many immigrants increase their reliance on private vehicles for their mobility—a process also called “transportation assimilation” (Portes & Zhou, 1993). Acculturation as well as improvements in socioeconomic status (SES) and the acquisition of new family obligations were some of the factors mediating and/or moderating such evolution (Smart, 2010; Lovejoy & Handy, 2011; Alba et al., 2000).

Research on immigrants’ transportation needs and usage has been based largely on those who have been in the United States for an extended period of time (Asgari et al., 2017; Barajas, 2019; Barajas et al., 2018). Only a limited number of studies have examined immigrants’ transportation behavior during their initial years in the United States (Smart, 2015; Lovejoy & Handy, 2011; Kim, 2009) and to our knowledge, only three studies examined immigrants’ transportation patterns within their first year of arriving to the United States (Chatman & Klein, 2009; Kim, 2009; Romano et al., 2021). Focusing on recent Latinx immigrants (RLIs) Romano et al. (2021) examined the travel patterns of those who had recently (i.e., within one year) arrived to Miami-Dade County (MDC), Florida. The authors reported that coincidentally with previous research from other hosting locations, RLIs relied on multiple modes of transportation during their first year in MDC, with the modes most frequently used being largely influenced by RLIs’ sex, income, legal residency status, and by transportation behaviors in the country of origin (Romano et al., 2021). Unlike previous studies, however, the sample of RLIs to MDC showed high driving rates during their initial year in the United States (Romano et al., 2021). About 64% of the RLIs have driven a vehicle within one year of arrival to the United States, a figure substantially higher than the 25%–33% of immigrants who reported driving within the first year of arrival to the United States reported by Chatman and colleagues (Chatman, 2014; Chatman & Klein, 2009). The authors argued the RLIs’ initial high driving rates may be indicative of the uniqueness of MDC as a hosting location for Latinx immigrants. MDC is a well-established Latinx immigrant-receiving community with dense ethnic enclaves (Sanchez et al., 2016b; Schwartz et al., 2010b; McGlynn, 2005) that provide support networks that may have increased RLIs’ availability to private transportation soon upon arrival (Romano et al., 2021).

This manuscript extends the examination of the transportation patterns for the sample of RLIs upon arrival to MDC for two more years. Rationale for the present study is fourfold. As indicated, rates of driving among the RLIs during their first year in MDC were relatively high. The first motivation for this manuscript reflects the need to describe this phenomenon in more detail, assessing whether these initially high driving rates remained at that level in the following years. Another related motivation is the need to describe RLIs’ use of transportation modes other than driving a vehicle. Typically, RLI’s are frequent users of transit upon arrival, a mode of transportation from which they shift away as time in the United States passes and access to a vehicle increases (Casas et al., 2004). Our previous examination indicated that the transportation patterns followed by the RLIs upon arrival to MDC differed from those typically reported for other jurisdictions. A second motivation for this study is to assess how the particular initial travel patterns shown by the RLIs to MDC evolve as they spend time in MDC. This information should help social and transportation planners to understand and accommodate the changing transportation needs of the population they serve. The need to examine the evolution of RLI’s transportation patterns in MDCs over time is compounded by the impact the COVID-19 pandemic may have had on RLIs’ travel. On March 12, 2020, the Major of MDC declared the Local State of Emergency (Ordinance No. 20–87), which initiated a lockdown of the MDC population. At that time, about 30% (N = 157) of participants had been interviewed for the second year follow up wave before the COVID-19 pandemic lockdown. The remaining 70% were interviewed during the lockdown. Subsequently, a third motivation of this study is the need to examine the impact of the pandemic lockdown on RLIs’ transportation modes. Finally, rates of immigrants from South American countries such as Venezuela have increased up to 76% to 421,000, while immigrants from Central American countries such as Guatemala have increased by 37% to 1.4 million (Noe-Bustamante et al., 2019). The fourth motivation for this study responds to the need to examine RLIs’ transportation patterns as they may have been impacted by the recent shifts in immigration patterns.

2. Materials & Methods

2.1. Data

Data for this study come from an ongoing National Institutes of Health (NIH)-funded longitudinal study examining pre- to post-immigration travel and drinking and driving trajectories among 540 young adult recent Latinx immigrants to MDC. Data were collected from November 2018 through December 2021. Retrospective pre-immigration data (T0) were obtained simultaneously with the current first-year post-immigration data (T1). The timeframe requested for assessing T0 variables included the last year spent in the country of origin. A follow up survey was conducted one year later, with about 30% (N = 157) of participants interviewed before the pandemic lockdown (T2-BC) and the remaining 70% (N = 372) during the lockdown. Another survey was collected a year later, after the pandemic lockdown ended (T3). Four trained bilingual interviewers, with experience conducting research surveys in the target population, were recruited from the community and extensively trained in the data collection procedures for the present study. All surveys were administered in Spanish and completed either at a confidential, safe location agreed upon by both the interviewer and participant (i.e., information for T0, T1, and T2-BC) or virtually (T2-DC, and T3). For their participation, participants received $50. Informed consent procedures were completed by all participants prior to data collection. This study was approved by the Institutional Review Board of Florida International University.

Inclusion criteria were determined by the parent study and involved being a Latinx immigrant, 18–34 years old, who at T1 had recently immigrated (within one year prior to assessment) to the United States from a Latin American country with the intention of staying in the United States at least three years beyond T1. Because the study consisted of recently arrived immigrants, many of whom had either temporary or undocumented immigration status, respondent-driven sampling was deemed to be the most appropriate sampling approach. We asked each participant (the seed) to refer three individuals in his or her social network who met eligibility criteria. Seeds were recruited via flyers and in-person throughout MDC neighborhoods and businesses with substantial RLI populations, in community-based agencies serving RLIs, and during Latinx health fairs in MDC. This procedure was followed for a maximum of three legs per seed.

2.2. Demographic and socioeconomic measures

At each interview, participants were asked to inform about their gender (all participants self-reported a binary outcome – either male or female. Subsequently, we renamed this variable as “sex”), age (“18–20 y/o,” “21–24 y/o,” “25–29 y/o,” and “30–34 y/o”), education attainment (“Less than High School (HS),” “High School Diploma,” “Some post HS training/college,” and “Bachelor degree or Higher”); monthly household income (“Less than $2,000,” “$2,000-$3,999,” and “$4,000 or more”); Country/Region of Origin (responses were grouped into “Venezuela,” “Other South American Country,” “Central America,” “Other Latin American Country”); Legal Residency Status (“Legal Permanent,” “Without papers,” “Legal Temporary,” “Asylum”); Marital Status (responses were grouped into “Living with a partner/married” and “other”); and Employment (“Employed” versus “Not employed”).

2.3. Transportation measures

Participants were asked to report on the many modes of transportation they used. To assess transportation in the country of origin, we asked participants whether “in the past year before coming to the United States, did you use [MODE] to get from place to place?” In separate questions, we also asked whether participants “used public transportation (bus, railway),” “used a bicycle,” “used ride-share services (Uber or Lyft),” “used para-transit services (taxi),” and/or “walked” to reach their destinations. Post-immigration transportation modes were assessed by asking participants: “Since you came to the U.S. did you use [MODE] to get from place to place?” with each of the abovementioned options subsequently presented. For each mode of pre- and post-immigration transportation mode that was endorsed, a follow-up item probed for frequency of use (“every day,” “several days a week,” “once a week or less,” “only certain times a year”). For analytical purposes (repeated measures mixed-model regressions), we converted these responses into the following numeric values (days/year): 365 d/y, 208 d/y, 52 d/y, and 10 d/y, respectively. Albeit arbitrary (an estimation of the number of days per year RLIs use a certain transportation mode would be meaningless), the use of these numeric values in our analyses allows for an interpretation of the relative use of transportation modes (the use of a certain transportation mode relative to another) while reducing the demand for degrees of freedom. We also asked participants at each time period whether they had a valid driver’s license. Foreign born visitors to Florida (as well as to most U.S. States) who wish to drive are required to have in their immediate possession a valid driver’s license issued in their name from their country of residence. This license will be valid for up to 3 months upon arrival (FLHSMV, 2014b). More specifically, we elicit the T0 information by asking “Last year, did you have a valid driver’s license in your country of origin?”, and for the remaining periods “do you have a valid driver’s license in the US?”

2.4. Statistical analyses

We first applied descriptive analyses to estimate the distribution of the modes of transportation under examination (Driving, Riding in a Passenger Vehicle, Transit, Bicycle, using a taxi, using a ride-share service, walking) by immigrants’ demographics (age, sex), country or region of origin (Venezuela, Other South American, Central American, Other Latin American countries), education attainment, legal residence status, marital status, employment, and monthly household income.

Next, we used regression analysis to examine the frequency on which each transportation mode was used as a function of the factors included in the bivariate analyses, as well as whether the RLIs had a valid driver’s license. To account for changes in the outcome and explanatory variables over time (from T0 to T3), transportation modes were modeled under a repeated measured framework. The PROC MIXED command in SAS^® v.9.4. was used to perform the analyses. We were also interested in identifying variables with large effect on outcome. Estimating effect size in mixed models with variables repeatedly measured over time is a complex feature that is not provided within the SAS framework. Nevertheless, we computed ad-hoc estimates of Cohen’s d measure by subtracting the means of each fixed effect at each level with respect to the reference level, and approximating the standard deviation by multiplying the standard error provided by LSMEANS procedure in SAS by the square root of the involved degrees of freedom plus 1. This approach allowed us to identify variables with a relatively large impact on the outcome.

3. Results

Table 1 shows that the sample of RLIs is young (about one third of them aged 30–34 at T1, a third 25–29, and another third aged 18–24), evenly distributed by sex.

Table 1.

RLIs’ Demographics and Percent with a valid Driver License, drove, and the mean number of days they drove at each time period.

		T0					T1					T2					T3
		N	col %	%	%	Mean	N	col %	%	%	Mean	N	col %	%	%	Mean	N	col %	%	%	Mean
				valid DL	drove	Driving Days			valid DL	drove	Driving Days			valid DL	drove	Driving Days			valid DL	drove	Driving Days
All		528	100	72.7*	75.0*	219.6*	527	100	49.9	60.7	168.4	528	100	64.0*	75.3*	242.9*	517	100	57.9	66.7	208.9*
Age	18–20	92	17	50.0	60.4	135.0	68	12.9	42.7	53.0	148.9	43	8.1	65.1	67.4	202.4	18	3.6	38.9	44.4	127.4
	21–24	94	18	70.2	73.1	198.0	98	18.6	45.9	55.1	148.3	102	19.3	63.7	72.7	239.3	98	19.5	60.2	66.7	210.3
	25–29	158	30	77.2	79.2	234.8	138	26.2	50.0	59.9	165.5	137	26.0	60.6	73.9	237.4	119	23.7	58.0	66.7	214.3
	30–34	184	35	81.5	79.7	260.2	223	42.3	53.8	65.9	184.8	246	46.6	65.9	78.5	254.5	268	53.3	59.0	68.4	216.2
Sex	Female			68.9	67.8	189.7	259	49.2	42.5	55.1	137.9			63.5	74.0	230.2			59.9	69.0	213.5
	Male			76.5	82.1	249.1	268	50.9	57.1	66.0	197.6			64.6	76.6	255.3			56.7	64.7	209.5
Legal	Asylum			82.5	82.5	268.5	57	10.9	52.6	74.6	167.1	80	15.2	87.5	91.3	295.3	67	13.0	79.1	76.1	236.9
Residency.	Permanent			54.7	55.8	158.0	86	16.4	76.7	76.5	245.5	109	20.7	93.6	89.0	298.5	102	19.7	80.4	80.2	259.3
Status	Temporary			78.2	82.7	237.8	284	54.3	50.4	62.7	173.1	193	36.6	77.7	83.7	275.7	202	39.1	71.3	74.5	231.1
	Undocumented			65.6	65.3	189.4	96	18.4	20.8	33.3	88.2	145	27.5	10.3	45.2	129.0	146	28.2	13.7	41.8	129.7
Month. Income	<$2,000	424	81.5	69.8	77.3	209.9	327	62.9	37.9	54.0	138.0	253	48.1	54.9	68.4	224.1	213	41.2	44.1	49.5	146.3
	$2,000 - $3,999	72	13.9	90.3	91.7	282.8	171	32.9	70.2	75.7	229.4	249	47.3	70.3	80.3	257.9	252	48.7	65.1	23.2	246.6
	$4,000 or more	24	4.6	79.2	91.7	233.2	22	4.2	68.2	50.0	168.2	24	4.6	91.7	95.5	298.5	52	10.1	78.9	15.4	283.9
Country. of Origin	Central Amer.			69.1	71.0	223.8	126	23.9	41.3	53.2	147.1			49.2	67.5	228.2			39.5	50.0	162.0
	Other			43.7	40.0	125.2	71	13.5	66.2	66.2	206.3			67.6	78.9	261.4			71.0	64.5	195.0
	Other South Am.			75.4	81.5	220.9	175	33.2	45.7	63.4	161.5			64.0	75.4	240.3			57.1	72.5	234.1
	Venezuela			85.8	86.8	257.2	154	29.2	53.9	61.0	174.8			74.2	79.7	248.5			70.3	76.0	235.9
Education.	Less than HS			47.1	40.0	78.0	17	3.2	11.8	11.8	30.2			41.2	52.9	137.9			7.1	0.0	0.0
	HS Diploma			54.4	61.4	152.9	195	37.0	47.7	55.7	156.2			41.8	67.0	223.5			56.0	61.3	199.3
	Bachelor/Higher			92.2	89.8	292.1	149	28.3	43.6	71.1	212.4			34.2	81.3	253.1			66.4	75.7	231.4
	Some Train./			78.0	79.3	225.2	166	31.5	62.1	61.2	138.0			28.3	81.9	268.6			56.8	72.5	221.0
	College
Employed.	No			71.4	70.6	208.1	203	38.6	34.0	45.3	111.3	127	24.1	59.1	65.1	185.5	118	22.8	44.1	49.1	140.4
	Yes			73.4	77.6	225.8	323	61.4	59.8	70.2	203.4	401	76.0	65.6	78.5	260.9	400	77.2	61.8	71.7	228.0

“% with valid DL”, % drove”, and “Mean # Driving Days” denote the percentage of RLIs who had a valid driver’s license, percentage who drove (regardless they had a valid license), and the mean number of days they drove in each of the time period. Time period are the year before immigration (T0), the first year (T1), the second year (T2), and the third year in MDC (T3). Cells in grey with entries in bold denote a factor whose levels are statistically different (p<.01). For instance, the % of RLIs with a valid driver’s license at T0 varies significantly (P<.01) with RLIs’ age. Non-grey cells with bold entries indicate significance at the 5% level. T0, T2, and T3 cells with an asterisk (*) in the “All” row denote statistical significance with the T1 Cells. Empty cells indicate a variable measured at T1.

Upon arrival at MDC (T1), about half of the RLIs (54.3%) had a temporary visa (e.g., working permit, tourist visa, student visa), 18.4%, 16.4%, and 10.9% were undocumented, permanent resident, and asylum-seeking individuals, respectively. RLIs’ residency status changed over time as they navigated the legal system, shifting between statuses. For instance, of the 284 RLIs who entered the United States with a temporary visa, 124 (43.7%) entered as tourist (not shown in Table 1). Two years later (T3), 48.4% of them had their visa expired, 13.7% moved into some temporary or permanent residency status (e.g., work permit, student visa, asylum), and 37.9% still held a tourist visa. Regarding income, 81.5% reported earning less than $2000 a month at T1. This figure reduced to 41.2% at T3. About a third of the sample came from Venezuela. About 40.1% of the sample had a high school degree or lower. About 61.4% were employed at T1, a figure that increased to 77.2% at T3.

3.1. Driving

Table 1 shows that the percentage of RLIs with a valid driver’s license, drove a vehicle at least once in each period, and the mean number of days they drove. Compared with the measures reported for the country of origin (T0), all driving measures significantly declined upon arrival to MDC (T1), and significantly increased a year later (T2). During the third year in MDC (T3), all driving measures declined closer to T1 levels, albeit the mean number of driving days at T3 remained significantly higher at T3 than at T1.

Although age had a significant impact on driving in RLI’s country of origin, age was not a significant factor after immigration. Males more than females drove both at T0 and T1. Sex was no longer a significant driving factor both at T2 and T3. RLIs’ legal residency status, income, education, and employment were significant driving factors in each period. Compared with other RLIs, those who were undocumented, earned less than $2000 a month, had less than a high school diploma, and/or were unemployed were significantly less likely to drive at any time period.

3.2. Transportation modes most frequently used

Table 2 and Fig. 1 show the percent of RLIs who frequently used (i.e., every day/several days a week) the transportation modes at each time period. Driving was the transportation mode most frequently used before immigration (64.4% at T0), dropping significantly upon arrival to MDC (48.5% at T1), increasing a year later (69.8% at T2), declining again at T3 (59.2%). The oscillation in driving rates over time was inversely mirrored by oscillations in the other transportation modes. Thus, the drop in driving, biking, and the use of taxis between T0 and T1 was compensated by significantly increases in the percent of RLIs who rode as passengers in private vehicles (from 45.9% to 60.2%) or used ride sharing services (from 6.3% to 26.0%). This pattern reversed the following year (T2). In their third year in MDC (T3), RLIs showed a reduction in the use of almost every transportation modes, which decreased to T1 levels (e.g., driving), or lower (e.g., riding as passengers in private vehicles, using transit, taxi, or walking). The sole exception was the percentage of RLIs who used a Ride Sharing Services most/every day, which was significantly higher at T3 (15%) than in the country of origin (6.3% at T0).

Table 2.

Percent of RLIS who Used a Transportation Mode Most/Every Day.

Transportation Mode	Time Period	%	95% LCL	95% UCL
Driving	T0	64.4%	60.3%	68.5%
	T1	48.5%	44.2%	52.8%
	T2BC	74.4%	67.5%	81.2%
	T2	69.8%	65.9%	73.7%
	T2DC	67.8%	63.1%	72.6%
	T3	59.2%	55.0%	63.4%
Ride	T0	45.9%	41.7%	50.2%
	T1	60.2%	56.1%	64.4%
	T2BC	36.5%	29.0%	44.1%
	T2	34.0%	29.9%	38.0%
	T2DC	32.9%	28.1%	37.7%
	T3	27.2%	23.4%	31.0%
Transit	T0	32.3%	28.3%	36.3%
	T1	36.2%	32.1%	40.3%
	T2BC	12.8%	7.6%	18.1%
	T2	12.0%	9.2%	14.7%
	T2DC	11.6%	8.3%	14.9%
	T3	16.4%	13.2%	19.6%
Bicycle	T0	13.5%	10.5%	16.4%
	T1	8.9%	6.5%	11.3%
	T2BC	3.2%	0.4%	6.0%
	T2	5.9%	3.9%	7.9%
	T2DC	7.0%	4.4%	9.6%
	T3	4.2%	2.5%	6.0%
Taxi	T0	14.0%	11.0%	16.9%
	T1	7.4%	5.2%	9.6%
	T2BC	0.6%	0.0%	1.9%
	T2	1.3%	0.4%	2.3%
	T2DC	1.6%	0.3%	2.9%
	T3	2.9%	1.4%	4.3%
Uber	T0	6.3%	4.2%	8.3%
	T1	26.0%	22.2%	29.7%
	T2BC	10.2%	5.5%	14.9%
	T2	11.2%	8.5%	13.9%
	T2DC	11.6%	8.3%	14.9%
	T3	15.0%	11.9%	18.0%
Walk	T0	32.1%	28.1%	36.1%
	T1	40.1%	35.9%	44.3%
	T2BC	16.0%	10.3%	21.8%
	T2	18.9%	15.6%	22.3%
	T2DC	20.2%	16.1%	24.2%
	T3	16.9%	13.7%	20.1%

Transportation modes use self-reported by Recent Latinx Immigrants (RLIs). “%” indicates the percent of recent immigrants who used a transportation mode either every day or several days a week at each time period. Time period indicate the year before immigration (T0), the first year in MDC (T1), the average second year in MDC (T2), separately for those interviewed before the COVID lockdown in Miami (T2-BC) and during the COVID lockdown in Miami (T2-DC), and the third year in the MDC (T3). 95%LCL and 95%UCL stand for 95% lower and upper confidence limit of %, respectively.

Fig. 1.

Transportation modes use self-reported by Recent Latinx Immigrants (RLIs). Source: Table 2. “%” indicates the percent RLIs who used a transportation mode either every day or several days a week. Time period indicate the year before immigration (T0), the first year in MDC (T1), the second and third year in MDC (T2, and T3, respectively). Columns labeled “BC” and “DC” denote measures for RLIs who at T2 were interviewed before and during the COVID lockdown. 95%LCL and 95% UCL stand for 95% lower and upper confidence limit of %, respectively.

Also indicated in Table 2 and Fig. 1 is the use of the different transportation modes at T2, but separately for the RLIs who were interviewed before and during the pandemic lockdown (denoted as BC, and DC, respectively). Compared to those who were interviewed before the lockdown, those interviewed during the lockdown reported less use of reported less mobility regardless of almost any transportation mode. The exceptions were biking and walking, which increased during the pandemic lockdown.

3.3. Factors Contributing to transportation Modes’ frequency of use of

Table 3 shows the outcome of the mixed-model regressions modeling the frequency of use of the different modes of transportation. RLIs’ use of driving was inversely associated with the frequency of all other transportation modes, albeit the association between bicycling, walking, and driving was not significant. RLIs who frequently used transit were those who did not use a car, either as drivers or passengers. The RLIS who frequently used transit also frequently biked, walked, and/or used a taxi. Those who frequently used a taxi also frequently used RSS.

Table 3.

Factors Contributing to the Frequency of Use of Transportation Modes.

		Drive		Transit		Ride as Passenger		Bicycle		Taxi		RSS		Walk
		Coeff.	Pr > |t|	Coeff.	Pr > |t|	Coeff.	Pr > |t|	Coeff.	Pr > |t|	Coeff.	Pr > |t|	Coeff.	Pr > |t|	Coeff.	Pr > |t|
Transp. Mode	Intercept	165.81	<0.0001	117.62	<0.0001	219.98	<0.0001	6.26	0.40	−8.08	0.21	22.64	0.01	42.26	<0.001
	Drive			−0.21	<0.001	−0.38	<0.0001	−0.02	0.17	−0.04	<0.001	−0.06	<0.001	−0.04	0.12
	Transit	−0.24	<0.0001			−0.14	<0.0001	0.10	<0.0001	0.09	<0.0001	0.01	0.54	0.37	<0.0001
	Ride as Passenger	−0.26	<0.0001	−0.08	<0.0001			−0.02	0.16	0.03	<0.001	0.03	0.02	0.05	0.01
	Bicycle	−0.05	0.18	0.22	<0.001	−0.06	0.16			0.08	<0.0001	−0.02	0.41	0.22	<0.0001
	Taxi	−0.13	<0.001	0.24	<0.0001	0.14	<0.001	0.10	<0.0001			0.12	0.00	0.09	0.04
	RSS	−0.11	<0.001	0.02	0.54	0.08	0.02	−0.02	0.41	0.06	<0.001			0.15	<0.0001
	Walk	−0.03	0.13	0.31	<0.0001	0.07	0.01	0.08	<0.0001	0.03	0.04	0.08	<0.0001
Driver License	Valid License	182.40	<0.0001	−11.84	0.08	−4.86	0.58	−6.49	0.16	12.21	<0.001	−8.22	0.14	2.15	0.77
	Ref: No valid license
Interview at T2	During lockdown	7.63	0.17	−8.90	0.09	−13.392	0.05	7.58	0.04	4.60	0.14	2.01	0.64	13.43	0.02
	Ref: Before COVID lockdown
Age	18–20	−6.48	0.48	10.72	0.20	33.02	<0.001	−1.41	0.81	21.28	<0.0001	−6.29	0.36	17.11	0.06
	21–24	−11.74	0.09	−5.31	0.41	−8.76	0.30	6.46	0.14	8.99	0.02	6.47	0.22	3.59	0.61
	25–29	−1.67	0.78	5.93	0.29	−2.18	0.77	−0.71	0.85	5.42	0.10	8.18	0.07	−1.85	0.76
	Ref: 30–34
Sex	Female	−11.63	0.02	2.20	0.64	14.82	0.02	−12.26	<0.001	0.32	0.91	18.14	<0.0001	0.48	0.93
	Ref: Male
Legal Residency. Status	Asylum			−10.99	0.16	−5.21	0.61	1.36	0.80	9.29	0.05	5.11	0.42	0.58	0.95
	Permanent	−10.66			0.14	−4.71	0.61	4.26	0.38	12.85	<0.001	−3.95	0.50	−12.31	0.12
	Temporary			−7.19	0.23	3.46	0.66	6.11	0.13	6.21	0.08	1.63	0.74	−1.74	0.79
	Ref: Undocumented
Monthly Income	$2,000 - $3,999	12.26	0.02	−26.59	<0.0001	−0.29	0.96	−2.34	0.49	−6.66	0.02	12.22	<0.001	−3.66	0.50
	$4,000 or more	6.63	0.53	−20.17	0.04	3.14	0.81	5.89	0.38	2.67	0.65	14.82	0.07	−10.47	0.33
	Ref: Less than $2,000
Country of Origin	Central American	3.62	0.62	4.26	0.54	−24.70	0.01	4.06	0.39	6.41	0.12	1.00	0.86	−1.48	0.85
	Other	−18.19	0.03	−6.78	0.42	−50.32	<0.0001	8.03	0.16	−3.87	0.44	−7.33	0.29	25.36	0.01
	Other South Am.	3.53	0.60	5.27	0.41	−22.23	0.01	0.74	0.86	2.95	0.44	6.71	0.20	13.38	0.06
	Ref: Venezuela
Education	College/Higher	−9.23	0.17	−4.93	0.43	−4.28	0.60	2.30	0.59	5.96	0.11	−2.54	0.62	7.41	0.28
	HS Diploma	−19.11	<0.001	1.09	0.85	−14.19	0.06	4.84	0.22	3.62	0.30	−2.29	0.63	2.41	0.71
	Less than HS	−59.37	<0.0001	12.61	0.36	14.86	0.41	29.07	<0.001	−21.84	0.01	−30.87	0.01	−2.01	0.90
	Ref: Some Training after H.S. but no Degree
Employment	Yes	13.06	<0.0001	1.75	0.72	27.02	<0.0001	5.97	0.07	−7.69	0.01	18.94	<0.0001	−22.81	<0.0001
	Ref: No
Interviewed at T2 during COVID-19	Yes	7.63	0.17	−8.90	0.09	−13.39	0.05	7.58	0.04	4.60	0.14	2.01	0.64	13.43	0.02
	No

Transportation modes use self-reported by Recent Latinx Immigrants (RLIs). Time periods indicate the year before immigration (T0), the first year in MDC (T1), the second year in (T2), and the third year in the MDC (T3).

Having a driver’s license was the factor most strongly associated with the frequency of driving. Examination of the pseudo effect sizes of the variables in the model for driving (not shown) showed that the ad-hoc Cohen’s d measure for having a driver’s license was estimated to be 0.66; while for all other variables the estimated effect size was lower than 0.10.

Decisions about how frequently transportation modes were used were influenced by RLIs’ sex, income, education and employment. Compared to male RLIs, female RLIs were less likely to drive or bike, but more like to ride as passengers in private vehicles or use RSS. Those living with a monthly income less than $2,000 were less likely to drive, ride as passengers in private vehicles, or use RSS than RLIs with a monthly income of $2,000-$3,999, $4,000 or more, or $2,000-$3,999, respectively. On the other hand, those living with less $2,000 a month were more likely to use transit, ride as passengers, or use taxi than RLIs with a monthly income of $2,000 or higher (transit) or, between $2,000 and $3,999 (ride as passengers and taxi users). RLIs with less than a high school diploma were less likely to drive, take a taxi, or RSS, but more likely to use a bicycle than those with some post high school training. Compared with those unemployed, employed RLIs were more likely to use a car (either as drivers or passengers) or RSS, but less likely to take a taxi or walk to reach their destination.

4. Discussion

The present study confirms previous reports indicating the relatively high driving rates shown by RLIs upon arrival to MDC (Romano et al., 2021). As described by Romano et al. (2021), RLIs’ high driving rates upon arrival may be in part related to distinctive characteristics of the MDC hosting area. MDC is a well-established Latinx immigrant-receiving community with support networks that may have increased recent immigrants’ availability to private transportation (Sanchez et al., 2016a, Schwartz et al., 2010a, McGlynn, 2005). RLIs’ relatively high driving rates upon arrival may also be related to the noted shift in immigration patterns to the United States. Compared to previous immigration patterns, there is a relatively large number of Venezuelans in the sample. It has been noted that Venezuelans have the highest levels of education of any U.S. Latinx immigrant group in the United States (Noe-bustamante, 2019).

Regardless of the relatively high driving rates upon arrival, the present study also confirms previous reports showing that upon arrival to MDC, RLIs showed a reduction in their driving coupled with the use of multiple modes of transportation (Romano et al., 2021). Our study shows that the main reason for such a reduction was the need to acquire a valid driver’s license. The percentage of RLIs who reported having a valid driver’s license dropped from 73% at T0, to 50% at T1. Lacking a valid driver’s license was especially challenging for undocumented RLIs. This finding is not surprising, since individuals applying for a driver’s license in MDC must show proof of Social Security (or a W-2 form) and proof of residential address among other requirements (FLHSMV, 2014a).

Despite not having a valid U.S. license, however, about a third of the undocumented RLIs drove a vehicle at T1. Some authors have argued that undocumented, unlicensed immigrants have high impaired driving and crash risk (Caetano & Clark, 2000; Cherpitel & Tam, 2000; Caetano & McGrath, 2005). Other authors, however, have argued that undocumented immigrants, in their attempts to remain “under the radar,” are actually some of the safest drivers in America (Arce & Sherrets, 2004). Further research on this issue is needed.

The overall reduction in driving upon arrival to MDC (T1) was compensated by an increase in the frequency they traveled as passengers in private vehicles, used transit, or traveled by walking. A year later (T2) RLIs’ driving increased. This increase was paralleled with an increase in the percent of RLIs who had a valid driver’s license: from 50% at T1 to 64% at T2. The increase in driving at T2 was parallel with a decrease in the use of all other transportation modes. Immigrants’ travel decisions are strongly influenced by income. Compared with lower-income immigrants, higher-income immigrants are more likely to find their access to a vehicle to be easier. Subsequently, lower-income immigrants tend to use public transit, bicycles, and carpool more often than their higher-income counterparts (Liu & Painter, 2012; Blumenberg & Smart, 2010; Barajas et al., 2018; Smart, 2010; Barajas et al., 2018). RLIs’ increase in driving at T2 could at least in part be explained by an increase in income achieved during their second year in MDC. The increase in RLIs driving at T2 may have been further stimulated by limited public transit opportunities. As indicated, Miami (FL) remains a car-dependent city that has witnessed a decrease in the use of public transit even before the pandemic lockdown (Miami-Dade Transportation Planning Administration, 2018).

As indicated, RLIs’ driving increased during the second year in MDC, but the increase was larger from those who were interviewed before the lockdown (T2-BC) than during the lockdown (T2-DC), a result that can be explained by the travel restrictions imposed by the lockdown. A year later (at T3), RLIs’ driving rates kept declining, although they remained above what they were at T1. Reasons for such further decline in driving are unclear. Not having a valid driver’s license shouldn’t have been more of a barrier for driving at T3 than it was a year later. It could be argued that the impact the pandemic lockdown had (and it is still having) on people’s working and traveling preferences (such as increased online activities and less travel requirements; Javadinasr et al., 2022) explains at least in part RLIs’ reduction in driving at T3. Researchers and urban planners should investigate this possibility. Moreover, during the COVID-19 pandemic, RLI were faced with not only navigating the stressors related to the immigration experience, but those of an unprecedented pandemic. How this experience impacted COVID-19 related experiences among RLIs, as well as the impact the pandemic had on immigration stressors remains unknown. Research can investigate whether there is a lasting interactive impact of the pandemic and immigration stressors to RLIs’ mental and physical health. The relative high frequency of driving among RLI’s, even upon arrival to the United States, suggests that transit is being underused in MDC. Researchers and urban planners concerned with city pollution and traffic congestion, as well as the social impact of driving, should investigate and address barriers to increase the use of transit by RLIs to MDC.

Finally, research on immigrants’ acculturation process has recently favored the pursuit of integrative frameworks and approaches that take into account the multiplicity of layers and factors influencing this process (Abraído-Lanza et al., 2016). There is consensus that limited transportation options also limits immigrants access to education, health, jobs, or other social services (Atiles & Bohon, 2003). At a minimum, limited access to transportation makes it unnecessarily difficult for Latino immigrants to adjust to their new environment (Bohon et al., 2008). However, the role that transportation plays in shaping the whole acculturation process is unclear, as such a role varies according to immigrants’ intrapersonal context (e.g., norms and beliefs), social support, local and international context, attitudes, and policies, among other factors (Abraído-Lanza et al., 2016). Although not its focus, the present study adds to our understanding of the impact transportation has on immigrants’ acculturation process, as they are pertinent to RLI’s to MDC.

4.1. Study limitations

Our findings should be interpreted in light of certain limitations. Eliciting retrospective information about participants’ pre-immigration behavior could have made responses susceptible to recollection bias. However, our ability to access and interview immigrants relatively soon after their arrival reduces the occurrence of such bias and increases the reliability of the information collected by this study. Another limitation of this manuscript relates to how participants may have interpreted the concept of “having a valid driver’s license.” Tourists are allowed to drive in the United States for several months as long as they bring a valid driver’s license from their country of origin. Although we were interviewing immigrants rather than tourists, it is possible then that some of the RLIs who drove at T1 despite not having a valid U.S. license may have been relying on the license they brought from their country of origin. It is unclear how safely these RLIs drove at T1. Future studies can assess and clarify this concern. RLIs’ employment was based on what was available, and information about length of commute was absent and could not be included in our analyses. Another limitation is that the information collected, and the interpretation of the findings pertain specifically to Latinx immigrants to MDC. Miami–Dade County presents unique characteristics and challenges for Latinx immigrants, culturally, legally, as well as environmentally. It is therefore likely that similar research efforts conducted on Latinx groups in other immigrant-receiving communities would yield different results. Although respondent-driven sampling is a preferred method to recruit “hard-to-reach” populations such as recent and undocumented immigrants, it does not ensure a representative sample. Nevertheless, the findings of this study are relevant to an important U.S. community and metropolis. Moreover, despite their limited generalizability, findings from the present study could be highly relevant to transportation planners in other immigrant-receiving communities, as it facilitates the adoption of early measures to address U.S. immigrants’ travel needs.

5. Conclusions

The present study aimed to understand the travel patterns of RLIs during the initial years in MDC. As such, this study confirms and assesses the uniqueness of the hosting location-immigrant association prevailing in MDC, which allows for high driving rates among RLIs straight upon arrival. Such elevated driving rates suggest the need for transportation planners to understand barriers in the use of transit and other transportation modes in MDC. The study also calls attention to the impact the COVID-19 pandemic has on RLIs’ mobility. A year after the pandemic lockdown, driving as well as the broad use of all transportation modes declined. We speculated that this decline relates to changes in people’s working and traveling preferences, such as increased online activities and less travel requirements, but this speculation needs to be explored. Finally, this study offers a picture of RLIs’ initial three years in MDC. The way that the RLIs to MDC are coping with immigration stress is compounded by how they cope and adapt to post-pandemic changes, as Latinx immigrants and the entire population of MDC adapt to the new post-COVID-19 reality. Future research can examine how the behaviors documented by this effort progress over time, including the need to further assess the impact of the COVID-19 outbreak on such progression.

6. Practical applications

This study alerts transportation planners about the need to address barriers in the use of transportation modes other than driving among RLIs in MDC, and helps them understand reasons for the broad decline in transportation one year after the pandemic lockdown. Understanding these phenomena is crucial to the development of safe and equitable transportation systems for RLIs’ in MDC.

Biographies

Eduardo Romano, Ph.D., is a Senior Research Scientist at the Pacific Institute for Research and Evaluation. An economist by training, his research interests have focused on risk-related behaviors as well as on the analyses of risk-reducing and risk-managing policies. His recent interests include risk perceptions and impaired driving, particularly among children, adolescents, women, and minorities. He is conducting research on issues involving substance use and risky behaviors among Latinx immigrants. Among others organizations, he is a member of the Research Society on Alcoholism, the Research Society on Marijuana, ICADTS, National Hispanic Science Network, and the TRB Committees on Impairment in Transportation, and Women and Gender in Transportation

Mariana Sanchez, Ph.D., has over 20 years of experience in conducting longitudinal and community-based health disparity research across the lifespan including youth, emerging adults, and adults. The bulk of her research has focused on examining how sociocultural determinants influence substance use and mental health among vulnerable Latinx. Dr. Sanchez is currently a co-principal investigator of a mixed-methods study funded by the National Institute for Alcohol Abuse and Alcoholism aimed at understanding the environmental, demographic, and sociocultural factors influencing drinking and driving trajectories among recent Latino/a immigrants. She also serves as a co-investigator on the Adolescent Brain and Cognitive Development (ABCD Study^®). Dr. Sanchez is also the Program Director for FIU’s Doctoral Program in Public Health with a Concentration in Health Disparities.