Quantifying real and perceived barriers to EV adoption in rural Michigan

Abstract

We survey a representative sample of 1000 rural Michigan residents to understand their constraints and attitudes around electric vehicle (EV) adoption and find that only 5% of respondents would choose a battery EV (BEV) as their next vehicle, a smaller number than is found in national surveys. Rural residents face real constraints to adopting EVs. However, their attitudes also correlate with their interest in EVs and they can be misinformed: 42% of respondents think they cannot meet their driving needs by a level 2 home charger while we estimate 75% of those respondents could; 30% of respondents think there is no public charger near their home while we estimate there is a public fast charger within 5 miles for 49% of these respondents; and 35% prefer an internal combustion engine vehicle because of cost, while we estimate a suitable BEV exists within their reported budget for 65% of these respondents. In general, there is room for information interventions that reduce inconsistent perceptions around BEVs. We find that respondents may be falsely pessimistic about BEV feasibility (56% of all respondents). A smaller percentage of respondents are also falsely optimistic (15% of all respondents).

Even as the popularity of electric vehicles (EVs) has risen¹, overall adoption in the United States is lower than in other nations^2,3. This may be particularly true in rural areas where there is a low density of populations (e.g. high-income Democrats) that are more open to adoption than others⁴.

Existing work on understanding the feasibility of mass EV adoption has considered situational constraints arising from both the vehicle technologies themselves (e.g. is the range sufficient for driving needs) and underlying public infrastructure for long-distance driving and touristic travel (e.g. are there enough accessible chargers)^5–7. Other work has focused on self-reported factors that consumers say influence their decision to purchase EVs^8,9. Surveys with self-reported preferences often exhibit an intention-behavior gap, where significantly more consumers seem willing to purchase EVs than actually do¹⁰. For example, existing work has found that up to 30% of consumers report being interested in purchasing an EV¹¹, while across the US roughly 9% of all new vehicle purchases were EVs in 2022¹².

Thus, we are left with a gap between societal estimates of feasibility¹³, and granular self-reported levels of interest¹⁴. This gap in the research between estimates of adoption based on aggregated public data and individual survey research is particularly apparent in rural communities, which are often left behind by both strands of work.

To address this gap, we investigate both constraints and attitudes. We sent letters to 43,000 Michigan residents with home addresses in rural zip codes as shown in Fig. 1 (see Appendix A for more details on the survey and response rates). Each letter contained an invitation to participate in a survey and we present 1063 responses to this survey here, where these responses are augmented with public data. We quantify different constraints and attitudes relevant to BEV adoption. Respondents may face real constraints that restrict the ability of a BEV to meet their needs. For example, a BEV won’t be feasible for someone whose driving requirements exceed the range that can be achieved with a convenient charge. Alternatively, respondents’ attitudes on a range of items, e.g., pro-environmental disposition, may lead to over-optimism about purchasing a BEV.

Fig. 1.

Survey invitation by zip code: Coverage and response rates.

We focus on three aspects of BEV ownership: 1) overnight charging at home, 2) convenient public charging, and 3) availability of new and used BEVs within self-reported budget and vehicle type preferences. We then estimate if respondents would be able to charge overnight given their self-reported driving behavior, if there is a public charger near their home, and whether there are BEVs for sale which meet self-reported constraints such as budget. Next, we investigate respondents’ self-reported attitudes and explore whether these attitudes align with our estimates of their constraints, based on self-reported driving needs and ability to charge at home. We then assess how these constraints and attitudes correlate with interest in purchasing EVs.

To do so, we augment the results of the survey with data we collect on public charging stations and vehicle sales listings. These auxiliary datasets allow us to complement self-reported constraints and attitudes with external observations. Unlike prior work we measure the real distance from respondents’ homes to the nearest charger, a measurement we view as an improvement over other work which has relied on shape files to estimate access within a given geographic boundary¹⁵ or to estimate the density of chargers within census areas, zip codes, or per person^16,17. Our measurement more clearly relates to how a person may think about accessing a charger, that is, how far they need to travel to reach one.

Our work produced several novel insights. We estimate that 75% of all respondents, who reported they could not meet their daily driving needs with overnight charging at home, would in fact be able to do so on average, assuming a level 2 charger. If a home level 2 charger was installed, it would support their reported range even if we assumed that all weekly driving happened in 3 days and that the BEV had reduced range due to cold winter conditions and an aged battery pack.

We find that 65% of all respondents would have at least 1 BEV purchase option matching their desired vehicle type and condition within their reported desired price range (and 57% would have at least 5). There is a public level 2 charger within 5 miles of the homes of 70% respondents. However, direct current (DC) fast chargers are available at commonly stopped-at locations (such as shopping centers) within 5 miles from home for only 29% of respondents. Our findings suggest that low-cost information interventions could ameliorate the effect of attitudes around range, charging at home, and affordability, while policy initiatives should focus on improving access to fast public chargers. We caution that our feasibility estimates are based on technical feasibility and do not imply behavioral readiness or even a complete characterization of the ability to integrate BEVs into specific lifestyles.

Background

Both real constraints on the ability of an EV to meet one’s transportation needs, given one’s lifestyle and living context⁵, and attitudinal factors that influence ownership, have been explored in the literature. Our work makes several contributions: we focus on residents of rural areas, we explore both constraints and attitudes on an individual level, and we inspect when and if attitudes are aligned with constraints.

Constraints on EV ownership

Barriers to EV adoption have been well studied in general. What is less well known is the affect of different constraints on particular populations, such as residents of rural areas. Factors restricting adoption include: battery range, access to convenient public charging, and total cost of ownership¹⁸. Additionally, the absence of standardization of charging stations has been recognized as an obstacle to adoption^18,19.

Existing work has focused on whether battery technology is advanced enough for BEVs to be feasible for certain populations on average, or for special demanding use cases^20,21. Similarly, others have modeled the efficacy of public charging networks^22,23 and proposed new approaches for their design^17,24 and even related access to chargers to interest in EVs²⁵. Others have quantified the effects of large-scale electrification²⁶ and analyzed the limitations of power grids in the United States²⁷. There has also been extensive work on estimating the total cost of ownership^28–30. However, whether BEVs are feasible for people on average may not be as important as if a person considers them to be a viable option³¹. Consequently, there have been efforts to understand the effect of attitudes on reported interest in purchasing EVs.

Attitudes towards EV ownership

Important attitudinal factors include: reliability and safety of EVs, concerns over battery range and cost of replacement, and general anxiety around the lack of familiarity with EV technology^31–34. People have expressed concern over finding maintenance and repair options, and the resale value of an EV^35,36. Another significant factor is the number of options available on the EV market and especially in the used EV market^33,35. Social norms have also been found to have an impact on interest in adoption³⁷.

EVs are perceived as having a much higher purchase price than ICEVs^33,35. Yet, environmental concerns and the perceived sustainability of EVs may be a stronger factor in determining purchase intention than range anxiety or price value³⁸. However, others have found that the only effect of environmental values is to increase the price one is willing to pay for an EV, rather than to impact interest directly¹⁴. Thus while there has been a wealth of research on understanding the relationship between environmental values and purchase intent^14,39–43, there is still ambiguity about the existence and magnitude of the relationship between environmental values on EV interest⁴⁴, especially for rural populations. An additional challenge is that while some perceive EVs as being better for the environment than ICEVs, others may question the sustainability of battery technology production⁴⁵, mining of natural resources, and the pollution burden from charging⁴⁶, which may result in a more ambiguous relationship between environmental values and interest in EVs.

Constraints and attitudes

Thus, both constraints (such as cost) and attitudes (such as range anxiety) have been found to be important in impacting interest in EVs⁴⁷. To estimate how constraints and attitudes relate to self-reported interest in purchasing a BEV, we collect data on both sets of factors. For example, we collect variables related to charging overnight, the presence of public chargers, and the cost of purchasing a BEV. These three constraints and attitudes have been consistently shown to be indicative of interest⁴⁸. For each of these constraints we also measure attitudes towards them. For example, do respondents think they can charge at home? Unlike existing work which relies only on survey responses or societal level measurements, we combine survey responses with relevant public data we collect from three different vehicle sale websites, and a public charging database.

Results

We begin by investigating three types of constraints among the 1063 survey respondents: charging at home to meet daily driving needs, public chargers, and one’s budget. Next, we inspect attitudes towards those constraints. Finally, we analyze the factors which are predictive of interest in purchasing BEVs.

Interest in different vehicle types

We measured interest in ICEVs and EVs in two ways. First, for each of the 8 vehicle types used or new ICEV, BEV, PHEV, or HEV, we asked the respondent to indicate purchase interest on a five-point scale where the options were: Extremely likely, Somewhat likely, Neither likely nor unlikely, Somewhat unlikely, and Extremely unlikely. We show the percentage of respondents who indicated that they would be at least somewhat likely to purchase as the purple dots in Fig. 2.

Fig. 2.

Interest in different vehicle types. There is a clear preference for used ICEVs over electric powertrains. Also, used ICEVs are preferred to new, whereas this pattern is reversed (new vehicles are preferred) for the three electrified powertrains.

Next, we asked respondents to indicate which of the 8 vehicle types would be their first choice. The exact question wording was “Considering the graphic above, if you had to choose one, which of one the vehicle types do you think you will choose when you purchase your next vehicle?” (this graphic is shown in Figure B2). The responses are the orange dots in Fig. 2. As shown in Fig. 2, 13% of respondents indicated interest in either a new or used BEV, while 5% indicated that a new or used BEV would be their first choice.

Ability to meet driving needs with charging overnight at home

We estimate if a respondent can meet their driving needs by charging overnight at home with level 2 charging, assuming that there are not installation constraints. We assume that drivers do not have access to public chargers and must rely on overnight home charging - an assumption that gives us an estimate of the feasibility in the case where public charging infrastructure is non-existent or not conveniently accessible.

Respondents selected the approximate number of miles they drive each week, expressed as a range, among the following driving options: A small amount (Fewer than 200 miles a week), A moderate amount (Between 200 and 400 miles a week), or A great deal (More than 400 miles a week) as shown in more detail in Table 4 and Table 5. We then translate this driving range into three scenarios. Each scenario is described by how a respondent’s weekly driving is distributed throughout the week. For the first two scenarios, we take the total amount of weekly driving to be the maximum of the range they described. We set the maximum to be 600 miles for respondents who reported driving more than 400 miles a week.

Table 4.

The vehicle types refer to the type of vehicle respondents reported that they would purchase as their next vehicle. The most desired next vehicle type is an Sport Utility Vehicle (SUV) / Crossover Utility Vehicle (CUV), with the next most popular category being a pickup truck.

	Next vehicle will be used
	Yes	No	Not sure	Fraction of respondents
SUV / CUV	325	174	171	63%
Pickup Truck	108	60	52	21%
Sedan	61	24	21	10%
Economy	41	15	11	6%
Fraction of respondents	50%	25%	25%

Table 5.

The vehicle types refer to the type of vehicle respondents reported that they would purchase as their next vehicle. For each of these vehicle types we show the self-reported range of miles the respondent reported driving each week.

	Fewer than 200 miles a week	Between 200-400 miles a week	More than 400 miles a week
SUV / CUV	330	267	68
Pickup	90	97	33
Sedan	49	43	14
Economy	41	23	2

Next, we distribute this mileage in three different ways. In the worst case scenario, the respondent would drive the maximum miles from the self-reported range on a single day, meaning that there are BEV models of the stated type-preference with large enough capacity that even at 70% state of health they can obtain sufficient charge in a single night to cover this distance. In the average case, the respondent would drive the maximum miles from the self-reported range equally spread over three days of the week. In the best case, the respondent’s maximum miles from the self-reported range would be uniformly spread across seven days. We stress that to ensure that our estimates are not overly optimistic we take two factors that affect BEV range into account: range loss due to cold temperatures⁴⁹, and battery degradation with used vehicle batteries. Taking these factors into account, Fig. 3 shows that most respondents can meet their driving needs with level 2 charging even in winter conditions unless all of their weekly driving is done in one day. The full tables detailing each calculation are shown in Appendix C.

Fig. 3.

Number of respondents, out of a total of 1063, who can meet their driving needs with level 2 (10 kW) charging overnight (10 hours) at home, with a high winter energy consumption (dark) or non-winter conditions (light). Our analysis accounts for the range (miles) at the beginning of the vehicle’s life for the stated vehicle type preference of each respondent. We assume that a respondent would be able to charge overnight at 10 kW for up to 10 hours. Over half of the 1063 respondents would be able to meet their driving needs if their driving is spread over 3 days, even after assuming the compounded effects of 30% range loss in winter and 30% capacity loss for used vehicles assuming the highest degradation expected while at warranty. Almost 90% of respondents are able to drive in non-winter conditions as shown in the shaded bars.

Presence of public chargers

We filtered for public-access, level 2 and DC fast chargers registered within the state of Michigan in the National Renewable Energy Laboratory (NREL) database⁵⁰. We calculated the number of charging stations that fall within a radius of 0 miles to 25 miles from the geo-coordinates of respondents’ home addresses. Figure 4 shows that 52% of respondents have at least one fast DC charger within 5 miles of their home.

Fig. 4.

Presence of public chargers. 70% of respondents have at least 1 public level 2 charger within 5 miles from home, and 52% respondents have at least 1 public DC fast charger within the same radius. However, considering chargers available at common locations 5 miles away from home, such as grocery stores, these numbers drop drastically to 26% and 29% for level 2 and DC fast chargers respectively. Note, although additional level 2 chargers are available after 5 miles, we consider this impractical for most drivers. We calculate the distance to chargers from the geo-coordinates of each respondent’s home address (latitude and longitude).

However, we note that presence is not access. That is, we consider only chargers in locations we expect people to stop at in the course of their everyday routines, such as grocery stores and gas stations. For example, there is a public charger at a routine location within 5 miles of the homes of only 29% of respondents.

Affordability: Ability to find a BEV within the stated budget range

To estimate respondents’ ability to afford a BEV as their next vehicle, we conduct web searches on three platforms (Cars.com, Edmunds.com, and Autotrader.com) for vehicle sales based on each respondent’s stated preferences for their next vehicle purchase (e.g. vehicle body type (SUV/CUV, Pickup, Sedan, etc.), budget range, and zip code).

While other characteristics could be incorporated in the future, this set of characteristics were available on all three websites. Again, we consider three scenarios, each corresponding to a different price limit. We assume that the respondent is willing to spend no more than the minimum of their price range in the worst case scenario; the midpoint of their price range in the average case; and the maximum of their price range in the best case. The full algorithm for this procedure is shown in Appendix D. Figure 5 shows that half of all respondents can find >25 used or new vehicles within their budget and within 200 miles of their home zip code. How far people are willing to drive to purchase a car does not seem to be well studied, but one survey found that the average distance was 469 miles⁵¹ and we decided that 200 was thus a conservative estimate.

Fig. 5.

Ability to afford a BEV in one’s desired vehicle class. Searching across three online sales platforms, we find that 65% of drivers could purchase at least one used BEV, within their stated budget and desired vehicle class, and 33% could find at least one new BEV.

Attitudes around constraints

While we see that there is a real shortage of public chargers in rural areas, we estimate that the ability to meet driving needs with overnight level 2 charging at home and to afford a used BEV impose less severe constraints than does the relative inaccessibility of public chargers. Next, we would like to estimate whether respondents’ attitudes about overnight charging, public charging, and affordability are aligned with the circumstances of those constraints. First, we explore what we term potentially movable populations. These are respondents who we estimate may have perceptions that do not accurately reflect the constraints they really face. These respondents may be affected by low-cost information interventions. Subsequently, we seek to estimate the effect of different constraints and attitudes around those constraints, on overall interest towards purchasing BEVs. Together, these two analyses disentangle the constraints respondents face from their perception of those constraints.

Potentially movable populations

We estimate the fraction of respondents who might be more inclined to purchase BEVs in response to informational interventions in Fig. 6. These are respondents whose attitudinal responses suggest they believe that a BEV would not meet their needs or that there is a reason they prefer an ICEV, while our estimates of their constraints suggest that BEVs may be a feasible option. Details on the derivation of Fig. 6 are provided in Appendix B.3.

Fig. 6.

Potentially Movable Populations. For each of the three constraints of overnight charging, public charging, and affordability, we show the number of respondents who report a belief suggesting that this is a perceived constraint on BEV ownership in purple/grey bars, and the subset which we estimate would not face that constraint based on best (light orange), base (medium orange), and worst (dark orange) assumptions. For example, in the lowest bars we find that nearly 400 respondents report that cost is a reason they would not purchase a BEV; of these 400 respondents we estimate that more than half could find an affordable BEV of their desired vehicle type in the average case.

We see in Fig. 6 that the largest proportion of respondents may underestimate their ability to cover their driving needs with overnight (10 hours) level 2 charging at home (we inspect this population in Appendix H). That is, 53% of respondents report that they would not be able to fulfill their driving needs with level 2 home charging, yet we estimate that 75% of them would likely fulfill their driving needs with charging at home even if all their weekly stated driving is done across only three days in winter conditions (assuming higher fuel consumption) and considering used battery degradations (70% capacity retention). We note that not all respondents would be interested in or able to install a level 2 charger, and that the 75% is thus likely a generous estimate. When we repeat the analysis with a level 1 charger, we see that 100% of respondents who reported that charging with level 1 is a barrier cannot meet their driving needs when their driving is spread across three days. When their driving is spread across seven days, 31% of them would would be able to meet their needs with overnight level 1 charging. Thus, level 1 charging is not feasible for most of rural Michigan (see G3). We further explore this population in Appendix H. For example, in Figure H5 we see that relative to other survey respondents this potentially movable population is less likely to think that people they know approve of driving BEVs and to know fewer people who drive BEVs.

How common is false optimism about BEV feasibility? Another quantity with policy relevance is the population of respondents who indicate interest in purchasing a BEV, but for whom we estimate a BEV would not be feasible. We see that 144 respondents indicate interest in a new or used BEV, and 35% of them may be falsely optimistic about BEV feasibility. In Table 1 we explore which constraints respondents may be falsely optimistic about. Of the interested population who has seen public chargers near their home, we estimate that 21% do not have DC or level 2 fast chargers within 5 miles of their home. In Table 1 we also see that the those interested in BEVs tend to be falsely optimistic less often those who aren’t interested.

Table 1.

We consider what fraction of the population may be falsely optimistic both for those interested and not interested in BEVs. In all cases one is labeled as falsely optimistic if we estimate they cannot do the behavior they think is feasible, i.e. if we estimate they cannot meet their daily driving needs with overnight charging but they report that they can. Here we consider the average case for all scenarios, that is the case where driving is spread across 3 days, public chargers (either DC or level 2) are within 5 miles of one’s home, and one’s budget is at the midpoint.

	Interested in BEV		Not interested in BEV
	Total respondents	% falsely optimistic	Total respondents	% falsely optimistic
Respondents reported they would be able to meet daily driving needs with level 2 charging	115	15%	500	19%
Reported seeing a public charger near home	120	21%	628	28%
Respondents reported they would be able to afford a BEV	90	19%	597	30%

Overall, considering whether respondents are more likely to be falsely optimistic or falsely pessimistic we estimate that false pessimism is a larger problem. Across all respondents, those we estimate that 77% of those who report any perceived constraint may be misinformed while only 56% of those who do not report any perceived constraint may in fact be subject to a barrier. Thus the total population of those who are falsely pessimistic (56% of all survey respondents) is much larger than that of the falsely optimistic (15% of all survey respondents).

Factors that predict interest in purchasing BEVs

We have seen that respondents do face real constraints around purchasing BEVs as their next vehicle. However, there is also a significant number of respondents whose attitudes around those constraints may be misinformed, as shown in Fig. 6. To understand how both constraints and attitudes about those constraints predict interest in BEVs, we fit a linear regression model. This model takes the following form:

1

where is a ordinal variable in the range of -2 to +2 and +2 indicates high interest in purchasing a BEV as the next vehicle, is a vector of variables that express constraints and is a coefficient vector for those constraints, is a vector of variables that express attitudes around constraints and is a coefficient vector for those attitudes, and is a vector of control variables with coefficient vector .

First, we consider constraints that are related to battery range. For example, we consider both the amount that one drives in a week, and how far they need to travel to reach essential services. As battery health can be affected by colder temperatures, for each respondent’s zip code, we calculate the number of days each year that the location experiences a below-freezing temperature. Similarly, to take into account the physical environment, we consider whether each respondent lives in a remote or distant area⁵². While the remoteness of one’s residence is likely aligned with their exposure to public chargers, we also utilize the collected data on public chargers to determine if there is a public charger (either level 2 or a DC fast charger) within 10 miles of the respondent’s home address. Together these constraints provide an idea of the public services around where the respondent lives, their exposure to the cold, and how much they drive. Finally, we include self-reported budget range for one’s next vehicle within the set of constraints.

Next, we consider attitudes around constraints. We begin by including attitudes towards charging at home and public charging. We then include responses to the question, I would prefer an ICEV to a BEV because.... These capture attitudes directly related to BEVs. Next, we draw on the literature highlighted above, to include standard demographics, the vehicles owned in the household, one’s electrical configuration, government policies around BEVs, environmental attitudes and finally trust in US institutions (as institutionalized trust has been found to be associated with a greater willingness to sacrifice for the environment⁵³). A full description of all variables included in the regression can be found in Appendix E. We include all 920 respondents who completed every survey question.

Linear regression results

Our focus is on the coefficients around constraints () and around attitudes () (additional regression results can be found in Appendix F).

We see that both constraints and attitudes likely influence interest. Respondents’ budgets are consistently significantly positively correlated with interest. That is, the higher a respondent’s budget the more interested they are in purchasing a new BEV. Other factors, such as one’s distance to the nearest doctor, and whether they live in a remote locale are negatively correlated with interest, however the strength of these relationships varies and is less consistent than that of budget.

Several attitudes are consistently correlated with interest. Both whether one reports seeing public chargers and whether they express that they believe they cannot charge overnight are negatively correlated with interest. Similarly, preferring ICEVs because of their range is consistently significantly negatively correlated with interest in BEVs. Thus, some form of range anxiety does appear to be a significant attitude with a negative impact on interest in BEVs. We see very little impact of the attitudes that BEVs are unsafe or that batteries may catch fire. Finally, while the cost of purchasing or financing was not a significant barrier to BEV interest, the cost of replacing the battery was. This implies that people are sensitive to the total cost of ownership and not just the cost of purchasing BEVs.

We included a large set of controls which do effect the magnitude of the relationship between different constraints and attitudes, and interest in purchasing a new BEV. Still, even with these controls we see strong relationships still emerge. Both the total cost of ownership of BEVs, and aspects of range anxiety are predictive of interest, even with extensive controls.

Discussion

We find little interest among a representative sample of residents of rural zip codes in choosing a BEV as their next vehicle. Overall, while 13% of respondents would be interested in either a new or used BEV, only 5% reported that either would be their first vehicle choice (far below the 30% figure reported in⁴⁰). Instead, our result is much closer to the 7% of strong interest found in⁵⁴. We are the first to explicitly focus on rural populations and these figures demonstrate that rural populations lag behind others in BEV interest.

We estimate that of those 42% of all respondents who report that they do not think they can charge overnight at home to meet their driving needs, 75% may benefit from an information intervention. Unless they complete all their driving in one day of the week, they would be able meet their driving needs by charging at home with a level 2 charger, even in the winter. While there is a possibility that respondents would not be able to use level 2 charging at home, we do not think this is the case. That is, we asked respondents who stated that they would not be able to charge at home overnight the following question: “Why do you think you wouldn’t be able to charge a Battery Electric Vehicle overnight at your home?”. The most common response which was given by 55% of respondents was, “I drive too much in a day”. This does not seem to be true given their self reported driving, and thus there is a need to better communicate the current range capacities of available BEVs.

Moreover, we find that over half of all respondents who reported the cost of BEVs as a reason to prefer ICEVs would likely be able to afford a BEV given their reported budget. The marketplace for BEVs is dynamic, and future work should track how the prices of BEVs evolve. Perceptions of affordability will likely lag behind price tags, making the need for up-to-date-information all the more pertinent. Finally, we also see that concern over the cost of replacing a battery is a factor in interest.

Thus, our work contains several policy implications. First, there is a clear need for better information around the ability of level 2 overnight home charging to meet daily driving needs with current BEV technology. For example, given access to location history stored on mobile phones it would be possible to deliver fine-grained estimates of whether level 2 charging would be sufficient for one’s average weekly driving needs⁵⁵. Personalized information interventions have been shown to decrease range concern⁵⁶ and tools which allow consumers to explore vehicle options tailored to their needs may close information gaps.

Second, there is a need for increased infrastructure for public charging stations. Compared to charging overnight at home and BEV affordability, we see less opportunity for low-cost interventions around public charging, as there is a real lack of charging infrastructure near respondents. We see that 70% respondents do not have a public DC fast charger within 5 miles of their home at the kind of locations where they would be likely to stop. Furthermore, our analysis focused on public databases of charging infrastructure, which, even when combined with personal survey responses, do not provide enough fine-grained information about whether one could actually use these public chargers in their daily life; that is whether they are affordable, consistently available, or at safe and convenient locations where one stays for a sufficient amount of time to charge a vehicle. Considering that even without imposing these filters, we see limited public charging availability in these rural locations, real infrastructure interventions are needed to improve access to public chargers in rural areas.

We found that exposure to public charging stations is correlated with interest. Not only is such infrastructure important for driving BEVs, increased visibility of charging stations has been recognized as an important policy avenue for improving awareness⁵⁷. One example intervention would be to support the test driving of BEVs for a sustained period of time (several weeks), at no financial cost (see https://rural-reimagined.com/ for one example program). Such programs can affect not only drivers, but others who observe these vehicles in their community.

Our analysis, while a valuable snapshot of the current state of the infrastructure, describes the situation at the earliest stages of adoption (fewer than 2% of U.S. cars have a plug⁵⁸). It also represents a conservative analysis of charger availability and willingness to adopt. As fast chargers are both available and visible at locations that are accessible to a wide swathe of the population, including rural populations, our analysis (see Table 2) suggests that attitudes will shift. In the past, retailers like Walmart have played a significant role in consumer adoption of green technologies^59,60.

Table 2.

Factors correlated with interest in BEVs.

	Dependent variable:
	Interest in purchasing a new BEV
Constraints ()
Driving amount (ordinal)	0.004	0.02	0.04	0.01	0.01	−0.03	−0.01	0.03
	(0.09)	(0.09)	(0.09)	(0.09)	(0.09)	(0.08)	(0.08)	(0.08)
Budget for my next vehicle purchase (ordinal)	0.33***	0.32***	0.32***	0.32***	0.33***	0.27***	0.27***	0.28***
	(0.07)	(0.08)	(0.08)	(0.08)	(0.08)	(0.07)	(0.07)	(0.07)
Lives in Upper Peninsula (binary)	0.01	0.04	0.03	0.04	0.04	0.04	0.04	0.06
	(0.11)	(0.11)	(0.11)	(0.11)	(0.11)	(0.10)	(0.10)	(0.10)
Level 2 Charger 10 miles from home (binary)	0.07	0.05	0.05	0.04	0.03	−0.02	−0.01	−0.02
	(0.09)	(0.09)	(0.09)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)
DC Fast Charger 10 miles from home (binary)	0.07	0.06	0.05	0.04	0.04	0.07	0.08	0.10
	(0.09)	(0.09)	(0.09)	(0.09)	(0.09)	(0.08)	(0.08)	(0.08)
Lives in distant or remote locale (binary)	−0.18**	−0.14*	−0.12	−0.10	−0.10	−0.11	−0.11	−0.11
	(0.08)	(0.08)	(0.08)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Distance to nearest gas station (ordinal)	0.06	0.05	0.04	0.03	0.03	0.04	0.04	0.05
	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.07)	(0.07)	(0.07)
Distance to nearest grocery store (ordinal)	0.09	0.07	0.07	0.07	0.07	0.03	0.04	0.02
	(0.09)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)
Distance to nearest doctor (ordinal)	−0.15*	−0.11	−0.11	−0.12	−0.12	−0.11	−0.12	−0.13*
	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.07)	(0.07)	(0.07)
Distance to job (ordinal)	−0.04	−0.02	0.002	−0.02	−0.01	−0.01	−0.04	−0.06
	(0.09)	(0.09)	(0.09)	(0.09)	(0.09)	(0.08)	(0.08)	(0.08)
Distance to nearest emergency room (ordinal)	0.001	0.01	0.01	0.01	0.01	0.03	0.04	0.03
	(0.09)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)
Distance to nearest town (ordinal)	−0.14*	−0.10	−0.10	−0.10	−0.11	−0.05	−0.04	−0.03
	(0.08)	(0.08)	(0.08)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Days per year with max temperature less than (continuous)	−0.01	−0.01	0.01	−0.01	−0.003	0.01	−0.001	−0.02
	(0.11)	(0.11)	(0.11)	(0.11)	(0.11)	(0.10)	(0.10)	(0.10)
Attitudes related to constraints ()
Have not seen public chargers	-	−0.14**	−0.13*	−0.14*	−0.15**	−0.10	−0.11*	−0.11
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Can’t charge overnight at home	-	−0.37***	−0.35***	−0.29***	−0.31***	−0.19**	−0.17**	−0.17**
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Would prefer ICEV because...
Concern over battery fires	-	−0.02	−0.01	−0.002	−0.01	0.09	0.09	0.09
		(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.07)	(0.07)
The cost of charging	-	−0.18**	−0.15**	−0.14**	−0.15**	−0.06	−0.05	−0.04
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
The cost of purchasing / financing	-	0.13	0.11	0.13	0.14*	0.09	0.09	0.11
		(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)	(0.08)
The cost of replacing batteries	-	−0.18**	−0.18**	−0.17*	−0.17**	−0.21**	−0.19**	−0.18**
		(0.09)	(0.09)	(0.09)	(0.09)	(0.08)	(0.08)	(0.08)
Of concern over safety	-	0.07	0.05	0.07	0.07	0.08	0.07	0.06
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Of the environmental impact of lithium batteries	-	−0.27***	−0.26***	−0.24***	−0.23***	−0.10	−0.09	−0.07
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Unable to charge at home	-	−0.02	−0.03	−0.04	−0.05	−0.03	−0.03	−0.02
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Unable to charge during long trips	-	0.001	−0.01	0.01	0.01	0.03	0.02	0.03
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Of the range	-	−0.18**	−0.17**	−0.16**	−0.17**	−0.16**	−0.13*	−0.13*
		(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Demographics
Gender	-	-	0.000	−0.01	0.01	−0.02	0.03	0.03
			(0.07)	(0.07)	(0.08)	(0.07)	(0.07)	(0.07)
Race	-	-	0.09	0.12*	0.11	0.04	0.02	0.02
			(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Partisan affiliation	-	-	0.14***	0.14***	0.13***	0.04	−0.01	−0.04
			(0.04)	(0.04)	(0.04)	(0.04)	(0.04)	(0.05)
Education level	-	-	0.05	0.04	0.04	−0.06	−0.07	−0.08
			(0.07)	(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Vehicle status
Number of pickups already own	-	-	-	−0.04	−0.03	0.02	0.03	0.03
				(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Number of SUV already own	-	-	-	−0.02	−0.02	−0.04	−0.02	−0.04
				(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Number of minivans already own	-	-	-	0.01	0.02	−0.02	−0.01	−0.01
				(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Number of BEVs already own	-	-	-	0.39***	0.38***	0.32***	0.30***	0.32***
				(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Number of PHEVs already own	-	-	-	0.07	0.07	0.04	0.05	0.04
				(0.07)	(0.07)	(0.07)	(0.07)	(0.07)
Awareness of electrical configuration at home
Need upgrade home electricity panel (Y/N)	-	-	-	-	−0.04	−0.08	−0.06	−0.03
					(0.07)	(0.07)	(0.07)	(0.07)
Know the capacity of my electrical panel (binary)	-	-	-	-	−0.07	−0.01	0.001	0.01
					(0.07)	(0.07)	(0.07)	(0.07)
Concerned about losing power at home (binary)	-	-	-	-	0.15**	0.12*	0.12*	0.11*
					(0.07)	(0.07)	(0.07)	(0.07)
Awareness and norms around EVs and EV policies
People I know drive BEVs	-	-	-	-	-	0.06	0.06	0.06
						(0.08)	(0.08)	(0.08)
People I know approve of driving BEVs	-	-	-	-	-	0.28***	0.22***	0.19**
						(0.08)	(0.08)	(0.08)
Owning a new BEV improves my self image	-	-	-	-	-	0.46***	0.39***	0.34***
						(0.07)	(0.07)	(0.08)
I’ve heard about rebates for purchasing new EVs	-	-	-	-	-	−0.01	−0.03	−0.02
						(0.07)	(0.06)	(0.06)
I’ve heard about rebates for purchasing used EVs	-	-	-	-	-	−0.02	−0.003	0.01
						(0.08)	(0.08)	(0.08)
I’ve heard about rebates for installing home chargers	-	-	-	-	-	0.12	0.12	0.15*
						(0.08)	(0.08)	(0.08)
I don’t approve of government policies to support EV adoption	-	-	-	-	-	−0.18**	−0.12*	−0.10
						(0.07)	(0.07)	(0.07)
Environmental attitudes
Concerned about global warming	-	-	-	-	-	-	0.35***	0.24**
							(0.08)	(0.10)
Trust in US Institutions
Trust in the federal government	-	-	-	-	-	-	-	0.21**
								(0.10)
Trust in the news media	-	-	-	-	-	-	-	0.13
								(0.09)
Trust in Republican	-	-	-	-	-	-	-	0.13
								(0.09)
Trust in Democrats	-	-	-	-	-	-	-	−0.07
								(0.10)
Trust in individual elected politicians	-	-	-	-	-	-	-	0.09
								(0.08)
Trust in Legislative bodies like Congress	-	-	-	-	-	-	-	0.08
								(0.09)
Trust in the Environmental Protection Agency (EPA)	-	-	-	-	-	-	-	−0.11
								(0.10)
Trust in the oil and gas industry	-	-	-	-	-	-	-	−0.10
								(0.08)
Trust in airlines and the air transportation industry	-	-	-	-	-	-	-	0.07
								(0.07)
Trust in researchers from universities	-	-	-	-	-	-	-	0.09
								(0.10)
Trust in researchers from organizations like NASA	-	-	-	-	-	-	-	−0.09
								(0.09)
Observations	920	920	920	920	920	920	920	920
	0.04	0.14	0.15	0.18	0.19	0.26	0.28	0.30
Adjusted	0.03	0.11	0.12	0.15	0.15	0.23	0.24	0.26

Note: ^*p<0.1; ^**p<0.05; ^***p<0.01.

Methods

This work utilizes a range of methods: a survey, auxiliary data on vehicle characteristics and battery life, auxiliary charging data, and auxiliary web data on vehicle sale listings. We describe our processes for collecting and incorporating these data below. As our primary data is collected through a survey, we first describe the design of this survey and the data we collected through it. Next, we describe the vehicle and battery life information we used, as well as the assumptions we made, in order to estimate whether a respondent would be able to meet their daily driving needs with an overnight charge. We then describe how we calculated one’s distance to a public charger, before detailing the process for collecting and incorporating web data on vehicles.

Survey design

In order to reach residents of rural Michigan we purchased a random sample of 50,000 addresses in zip codes classified as rural with the help of an internal institutional survey office. We sent 30,000 letters through the United States Postal Service in the months of September and October 2023 and 13,000 more letters in the months of January and February, 2024. Each letter contained a short invitation to participate in an online survey hosted on the platform Qualtrics (see Appendix A for the survey template).

The survey was designed to elicit respondents’ driving and fueling behavior, vehicle preferences, experiences with electrical utilities in their home, attitudes towards electric vehicles, knowledge of existing policies for increasing access to electric vehicles, attitudes towards the environment and institutions, and demographic profiles. Survey participants received $5 for their time. We obtained 2698 responses in total 1063 of which were by respondents who completed the survey and passed an attention check. Our results are based on an analysis of these 1063 responses. The full breakdown of our recruitment is shown in Table 3.

Table 3.

Survey Responses Breakdown. The final sample of 1063 that fed our analysis only includes respondents who consented to participate, and who contributed quality responses, e.g. they passed the attention check, and completed the entire survey at a reasonable rate (time spent taking the survey was greater than half of the median duration, i.e., 10 mins).

	Survey Period
	Sep 1 - Nov 14, 2023	Jan 19 - Feb 20, 2024
Letter Sent	30,000	13,000
Completed EV Survey	1,857	841
Completion Rate	6.19%	6.47%
Total Completed	2698
Consented	2685
Passed Quality Check	1063
Completed every survey question	920

All methods were carried out following the guidelines of the Institutional Review Board (IRB) of the University of Michigan. The University of Michigan IRB approved the experimental protocol. Informed consent was obtained from all subjects who participated in this study.

As part of the survey, to ensure that all respondents had a common point of reference around vehicle type definitions, we showed them a graphic with the four vehicle types mentioned in the survey: Battery Electric Vehicle (BEV), Plug-in Hybrid Electric Vehicle (PHEV), Hybrid Electric Vehicle (HEV), and Internal Combustion Engine Vehicle (ICEV). The ranges of the vehicle types were displayed as well as their fuel source: BEV (277 miles, Battery), PHEV (615 miles, Battery 25 mile, Fossil Fuel the remainder), HEV (588 miles, Small Battery and Fossil Fuel), ICEV (411 miles, Fossil Fuel). These statistics were calculated using the following data sources: emissions⁶¹, range⁶¹, fuel type⁶¹, total cost of ownership^62–68. Please see Figure B2 for reference.

Estimating the ability to drive with an overnight home charge

We assume that overnight charging takes about 10 hours. Here, we take respondents’ future vehicle type into account. We found that SUV/CUVs are by far the most popular future vehicle choice, for which there are multiple BEV options. Across all vehicle types the majority of respondents plan on purchasing a used vehicle. The intersection between whether a respondent reports planning on purchasing a used vehicle and the vehicle type they are interested in is shown in Table 4 and the intersection between respondents’ self-reported weekly driving range and the vehicle type they are interested in is shown in Table 5.

We grouped respondents into two categories, those who stated that their next vehicle would be a pickup and those who reported that it would not. If a participant’s desired car type for their next car was a pick-up truck, their range was estimated with a 2022 Ford F150 Lighting 4WD⁶⁹. If a participant’s desired car type for their next car was not a pick-up truck, their range was estimated with a 2022 Tesla Model Y AWD⁷⁰.

For overnight charging with level 2 chargers, a session of 10 hours of charging will exceed the full battery capacity and will enable the driver to start their day with a fully charged pack. Therefore, we estimate the overnight mileage gain as the certified range of the vehicle: 279 miles for Tesla Model Y, and 230 miles for Ford F150 Lighting. We assume this range could be reduced by a factor of 0.7 at worst due to battery capacity retention requirements in most BEV warranties^71–73. As shown in Table 6, this reduces the range estimates (worst case under warranty) for used vehicles to 195 miles for the Tesla Model Y, 161 miles for the Ford F150 Lighting. We describe the calculations for winter and level 1 driving in Appendix C.

Table 6.

Maximum vehicle range under different modeling conditions.

	New & Non-winter			Used Range (miles)	Used & Winter Range (miles)
	Range (miles)	Energy Consumption (kWh/m)	Fuel Efficiency (m/kWh)	Degradation Factor 0.7	Degradation Factors 0.7 x 0.7
Pick-up Truck 2022 Ford F150 Lighting 4WD	230	0.49	2.04	161	113
Other 2022 Tesla Model Y AWD	279	0.28	3.57	195	137

Once we obtained the mileage calculations that could be achieved from an overnight home charge under different modeling conditions (level 1 vs. level 2, non-winter vs. winter, new vs. used battery), we compared the mileage for each condition to the estimated daily driving range by survey respondents. If the mileage achieved from an overnight home charge exceeds a respondent’s estimated daily driving range then we count the respondent as able to drive with overnight charging, and vise versa. Recall from earlier, respondents selected the range of miles they drive each week among the following: A small amount (Fewer than 200 miles a week), A moderate amount (Between 200 and 400 miles a week), or A great deal (More than 400 miles a week). We then distribute the maximum of these mileage ranges. In the worst case, the respondent would drive the self-reported range on a single day, meaning that they would need to obtain sufficient charge to cover this distance in a single night. In the average case, the respondent would drive the self-reported range equally spread over three days of the week. In the best case, the respondent’s maximum miles from the self-reported range would be uniformly spread across seven days.

Additional summary statistics on which participants would be able to meet their driving needs with overnight charging according to different modeling assumptions are shown in tables C2, C3, and C4.

Distance to a public charger

To estimate the presence of chargers away from home we use a charger database maintained by the NREL dataset⁵⁰ from 2023. This database was accessed from May 24th to June 2nd 2024. We have access to the latitude and longitude of each respondent’s home address. We can thus construct a 25 mile radius from each respondent’s home.

Finding a BEV within Budget

Algorithm 1 shows our process for determining if there is an appropriate BEV within a respondent’s budget. We collected data from three different websites, Cars.com, Edmunds.com, and Autotrader.com. These three websites were chosen after inspecting site analytics which reported these as commonly used websites by people looking for new vehicles online. This data was collected shortly after the survey in the time period of August 2024.

Algorithm 1.

Find BEV within budget.

Robustness to model specification

We additionally repeat our regression analysis using an ordered logit model specification in F5. Doing so, we see the relationships we highlight in Table 2 are supported in this model specification as well. Two of the relationships we highlight have different magnitudes according to which model is chosen, however, we see that none have different directions. In particular, the negative relationship between whether one has seen public chargers and their interest in purchasing a new BEV is more pronounced in the ordered logit specification. The relationship between concern about the range and interest is more pronounced under the linear regression specification. Thus, these two relationships may warrant more study than others to explore how robust they are in other settings. Still, we see that in either specification, not having seen public chargers and being concerned about the range of BEVs are negatively correlated with interest.

The particular variables we included in the regression may influence the results. Our aim was to design a comprehensive survey which covered many items in one place which have previously been investigated in isolation. This comprehensiveness may come at the cost of depth into particular constraints or attitudes.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

S.T., A.S. and P. V. conceived and planned the research design. S.T. and S.Y. carried out the survey. S.T. and S.Y produced the visualizations. All authors contributed to the interpretation of the results. S.T. took the lead in writing the manuscript. All authors provided critical feedback and helped shape the research, analysis and manuscript.

Data availability

No datasets were generated or analysed during the current study. The data that support the findings of this study are available from the corresponding author but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-32478-w.