COVID-19 Risk Perceptions, Self-Efficacy, and Prevention Behaviors Among California Undergraduate Students

Abstract

Objective:

College students play a major role in the transmission of SARS-CoV-2, the viral agent responsible for COVID-19. We aim to understand risk perceptions, self-efficacy, and adoption of prevention behaviors in this population to inform prevention strategies.

Participants:

Undergraduate students attending a large public university

Methods:

A convenience sample of students were surveyed (April–June 2020). Participants self-reported risk perceptions, perceived risk of contracting COVID-19, self-efficacy, and prevention behavior engagement.

Results:

A total of 1,449 students were included in the analysis. The majority were women (71.2%) and aged 18–24 (86.6%). Freshmen had the lowest risk and threat perceptions, as did men; men also had lower self-efficacy. Women engaged significantly more in prevention behaviors compared to men.

Conclusions:

Perceived risk of contracting COVID-19 was low, but overall adoption of prevention behaviors was high due to local mandates. Freshmen men were identified as having the greatest need for changing perceptions and behaviors.

Introduction

Coronavirus Disease 2019 (COVID-19) is the contagious disease caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2).¹ From January 21, 2020, when the first COVID-19 case was documented in the United States (USA),² to March 13, 2020, when the President of the USA declared a National Emergency, the Centers for Disease Control and Prevention (CDC) reported 1,678 COVID-19 cases and 41 deaths across the country.³ Within days, universities and colleges halted nearly all in-person instruction and transitioned to virtual learning.⁴ On March 19, to slow the spread of COVID-19, the Governor of California issued an Executive Order directing all Californians to stay home except for essential work and needs.⁵

Because adherence to public health recommendations can vary, understanding the motives and enactment of prevention behaviors among college students in the pandemic is critical to effective mitigation of transmission in the broader community. College students, who tend to be young, or emerging, adults, are transitioning from adolescence to adulthood.⁶ The pre-frontal cortex, the region of the brain responsible for planning and controlling impulses, finishes maturing in the mid-to-late 20s,⁷ and critical connections between the emotional (amygdala) and cognitive (frontal cortex) parts of the brain are still developing.⁸ This stage of neural development leads to behavioral changes, of which the most commonly exhibited cross-culturally are increased novelty-seeking, risk-taking, and a social affiliation shift toward peer-based interactions.⁹ This developmental period often coincides with college-years, where students are living away from home outside of parental supervision and are adjusting to new people, environments, and rules.⁶ College students also tend to reside in communal living situations¹⁰ and to work in the service industry,¹¹ increasing their risk of exposure to COVID-19. A Fall 2020 study among University of Wisconsin undergraduates living in dormitories found that students were 1.5 times more likely to get COVID-19 if they were sharing a room and 1.8 times more likely if they were sharing a room and living space compared to students living alone in dorms.¹²

Early studies from China underscored the role that young people and asymptomatic cases play in the transmission of COVID-19. One study inferred that 86% of cases in China prior to January 23, 2020 were undocumented; despite lower transmission rates in China during this time, 66% of all COVID-19 cases in the country were attributable to undocumented cases.¹³ In a case study of seventy-eight COVID-19 patients from twenty-six exposure clusters in China, asymptomatic cases (40%) were significantly younger than symptomatic cases,¹⁴ and a contact tracing observational study in Italy showed that nearly 80% of cases among 0–19 year-olds were asymptomatic.¹⁵

Asymptomatic cases are more difficult to identify and quarantine, and consequently shed virus to people with whom they interact.¹⁶ This is problematic given the rate of secondary transmission is highly over-dispersed, where only about 10% of total cases cause 80% of COVID-19 transmissions (reproduction number, R₀: 2–3, dispersion number, k: ~0.1).¹⁷ These factors contribute to super-spreader events,¹⁸ with ramifications evidenced by counties with academic universities that had early spring break (i.e., before declared national emergency) having increased COVID-19 growth rates and 20% higher cases per capita in the two weeks following spring break compared to counties where universities had spring break later (i.e., after shutdowns).¹⁹ Lehnig et al.²⁰ found that alternative spring break schedules (i.e., multiple breaks for shorter durations throughout the semester), could reduce the incidence of COVID-19 in the college population by up to 38%.²⁰ These studies highlight that the behaviors of college students impact the broader surrounding community and underscore the need to disseminate appropriate prevention messaging, especially to younger people who may be asymptomatic and likely to attend super-spreader events.

Risk perception– the perceived personal vulnerability or likelihood of a health threat– plays a key role in most health behavior theories.^21–23 In a meta-analysis of risk perception and adult vaccination, people with higher perceived susceptibility and/or perceived likelihood of getting ill were more likely to get vaccinated.²⁴ While theories posit that risk perceptions shape health behaviors, the magnitude and direction of the association may vary across different behaviors and health outcomes depending upon one’s perceptions of risk and self-efficacy. For instance, health behaviors perceived as more difficult to perform may demonstrate a weaker association between risk perceptions and behavior enactment.^25–26 Similarly, one may have a low-risk perception and still enact a health behavior if self-efficacy is sufficiently high.

In response to the COVID-19 pandemic, complying with public health recommendations is imperative to reducing COVID-19 transmission in the community. This is especially important for young adults who are likely to be asymptomatic/mild cases and have a proclivity toward social gatherings. Understanding the risk perceptions and self-efficacy of college students is important for tailoring effective approaches for prevention messaging and interventions. As such, this study aims to measure risk perception, self-efficacy, and the adoption of prevention behaviors among undergraduate students, while assessing differences by gender, class standing and perceived risk.

Methods

Participants and Procedures

The study was conducted using data collected from undergraduate students at a large public university in California. Students were conveniently sampled via email listservs and social media and self-selected into the study without incentive. Individuals were directed to Qualtrics and were screened prior to completing a 36-item investigator-initiated questionnaire; current undergraduate students aged ≥18 years were eligible for study inclusion. The investigators modified and adapted measures from prior research and frameworks to capture demographic information, health information seeking and scanning behaviors,^27–28 COVID-19 knowledge²⁸, COVID-19-related symptoms,²⁹ perceived COVID-19 threat, COVID-19 situational risk perceptions, perceived risk of contracting COVID-19,^25,30 self-efficacy,^25,30 stage in Precaution Adoption Process Model,³¹ and adoption of prevention behaviors.³² The survey also had embedded validation questions to verify participants’ attention. The current manuscript focuses on a subset of these measures collected in the larger survey. The study (Protocol HS-2020-0097) was reviewed and verified as exempt in accordance with university assurance and federal requirements pertaining to human subjects’ protections within the Code of Federal Regulations (45 CFR 46.104, April 24, 2020, IRB 042220).

Between April 23 – June 13, 2020, a total of 3,555 individuals accessed the survey. Those who did not identify as current undergraduate students (n=105), were not ≥18 years old (n=57), did not complete the first two survey modules (n=1,483), failed the validation checks (n=14), or had incomplete risk perception data (n=447) were excluded from the study sample. The remaining 1,449 students were included in the analytic sample.

Measures

Demographic Covariates

Students self-reported their gender (i.e., male, female, non-binary, prefer not to answer), age (18–24, 25–34, 35–44, 45–54, ≥55), current class standing (freshman, sophomore, junior, senior or 5+ years), major declaration, current geographic location (Zip code and County) and type of living arrangement (i.e., residence hall/fraternity/sorority, rented apartment/house, owned house/condominium, living with friend/family, or unhoused). Age (18–24, 25–34, ≥35 years old) and current geographic location (within the county, different county in California, and outside of California) were recategorized.

Perceived Coronavirus Threat, Risk Perception, and Efficacy Belief

To measure perceived COVID-19 threat (investigator-initiated), participants ranked each of the following four statements on a scale from 1 (not worried at all) to 4 (very worried): “you/someone you know will be exposed to coronavirus”; “the coronavirus outbreak in the US will worsen over the next few months”; “the coronavirus will have a negative economic effect on the US”; and “your company/organization’s operations will be affected by the coronavirus outbreak”. Responses across the four statements were averaged for each participant to calculate a summated index quantifying the overall perceived coronavirus threat. To measure situational risk perception (investigator-initiated), participants ranked how risky they perceived the following situations on a scale from 1 (not risky at all) to 5 (extremely risky): “traveling to an area where there is a current outbreak of coronavirus”; “not following self-quarantine orders to prevent the contraction of the coronavirus”; “hanging out with friends in groups of more than 10+”; “sharing a space with someone who contracted the coronavirus recently”; and “going to the beach, park, hiking trail, or somewhere in nature”. A summated index was calculated for each participant to quantify their overall situational risk perception. The perceived risk of contracting COVID-19 was measured on a scale from 1 (very unlikely) to 7 (extremely likely) for four different timepoints: the next 3, 6, and 12 months, and over the lifetime.^25,30 Lifetime perceived risk was dichotomized using the sample median (4.2) to classify high vs. low lifetime perceived risk. To quantify self-efficacy, participants rated their confidence in successfully engaging in prevention behaviors on a scale from 1 (not confident at all) to 7 (very confident).^25,30 Prevention behaviors of interest included avoiding close social contact, keeping their hands clean, avoiding non-essential travel, and cleaning/disinfecting frequently touched surfaces. Efficacy responses were averaged for a summated index of overall self-efficacy.

Adoption of Prevention Behaviors

To assess the adoption of prevention behaviors,³² participants were asked if they were adhering to the stay-at-home order, whether they only patronized essential businesses, and how often they went out per week (0, 1–3, 4–6, 7–9, 10+ times). Participants were also asked to report their personal response to the outbreak, including whether they were handwashing or cleaning more frequently, self-isolating, wearing personal protective equipment (PPE), or had done nothing differently.

Statistical Analysis

Descriptive characteristics of the study sample were presented by lifetime COVID-19 risk perception (dichotomized using the sample median, where lifetime risk>4.2 is high and lifetime risk≤4.2 is low). Students who self-identified as male or female were included for stratified gender analysis. Sample means and standard deviations were estimated for individual questions and summated indices of perceived COVID-19 threat, perceived situational risk, and self-efficacy and were stratified by class standing and gender. Last, differences in engagement in COVID-19 prevention behaviors were examined by COVID-19 lifetime perceived risk and gender. Linear regression was employed to evaluate significant trends by class standing, and chi-squared tests were performed to assess differences in categorical variables. An alpha level of 0.05 was used to establish statistical significance in differences and trends.

Analyses were performed in SAS Studio 3.8 and figures were created in RStudio 1.3.1093.

Results

Among the 1,449 undergraduate students included in the analysis, the majority were women (71.2%), aged 18–24 years old (86.6%) and a junior class standing (32.7%) (Table 1). The sample reported residing in the county (69.9%) or a different county in California (21.1%) at the time of the survey. Significant differences were observed in lifetime perceived risk; a greater proportion of students with low lifetime risk perception were 18–24-years-old (88.2% vs. 84.1%; p=0.02) and freshmen (21.3% vs. 16.3%; p=0.01) compared to students with high lifetime risk perception (Table 1).

Table 1.

Demographic Statistics of the Study Sample by Lifetime COVID-19 Risk Perception: California Undergraduate Student COVID-19 Survey (Apr–June 2020)

	Undergraduate Sample	High Lifetime Risk Perception	Low Lifetime Risk Perception
	n = 1,449	n = 563	n = 886
Variable	n (%)	n (%)	n (%)	p-value
Gender				0.52
Male	382 (26.6)	154 (27.5)	228 (26)
Female	1024 (71.2)	390 (69.6)	634 (72.2)
Non-Binary	22 (1.5)	11 (2)	11 (1.2)
Prefer not to Answer	10 (0.7)	5 (0.9)	5 (0.6)
Missing	11	3	8
Age				0.02
18–24 years old	1250 (86.6)	472 (84.1)	778 (88.2)
25–34 years old	149 (10.3)	64 (11.4)	85 (9.6)
≥ 35–44 years old	44 (3.1)	25 (4.5)	19 (2.2)
Missing	6	2	4
Class Standing				0.01
Freshman	281 (19.4)	92 (16.3)	189 (21.3)
Sophomore	256 (17.6)	100 (17.8)	156 (17.6)
Junior	474 (32.7)	176 (31.3)	298 (33.6)
Senior	376 (26)	172 (30.5)	204 (23)
> 5 years	62 (4.3)	23 (4.1)	39 (4.4)
Undergraduate Major				0.40
Science	512 (35.5)	206 (36.8)	306 (34.7)
Humanities	929 (64.5)	353 (63.2)	576 (65.3)
Missing	8	4	4
Current Residential Location				0.49
Within County	1003 (69.9)	390 (69.73)	613 (69.8)
California State (Outside County)	318 (21.1)	118 (21.26)	200 (22.8)
Outside of California	115 (9)	50 (9.01)	65 (7.4)
Missing	13	5	8
Current Living Arrangement				0.27
Campus Housing*	78 (5.4)	23 (4.1)	55 (6.2)
Rented Apartment/House	507 (35)	212 (37.7)	295 (33.3)
Owned House/Condominium	91 (6.3)	39 (6.9)	52 (5.9)
With family member/friend	759 (52.3)	284 (50.4)	475 (53.6)
Unhoused	4 (0.3)	1 (0.2)	3 (0.3)
Other	10 (0.7)	4 (0.7)	6 (0.7)

^*Residence hall, fraternity, or sorority house

Perceived COVID-19 risk varied significantly over time and by class standing, and overall was low (Figure 1). On a scale from 0 to 7, the average perceived risk of contracting COVID-19 in the next 3, 6, or 12 months was 3.1 (SD=1.5), 3.2 (SD=1.6) and 3.4 (SD=1.8), respectively; lifetime risk perception was significantly higher than the other timepoints (mean_life=4.1, SD=2.0). Risk perception of contracting COVID-19 demonstrated a significant increasing trend for each class standing as the time-point increased (p<0.001). As shown in Figure 1, freshman had the lowest average perceived risk for getting COVID-19 in the next 3, 6, 12 months or in their lifetime (mean_3m=2.7, SD=1.5; mean_6m = 3.0, SD=1.6; mean_12m=3.1, SD=1.8; and mean_life=3.8, SD=2.0, respectively) and seniors had the highest perceived risk (mean_3m=3.3, SD=1.5; mean_6m=3.4, SD=1.5; mean_12m=3.6, SD=1.8; and mean_life=4.4, SD=2.0, respectively). While risk perception trended upward as class standing increased, students in their 5+ year had lower risk perception compared to the seniors (mean_3m=3.1, SD=1.5; mean_6m=3.4, SD=1.5; mean_12m=3.6, SE=1.8; mean_life=4.1, SD=1.8). Women had a higher perceived risk of contracting COVID-19 in the next 3 months compared to men (mean_3m=3.1, SD=1.5 vs. mean_3m=3.0, SD=1.6), but they had a lower perceived risk at every other time point (mean_6m=3.2, SD=1.5 vs. mean_6m=3.3, SD=1.6; mean_12m=3.3, SD=1.8 vs. mean_12m=3.6, SD=1.8; mean_life=4.1, SD=2.0 vs. mean_life=4.2, SD=2.0; data not presented).

Figure 1.

Perceived Risk of Contracting COVID-19 among California Undergraduate Students

Class standing was significantly associated with a positive trend in the summated indices for perceived coronavirus threat (p<0.001) and perceived situational risk (p<0.0001) (Table 2). The mean overall perceived coronavirus threat was 3.2 (SD=0.6) out of 4; the perceived threat was lowest among freshmen (mean=3.1, SD=0.6) and trended up to 3.3 (SD=0.6) among seniors and 5+ students (Table 2). Undergraduate students were most worried about the outbreak having a negative economic impact on the US (mean=3.4, SD=0.8). The mean overall situational risk perception was 4.1 (SD=0.7) out of 5 and ranged from 3.9 (SD=0.7) among freshman to 4.2 (SD=0.5) among +5^th year students. Students had the highest perceived risk of sharing a living space with someone who had contracted the coronavirus recently (mean=4.6, SD=0.8) and the lowest perceived risk of going to the beach, park, hiking trails, or somewhere in nature (mean=2.8, SD=1.1; Table 2). As presented in Table 2, overall self-efficacy was high in the sample (mean=5.9, SD=1.0 out of 7) and no statistically significant trend by class standing was observed. Undergraduate students had the lowest confidence in their ability to avoid close social contact (mean=5.6, SD=1.4; Table 2).

Table 2.

Perceived Coronavirus Concern, Situational COVID Risk and Self-Efficacy by Class Standing: California Undergraduate Student COVID-19 Survey (Apr–June 2020)

	Total	Freshman	Sophomore	Junior	Senior	5+ Years
	(n = 1,449)	(n = 281)	(n = 256)	(n = 474)	(n = 376)	(n = 62)
Variable	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	p-value†
Perceived Coronavirus Concern	(n = 1,337)	(n = 259)	(n = 226)	(n = 440)	(n = 354)	(n = 58)
Please indicate the extent to which you are worried with the following statements about how you feel regarding the coronavirus…
You or someone you know will be exposed to coronavirus	3.1 (0.9)	2.9 (0.9)	3 (0.9)	3.1 (0.8)	3.2 (0.8)	3.2 (0.8)	<0.001
The coronavirus outbreak in the US will worsen over the next few months	3.3 (0.8)	3.2 (0.8)	3.2 (0.9)	3.3 (0.8)	3.3 (0.8)	3.4 (0.7)	0.003
The coronavirus outbreak will have a negative economic effect on the US	3.4 (0.8)	3.3 (0.8)	3.3 (0.8)	3.4 (0.8)	3.5 (0.7)	3.3 (0.8)	0.02
Your company or organization’s operations will be affected by the coronavirus outbreak	3.1 (1)	2.9 (1)	3.1 (1)	3.1 (1)	3.1 (1)	3.2 (1)	0.01
Perceived Coronavirus Concern Index	3.2 (0.6)	3.1 (0.6)	3.1 (0.6)	3.2 (0.6)	3.3 (0.5)	3.3 (0.6)	<0.001
Perceived Situational Risk	(n = 1,413)	(n = 276)	(n = 251)	(n = 462)	(n = 365)	(n = 59)
For each of the following statements, please indicate how risky you perceive each situation…
Traveling to an area where there is a current outbreak of the coronavirus	4.4 (0.9)	4.3 (0.9)	4.4 (0.9)	4.5 (0.9)	4.5 (0.9)	4.5 (0.8)	0.02
Not following self-quarantine orders to prevent the contraction of the coronavirus	4.2 (1)	4 (1)	4.1 (1.1)	4.3 (0.9)	4.2 (1)	4.2 (1)	0.002
Hanging out with friends in groups of 10+	4.4 (0.9)	4.2 (1)	4.3 (1.1)	4.4 (0.9)	4.5 (0.9)	4.5 (0.8)	<0.001
Sharing a living space with someone who has contracted the coronavirus recently	4.6 (0.8)	4.5 (0.9)	4.5 (0.8)	4.6 (0.8)	4.6 (0.7)	4.7 (0.6)	0.002
Going to the beach, park, hiking trail, or somewhere in nature	2.8 (1.1)	2.6 (1.1)	2.6 (1.2)	2.9 (1.1)	2.9 (1.1)	2.9 (1)	<0.001
Perceived Situational Risk Index	4.1 (0.7)	3.9 (0.7)	4 (0.8)	4.1 (0.7)	4.1 (0.7)	4.2 (0.5)	<0.001
Efficacy	(n = 1,416)	(n = 275)	(n = 249)	(n = 464)	(n = 367)	(n = 61)
I am confident in my ability to successfully…
Avoid close social contact	5.6 (1.4)	5.5 (1.5)	5.5 (1.5)	5.6 (1.4)	5.6 (1.5)	5.7 (1.4)	0.35
Keep my hands clean	6.2 (1)	6.2 (1.1)	6.2 (1.1)	6.3 (1)	6.2 (1)	6.3 (0.9)	0.34
Avoid non-essential travel	5.9 (1.4)	6 (1.4)	5.9 (1.5)	6 (1.4)	5.9 (1.4)	6.2 (1.4)	0.90
Clean/disinfect frequently touched surfaces	5.8 (1.3)	5.7 (1.3)	5.7 (1.4)	5.9 (1.3)	5.7 (1.4)	5.6 (1.4)	0.29
Efficacy Belief Index	5.9 (1)	5.8 (1)	5.8 (1)	6 (0.9)	5.8 (1)	5.9 (0.9)	0.36

^a.Response Options from 0 (Not Worried at All) to 4 (Very Worried);

^b.Response Options from 0 (Not Risky at All) to 5 (Extremely Risky);

^c.Response Options from 0 (Not Confident at All) to 7 (Very Confident)

^†P-Value for Trend

Significant differences in perceived Coronavirus concern, situational COVID Risk, and self-efficacy were also observed by living situation and geography. Those living in a residence hall or fraternity/sorority house averaged the lowest perceived concern (mean=3.0, SD=0.6; p<0.01) and situational COVID risk (mean=3.7, SD=0.8; p<0.001) compared to those residing in other living situations. Undergraduate students living within the study county averaged the highest perceived concern (mean=3.2, SD=0.6; p<0.001), perceived situational risk (mean=4.1, SD=0.7; p<0.001) and self-efficacy (mean=5.9, SD=1.0; p<0.001) compared to students living in other parts of California or outside of California.

Women had significantly higher overall perceived threat (mean=3.2, SD=0.5 vs. mean=3.1, SD=0.6; p=0.01), perceived situational risk (mean=4.1, SD=0.9 vs. mean=3.9, SD=0.8; p<0.001), and self-efficacy (mean=6.0, SD=0.9 vs. mean=5.6, SD=1.1; p <0.001) compared to men, as presented in Table 3. Men and women perceived the negative economic consequences of the pandemic as the greatest threat (mean=3.4, SD=0.8) with no statistical difference by gender; women had higher perceptions that the coronavirus outbreak would worsen in the next few months (mean=3.3, SD=0.7 vs. mean=3.0, SD=0.9; p<0.001, Table 3). Women perceived situations as significantly riskier compared to men; not following self-quarantine orders (4.3±0.9 vs. 3.9±1.2; p<0.001) and hanging out in groups ≥10 friends (mean=4.5, SD=0.8 vs. mean=4.1, SD=1.0; p<0.001) demonstrated the greatest gender differences (Table 3). Additionally, women had significantly greater self-efficacy (p<0.001); women were more confident in their ability to avoid non-essential travel (6.1±0.3 vs. mean-5.6, SD=1.6; p<0.001) and to clean/disinfect frequently touched surfaces (mean=5.9, SD=1.2 vs. mean=5.3, SD=1.5; p <0.001) (Table3).

Table 3.

Perceived Coronavirus Concern, Perceived Situational Risk and Self-Efficacy Belief by Gender: California Undergraduate Student COVID-19 Survey (Apr–June 2020)

	Total	Women	Men
	(n = 1,406)	(n = 1,024)	(n = 382)
Variable	Mean (SD)	Mean (SD)	Mean (SD)	p-value†
Perceived Coronavirus Concern	(n = 1,299)	(n = 951)	(n = 348)
Please indicate the extent to which you are worried with the following statements about how you feel regarding the coronavirus…
You or someone you know will be exposed to coronavirus	3.1 (0.9)	3.1 (0.8)	2.9 (0.9)	<0.001
The coronavirus outbreak in the US will worsen over the next few months	3.3 (0.8)	3.3 (0.7)	3.0 (0.9)	<0.001
The coronavirus outbreak will have a negative economic effect on the US	3.4 (0.8)	3.4 (0.8)	3.4 (0.8)	0.09
Your company or organization’s operations will be affected by the coronavirus outbreak	3.1 (1)	3.1 (1)	3 (1)	0.69
Perceived Coronavirus Concern Index	3.2 (0.6)	3.2 (0.5)	3.1 (0.6)	0.009
Perceived Situational Risk	(n = 1,372)	(n = 1,003)	(n = 369)
For each of the following statements, please indicate how risky you perceive each situation…
Traveling to an area where there is a current outbreak of the coronavirus	4.4 (0.9)	4.5 (0.8)	4.2 (1)	<0.001
Not following self-quarantine orders to prevent the contraction of the coronavirus.	4.2 (1)	4.3 (0.9)	3.9 (1.2)	<0.001
Hanging out with friends in groups of more than 10+	4.4 (0.9)	4.5 (0.8)	4.1 (1.1)	<0.001
Sharing a living space with someone who has contracted the coronavirus recently.	4.6 (0.8)	4.6 (0.8)	4.5 (0.9)	0.01
Going to the beach, park, hiking trail, or somewhere in nature.	2.8 (1.1)	2.8 (1)	2.6 (1.2)	0.002
Perceived Situational Risk Index	4.1 (0.7)	4.1 (0.7)	3.9 (0.8)	<0.001
Self-Efficacy	(n = 1,375)	(n = 1,002)	(n = 373)
I am confident in my ability to…
Successfully avoid close social contact	5.6 (1.4)	5.6 (1.4)	5.4 (1.6)	0.04
Successfully keep my hands clean	6.2 (1)	6.3 (1)	6 (1.1)	<0.001
Successfully avoid non-essential travel	6 (1.4)	6.1 (0.3)	5.6 (1.6)	<0.001
Successfully clean/disinfect surfaces	5.8 (1.3)	5.9 (1.2)	5.3 (1.5)	<0.001
Self-Efficacy Index	5.9 (1)	6 (0.9)	5.6 (1.1)	<0.001

^a.Response Options from 0 (Not Worried at All) to 4 (Very Worried);

^b.Response Options from 0 (Not Risky at All) to 5 (Extremely Risky);

^c.Response Options from 0 (Not Confident at All) to 7 (Very Confident)

^†P-Value from Chi-Squared Test

There was high endorsement of prevention behaviors in the undergraduate sample. The majority of students reported adhering to stay-at-home orders (93.3%), visiting only essential businesses (84.1%), and going out 1–3 times per week (71.2%). Among students who did not adhere to the stay-at-home order (n=96), the most common reasons for not complying included: work (33.3%), wanting/needing to see family/friends (12.5%), low perceived risk (11.5%), and mental/physical health (10.4%) (data not shown). Despite high endorsement of prevention behaviors, significant gender differences were observed (Table 4). More women reported adhering to the self-quarantine order (94.9% vs. 88.4%; p<0.001), going only to essential businesses (85.9% vs. 79.6%, p<0.001) and going out 1–3 times/week (74.6% vs. 62.3%; p<0.001) compared to men (Table 4). Regarding personal response to the outbreak, more women adopted each of the prevention behaviors. There were no observed significant differences by high vs. low perceived lifetime COVID-19 risk (data not shown).

Table 4.

Adoption of WHO Prevention Behaviors by Gender: California Undergraduate Student COVID-19 Survey (Apr–June 2020)

	Total	Women	Men
	(n = 1,406)	(n = 1,024)	(n = 382)
Variable	n (%)	n (%)	n (%)	p-value
Adhering to Self-Quarantine Order				<0.001
Yes	1292 (93.2)	957 (94.9)	335 (88.4)
No	95 (6.9)	51 (5.1)	44 (11.6)
Missing	18	15	3
Going out Strictly to Essential Businesses				<0.001
Yes	1078 (84.2)	813 (85.9)	265 (79.6)
No	202 (15.8)	134 (14.2)	68 (20.4)
Missing	12	10	2
Number of times going out per week				<0.001
0 times	978 (71.2)	745 (74.6)	233 (62.3)
1 – 3 times	288 (21)	196 (19.6)	92 (24.6)
4 – 6 times	71 (5.2)	44 (4.4)	27 (7.2)
7 – 9 times	22 (1.6)	5 (0.5)	17 (4.6)
10+ times	14 (1)	9 (0.9)	5 (1.3)
Missing	33	25	8
Personal Response to the Outbreak *
Increased hand washing frequency	1230 (89.5)	918 (91.8)	312 (83.4)	<0.001
Cleaning more frequently	1006 (73.2)	780 (78)	226 (60.4)	<0.001
Self-isolated	1191 (86.7)	889 (88.9)	302 (80.8)	<0.001
Wear masks, gloves, or other PPE	1294 (94.2)	959 (95.9)	335 (89.6)	<0.001
Other	361 (26.3)	223 (22.3)	138 (36.9)	<0.001
Did nothing differently	17 (1.2)	7 (0.7)	10 (2.7)	0.003

^aAmong participants who reported adhering to self-quarantine order (n=1,292)

^*Check all that Apply

Discussion

The findings of this study are pertinent and applicable towards efforts to mitigate transmission of COVID-19 and to roll-out vaccination programs in undergraduate students living across the US. The analyses found low perceived risk of contracting COVID-19 in the undergraduate sample at each of the considered timepoints (3-months, 6-months, 12-months, and lifetime) with a significant increasing trend in perceived risk as the time increased. High lifetime risk perception was significantly associated with age and class standing, where higher proportions of the low lifetime risk perception group were freshman and in the youngest age group. In addition to having the lowest perceived risk of contracting COVID-19, freshman also had the lowest perceived coronavirus threat and situational risk perception; self-efficacy was high in the sample and did not demonstrate a significant trend by class standing. Perceived lifetime risk was not significantly associated with differences in risk reduction behaviors (i.e., adhering to the quarantine order, visiting only essential businesses); however, more students with high lifetime risk perception enacted a personal prevention response. Despite undergraduate women having lower perceived risk of contracting COVID-19 at most timepoints, they showed higher perceived coronavirus threat, higher situational risk perception, and higher self-efficacy compared to undergraduate men. Additionally, women were more likely to adhere to the mandated stay-at-home order, only go to essential businesses, and take a personal response to the outbreak.

While COVID-19 risk perception among college students in China during quarantine was high,³³ our study was more consistent with an online study conducted among the general population in the US at the beginning of the pandemic.³⁴ On a scale from 0–100, participants in the online study rated their risk perception of contracting COVID-19 as 43.1 and perceived themselves as having lower risk of getting infected than the average person,³⁴ consistent with unrealistic optimism bias. Studies have found high endorsement of prevention behaviors among the US and Chinese general populations;^34–35 Wise et al.³⁴ found that participants with higher perceived risk of contracting the virus were more likely to engage in hand washing and social distancing after adjusting for age.³⁴ Another study also found that high perceived COVID-19 severity was positively correlated with compliance to non-pharmaceutical intervention, as was constitutionalism, news exposure, and religiosity.³⁶ While the current study of undergraduate students finds similarly low perceived risk of contracting COVID-19 and high overall adoption of prevention behaviors, the association between high perceived risk and risk reduction behaviors was not observed in our study. The Wise et al. study³⁴ was performed in a broad population across the USA, which while young (median age=30), may not be representative of the undergraduate population.³⁴

Additionally, differences in local guidelines and public health orders vary across geographies where participants reside, which was not controlled for in the Wise study.³⁴ Our study found that students living within the study county had higher perceived concern, situational risk, and self-efficacy compared to students living in other counties or outside of California, which underscores the role of local policy on a student’s health perceptions. Along with geographic location, there were observed differences by living situation in the study. While only around 5% of the study population was living in residence halls or fraternity/sorority houses at the time of data collection, this group had lower perceived Coronavirus concern and perceived situational risk. Approximately 80% of students in this group were freshman or sophomore, who as a group had the lowest concern and risk perceptions. Bigouette et al.¹² demonstrated an increased risk of contracting and transmitting COVID-19 among individuals sharing a room or living space, and their estimates showed that 75% and 82% of students in a dormitory share a room or living space, respectively. The combination of increased risk from the living situation and having low risk perceptions puts this group at particularly high risk for infection and subsequent transmission of COVID-19 to the broader community, and as such, is a critical group to target for non-pharmaceutical interventions and vaccination initiatives.

This study provides valuable insights into to the health perceptions and behaviors of a high-risk group, which has implications for the strategy and messaging of prevention programs in the future. A study among South Carolinian undergraduates found that greater perceived disease susceptibility and lower negative attitudes towards vaccines were associated with greater COVID-19 vaccine acceptance, while having higher perceived risk of COVID-19 exposure was associated with lower vaccine acceptance – the latter being consistent with unrealistic optimism bias.³⁷ Kecojevic et al.³⁸ found in the summer of 2021 that 53% of Northern New Jersey students (undergraduate and graduate level) had the intention to get vaccinated against COVID-19; vaccination intention was associated with older age, White race, high COVID-19 knowledge, trust in official sources/news media, and having a vaccinated family member.³⁸ Our study provides relevant context for targeting high risk students who have low vaccine acceptance and who may be less likely to get vaccinated and/or adhere to non-pharmaceutical interventions.

Existing literature largely shows that high risk perception is associated with greater enactment of prevention behaviors.^21–23 In addition to subjective risk perception, the psychology of decision making, or choice, also plays a role in the adoption of health behaviors.^39–40 Relevant context for our study is that California issued a statewide stay-at-home order on March 19, 2020;⁵ non-essential businesses were closed, and many California universities transitioned to virtual learning, in some cases the week prior to the order.⁴ As such, during the time that our study was conducted, participants may not have had perceived choice or decision-making capacity in an environment where statewide and local regulations were in place. Regardless of risk perception, there was less choice in the matter of engaging in the risk reduction behaviors in our study context. This may, in part, explain the lack of significant association between risk perception and prevention behavior adoption in our study sample. As evidenced by the Wise³⁴ et al. study, engagement in these behaviors can change in as little as a week, so as public health mandates and guidance change, one may expect these behaviors to change as well.³⁴

The current study also found significant differences in coronavirus perceptions by gender. Others find that women tend to have higher perceived risk than men.⁴¹ Women are more likely to perceive health risks of cigarette smoking for themselves and others more highly than men; additionally, college women have higher perceived risk of sexual behaviors, alcohol, and drug-use than college men.⁴¹ However, in the current study, men had higher perceived risk of contracting coronavirus than women. Hitchcock⁴¹ and others note that the lack of differential unrealistic optimism by gender with outcomes such as cancer or nuclear war leads us to believe there are other underlying factors driving the differences other than gender.⁴¹ One potential factor for the differences in risk perception may be race/ethnicity and cultural differences, as proposed by Niño et al.⁴² Women had higher situational risk perception, perceived threat, and self-efficacy, as well as significantly greater engagement in prevention behaviors, but had lower risk perception compared to the men surveyed in the study. Not only were women significantly more likely to adhere to the mandated guidance, but they were also more likely to adopt personal responses to the outbreak, which suggests that women were more likely to follow health advice and guidelines. This is consistent with a study among Ethiopian health care workers (median age=30) which found that men were 2.5 times more likely to have low level of preparedness for the COVID-19 pandemic than women.⁴³ As posited by Brewer et al.,⁴⁴ the enactment of prevention behaviors may reduce risk perceptions, which is consistent with the differences observed in our study between men and women. These findings suggest that freshmen men in particular are in the greatest need for COVID-19 tailored messaging to change perceptions and behaviors. This group is in the first year of a transitional period, and therefore, they may underestimate risks and believe that they are less likely to experience negative events (i.e., unrealistic optimism bias); as such, they should be targeted in future prevention efforts.

Some limitations of this study merit discussion, namely the study is a single cross-sectional survey with self-reported measures, which is subject to social desirability bias. The study was conducted within the early months of the pandemic at only one time point, and as such, changes in risk perception and enactment of prevention behaviors could not be evaluated. Risk perceptions are influenced by various factors, such as an individual’s numeracy and ability to parse new information, personal experience, salient information about the threat, contextual factors, and one’s general affect.⁴⁴ During the time of this survey, information about COVID-19 was highly accessible and featured prominently on most news networks; in California, the state government perceived the threat was sufficiently serious for a stay-at-home mandate and for universities to transition to virtual learning. Despite the high perceived threat-level, the surveyed undergraduate students still had low risk perception. While changing information about COVID-19 may change deliberative (or logical) risk perception, affective and experiential risk perceptions tend to be more predictive of protective behaviors.⁴⁵ As such, changing health behaviors necessitates specific interventions that not only target deliberative risk perception, but more importantly, affective and experiential risk perceptions.⁴⁵ Without tailored interventions, it is unlikely that this population would have experienced drastic changes in risk perception or associated prevention behaviors from the time that the survey was completed.

Additionally, the survey was conducted among undergraduate students at a public university in California and did not include students at other types of schools or localities. While this age group is similar neurodevelopmentally regardless of the type of college enrollment, behaviors and influences may be different. Next, 6% of the undergraduate population at the sampled university were included in the analysis; compared to the university and national demographics, our sample overrepresented women (70.6% vs. 55.5% vs. 56.6%, respectively). Based on the existing literature which suggests that men have lower risk perception compared to women,³⁴ having men appropriately represented would have likely lowered the overall risk perception of the sample and led to greater gender differences. Last, the current study did not collect information regarding race/ethnicity, religion, or other socioeconomic factors, and differences in risk perceptions and behaviors could not be evaluated. Thus, the inability to generalize the data to certain sub-groups is a limitation of the study.

Despite these limitations, our study has many strengths. The survey was conducted among a priority population during a period in the pandemic when health orders were mandated in California. Identifying those who, despite being required, did not comply with health mandates helps to understand the profile for those students who are less likely to enact (or maintain) prevention behaviors when mandates are lifted. Additionally, this study looks at factors that inform health behaviors– such as risk perception, perceived threat, and self-efficacy– thus enabling institutions to tailor messaging to the identified gaps and groups in greatest need of intervention. Despite approval and distribution of COVID-19 vaccines, pandemic fatigue and the emergence of COVID-19 variants highlight the unrelenting effort necessary to stop the spread of COVID-19 in the USA. This study provides insight into the perceptions and beliefs of college students, an integral constituent of COVID-19 mitigation efforts.

Conclusion

In our study, undergraduate students had low risk perception of contracting COVID-19. Freshmen showed the lowest perceived concern, situational risk perception, and perceived risk of contracting COVID-19 compared to upperclassmen. Despite men having higher risk perception of contracting COVID-19 than women, they endorsed adhering to the health mandates and engaging in prevention behaviors significantly less. As such, freshmen men have the greatest need for tailored messaging to help curb the spread of COVID-19 in the university setting and surrounding communities. Future studies should consider racial/ethnic and cultural differences in risk perceptions and self-efficacy in the college population, which have been shown to influence health perceptions and behaviors. Understanding how the college population is seeking and acquiring COVID-19 related information will help testing and vaccination education efforts in the future.