403 Descriptive Statistics in Journal Article

Descriptive Statistics in Journal Article Exercise

Prepare and provide an APA-format citation for this article:

Using the article’s Results section, find one example of a descriptive statistic that was presented and answer the set of questions below about this descriptive statistic.

1. List the descriptive statistic you chose.

2. What variable was being measured?

3. Discuss how that variable was measured?

4. What was the level of measurement for that variable? Explain why it fits under this level of measurement.

5. Given the level of measurement for the variable, why was that descriptive statistic used? Was it the appropriate descriptive statistic to use for this variable? Explain why or why not. If not, explain what a more appropriate descriptive statistic would have been.

Health Behaviors at the Onset of the
COVID-19 Pandemic
Raven H. Weaver, PhD
Alexandra Jackson, MS, RDN, CDCES
Jane Lanigan, PhD
Thomas G. Power, PhD
Alana Anderson, MS
Anne E. Cox, PhD
Linda Eddy, PhD, RN, ARNP
Louise Parker, PhD
Yoshie Sano, PhD
Elizabeth Weybright, PhD
Objectives: We examined perceived behavior change since implementation of physical distancing restrictions and identified modifiable (self-rated health, resilience, depressive symptoms,
social support and subjective wellbeing) and non-modifiable (demographics) risk/protective
factors. Methods: A representative US sample (N = 362) completed an online survey about
potential risk/protective factors and health behaviors prior to the pandemic and after implemented/recommended restrictions. We assessed change in perceived health behaviors prior to
and following introduction of COVID-19. We conducted hierarchical linear regression to explore
and identify risk/protective factors related to physical activity, diet quality, and social isolation.
Results: There have been substantial decreases in physical activity and increases in sedentary
behavior and social isolation, but no changes in diet quality since COVID-19. We identified modifiable and non-modifiable factors associated with each health behavior. Conclusions: Negative
effects indicate the need for universal intervention to promote health behaviors. Inequalities in
health behaviors among vulnerable populations may be exacerbated since COVID-19, suggesting need for targeted invention. Social support may be a mechanism to promote health behaviors. We suggest scaling out effective health behavior interventions with the same intensity in
which physical distancing recommendations were implemented.
Key words: COVID-19; pandemic; risk/protective factors; physical activity; diet; social isolation
Am J Health Behav.™ 2021;45(1):44-61
DOI: doi.org/10.5993/AJHB.45.1.4
O
n March 11, 2020, the World Health Organization declared COVID-19 a global
pandemic.1 Profound changes due to the
pandemic, including ongoing physical (or social)
distancing recommendations, increases in unemployment, and therefore, changes in financial status,
and exacerbated racial and class inequities, likely
will have significant detrimental impact on long-
term health and wellbeing. In addition to the fear,
anxiety, and stress associated with disease outbreaks,
the implementation of physical distancing recommendations has altered interpersonal interactions.
These changes may impact health-related behaviors
and have long-term implications for physical and
psychosocial health, as evidence suggests long quarantines lead to poor psychological outcomes.2
Raven H. Weaver, Assistant Professor, Human Development, Washington State University, Pullman, WA, United States. Alexandra Jackson, PhD
Candidate, Human Development, Washington State University, Vancouver, WA, United States. Jane Lanigan, Professor, Human Development,
Washington State University, Vancouver, WA, United States. Thomas G. Power, Emeritus Professor, Human Development, Washington State University, Pullman, WA, United States. 5th authors listed in alphabetical order: Atlana Anderson, PhD Candidate, Human Development, Washington
State University, Pullman, WA, United States. Anne E. Cox, Associate Professor, Kinesiology and Educational Psychology, Washington State University, Pullman, WA, United States. Linda Eddy, Professor, Nursing, Washington State University, Vancouver, WA, United States. Louise Parker, Professor, Human Development and Extension Youth and Families Unit, Washington State University, Seattle, WA, United States. Yoshie Sano, Associate
Professor, Human Development, Washington State University, Vancouver, WA, United States. Elizabeth Weybright, Associate Professor, Human
Development, Washington State University, Pullman, WA, United States.
Correspondence Dr Weaver; raven.weaver@wsu.edu
44
Weaver et al
We used a prevention science framework to examine perceived behavior changes since COVID-19
and identify potential modifiable and non-modifiable risk/protective factors associated with current
health behaviors. Modifiable risk factors have the
potential to change through prevention programs
(eg, self-rated health, resilience, social support).
Conversely, non-modifiable factors include predisposing (ie, age, race/ethnicity) and stable (ie, sex,
geographic location) factors less likely to be altered
by prevention programming, yet associated with
an unequal distribution and burden of COVID-19
morbidity and mortality.3 Higher mortality rates
from COVID-19 among vulnerable populations
have highlighted structural and cumulative disadvantages that intensify health inequality between
racial/ethnic groups. It is important to identify
social determinants of health that shape individuals’ opportunities and health outcomes during the
pandemic. Social determinants of health can serve
as key targets for prevention efforts to support positive health behaviors. Prevention science postulates
individual, intermediate, and structural factors either increase the risk for adverse health outcomes,
or are protective factors that mediate or moderate
exposure to risk and/or decrease the likelihood.4
We focus on health-related behaviors including
physical activity, diet quality, and social isolation,
which influence long-term health and reduce risk
for chronic disease and mortality.5-7
Hispanic individuals are more likely to engage in
occupational activity.11,12 White females and older
adults, as well as individuals who had lower income, experienced depression, and lived in a rural
area were at higher risk of regular physical inactivity.11,13-15 In addition, there are structural and
systematic factors serving as barriers to physical activity in minority populations.16 Life transitions associated with decreases in physical activity include
changes in employment status, residence, and family composition.17,18
Sedentary behavior has been associated with negative outcomes, including discrete impacts on health,
because of less physical activity.19 Therefore, most
characteristics positively associated with physical activity are inversely associated with sedentary behavior.13 Similar to physical activity, sedentary behavior
tends to remain unchanged throughout life stages.18
However, if COVID-19-related restrictions become
protracted, access to spaces to be physically active
may decrease, time at home may increase, and participation in concomitant sedentary behaviors, such
as screen time, may escalate as well.
Physical Activity
Achieving recommended amounts of physical
activity is a protective factor associated with a multitude of health benefits.8 In contrast, low levels of
physical activity increases risk of chronic disease.9
Prior to the pandemic, most adults in the United
States (US) did not meet the recommended physical activity guidelines.8 Current recommendations
during the COVID-19 pandemic acknowledge the
benefits of physical activity, encouraging physical
activity while maintaining physical distancing.10
With stay-at-home orders and closures of parks and
fitness centers, researchers anticipate reductions in
physical activity.5
Researchers have identified individuals that are
male, younger, White, married, employed, and
having high levels of social support are more likely
to participate regularly in physical activity while
Diet Quality
Healthier dietary patterns are associated with reduced risk for multiple chronic diseases, many of
which are associated with risk of hospitalization
due to COVID-19.20 Specifically, increased fruit
and vegetable consumption and decreased added
sugar consumption are important characteristics of
healthy eating patterns.6 Higher fruit and vegetable
intake is preventive of cardiovascular disease and
cancer, inversely associated with the risk for depression, and closely related to psychological wellbeing
and happiness.6,21 Whereas the benefits of fruit and
vegetable intake are important, about 75%-85%
of individuals living in the US do not meet the
recommended intake. Fruit and vegetable intake
differs by sex, income status, geographic location,
employment status, age, and is higher among women (per 1000 calories) and individuals with higher
income.22 Urban populations consume an average
of nearly one-third servings more than rural populations.23 Fruit and vegetable intake is positively
associated with employment status.24 Households
without reliable access to affordable, nutritious
food may be food insecure, which is a significant
risk factor for low diet quality – specifically de-
Am J Health Behav.™ 2021;45(1):44-61
DOI: doi.org/10.5993/AJHB.45.1.4
45
Health Behaviors at the Onset of the COVID-19 Pandemic
creased intake of fruits and vegetables.25 Whereas
fruit intake increases through adulthood, vegetable
intake tends to increase from 20-59 years old and
then maintain throughout older adulthood.22
Added sugar intake is associated with the consumption of less nutrient-dense foods, and increased likelihood of dental caries, type 2 diabetes,
cardiovascular disease, and risk for depression.26-28
Lower added sugar intake may be associated with
improved psychological health.28 Between 2005
and 2010, over 70% of Americans exceeded the
recommendations of less than 10% of total calories per day from added sugars.29 Added sugar intake is higher among men, those with lower family
income, and physically inactive individuals.27,30 In
adulthood, added sugar intake tends to peak in
early adulthood (18-24-year-olds) and is lowest in
adults 40 years old and older.30 Added sugar intake
is lowest among non-Hispanic individuals, Asian
Americans, and families with higher income and
education levels.27,30
due to becoming unemployed. These disruptions to
social health and wellbeing may increase risk of experiencing social isolation. Social isolation is defined
as “inadequate quality and quantity of social relations with other people at the different levels where
human interaction takes place (individual, group,
community and the larger social environment)”42
and has been associated with detrimental health
outcomes. More specifically, this includes poor selfrated health,43 psychosocial health,43-45 cognitive
functioning,46 and risk of mortality.44 Risk factors
and experiences of social isolation vary across the
lifespan.47 Older adults are particularly vulnerable43
but younger age groups are also at risk of experiencing social isolation.48 Researchers have found
that race and sex moderate the detrimental impact
of social isolation. Social isolation had a greater association on all-cause mortality among Blacks, was
a significant factor for death from cancer among
Whites, and had less impact on cardiovascular disease for White males than other subgroups.49
Social Isolation
Social health encompasses the ability to form effective relationships, fulfill social roles, and adapt to
the social environment,31 and is a key component
of individual health and wellbeing across the lifespan. Social health can buffer against or worsen one’s
ability to adapt and manage stress or the challenges
of daily living and is associated with numerous psychosocial and physical health outcomes.32 Although
social connection is protective against a range of
disease morbidities and all-cause mortality,7 social
isolation is a risk factor associated with depression,
poorer cardiovascular health, and mortality.33,34
Both the quality and quantity of social interactions
have been associated with health and wellbeing,
with quality typically the stronger predictor.35 Positive social interactions affect physiological processes
such as immune functioning and inflammation,36,37
circulation of stress hormones,38 and health outcomes including cardiovascular disease,36 diabetes,39
depression,40 and post-traumatic stress disorder.41
Physical distancing measures associated with COVID-19 changed individuals’ social worlds by limiting in-person gatherings, suspending or limiting
operations of many social contexts, and redefining
personal roles such as parents adding the role of
teacher or losing the role of primary wage earner
Current Evidence of COVID-19 and Changes
to Health Behaviors
The evidence suggests contextual changes implemented/recommended to mitigate the spread
of COVID-19 have led to substantial changes in
health behaviors. There have been negative impacts on physical activity, mixed findings for dietary behaviors, and indications that COVID-19
has contributed to psychological distress. Several
international studies indicate decreases in physical
activity and increases in sedentary behavior during home confinement.50,51 Positive dietary behaviors include improving diet quality50,52 and eating
at home,53-55 while potentially negative behaviors
include increasing consumption of snacks and unhealthy foods and eating due to stress.50,55,56 Two
studies suggest negative emotional impacts of COVID-19 including distress57 and increases in work,
financial, and family stress.58 However, data on
social connection and COVID-19 indicate no significant changes in social isolation or support from
January 2020 to April 2020,59 although differences
existed by age, with older adults reporting less social isolation and greater social support, compared
to middle aged or young adults. These findings
align with past research on psychological impacts
of quarantine in previous disease outbreaks,2 im-
46
Weaver et al
pacts on health after experiencing a natural disaster,60 and after September 11, 2001 in New York.61
They also illuminate changes in health behaviors
during COVID-19 that may be worrisome and require further inquiry, particularly as relatively few
studies have examined the psychological impact of
physical distancing on health behaviors, and fewer
still have used validated measures.2
Purpose of Study
The purpose of this study was to assess perceived
changes to health behaviors before and during COVID-19 restrictions and identify modifiable and
non-modifiable risk/protective factors associated
with health behaviors during COVID-19. Our first
research question (RQ1) was: Among adults, how
have health behaviors (physical activity, diet quality,
and social isolation) changed with the implementation of physical distancing recommendations?
Based on the previous literature, we hypothesized
physical activity and fruit/vegetable intake will decrease, and sedentary behavior, added sugar intake,
and social isolation will increase. Finally, examining COVID-19 as a unique historical context, we
conducted an exploratory analysis to identify modifiable (self-rated health, resilience, depressive symptoms, social support, and subjective wellbeing) and
non-modifiable (demographics) risk/protective factors associated with physical activity, diet quality,
and social isolation. Our RQs were: In the context
of COVID19, what risk/protective factors are associated with physical activity? (RQ2); what risk/
protective factors are associated with diet quality?
(RQ3); and what risk/protective factors are associated with social isolation? (RQ4). As an exploratory
analysis, we did not develop a priori hypotheses for
risk/protective factors associated with physical activity, diet quality, or social isolation.
of the survey. We obtained a representative sample
with 402 participants, of which 362 completed at
least Part 1 and 2 of the survey and met the criteria
for completing survey attention checks. There were
no statistically significant differences in demographic characteristics between participants who
completed at least 2 parts of the survey and those
who did not. A post hoc power analysis indicated
that a sample of 362 participants was sufficient for
the planned analysis.64
Prior to completing Part 1, participants consented to participate in the 3-part survey. Part 1
was posted during April 21-23, 2020 and included
demographic, health history information, COVID-19 and physical distancing recommendations,
measures of social isolation and support, subjective
wellbeing, and resiliency. Part 2 was posted during
April 28-May 6, 2020 and included measures assessing eating behavior, physical activity, financial
stress, and employment. Part 3 was posted during
May 1-May 6, 2020. To reduce participant burden, parents of children between 18 months and
12 years old completed measures about family and
child health-related behaviors and parenting dimensions while all other participants completed a
dietary questionnaire.
Measures
We collected assessments of perceived health
behaviors (each dependent variable) prior to the
COVID-19 pandemic (retrospectively) and after
COVID-19 restrictions were implemented/recommended (current), and modifiable and non-modifiable risk/protective factors (current). Participants
shared COVID-19 specific information, which
provided insight about awareness of local mitigation efforts and salience of the virus.
METHODS
Participant Recruitment and Procedure
We used Prolific Academic, an established
method of online participant recruitment, that
cross-stratifies on age, ethnicity, and sex, to obtain
a representative US sample using Census Bureau
population group estimates.62,63 Individuals were
eligible for this study if they lived in the US, spoke
English, and were at least 18 years old. Participants
were compensated $8 for completion of all 3 parts
Dependent Variables
Physical activity and sedentary behavior. The
International Physical Activity Questionnaire
(IPAQ-SF) is a valid and reliable 7-item measure of
vigorous and moderate physical activity, walking,
and time spent sitting during the past week.65,66
Participants were asked to identify the level of
physical activity (vigorous, moderate, walking, and
sitting) and length of time (days per week and minutes per day) in each type of activity over the past
7 days. Data were scored using the IPAQ scoring
Am J Health Behav.™ 2021;45(1):44-61
DOI: doi.org/10.5993/AJHB.45.1.4
47
Health Behaviors at the Onset of the COVID-19 Pandemic
guidelines and calculated using the average metabolic equivalents (MET) score for each type. Total
weekly physical activity in MET-minutes per week
was estimated by adding the MET value for each
type of physical activity. This overall score allowed
us to include all types of physical activity in an
overall measure, acknowledging that physical activity habits may have changed due to COVID-19.
Sedentary behavior captured the total time in minutes participants spent sitting on a weekday.
Diet quality. Intake of fruits/vegetables, and
added sugar was obtained using the dietary screener questionnaire (DSQ).67 The DSQ asks participants about the frequency of intake for selected
food items and beverages. Responses are then converted into cups of fruit and vegetables (excluding
French fries) and teaspoons of added sugar per day
using the National Cancer Institute validated scoring procedures. The DSQ has been validated across
large, diverse national samples.67
Social isolation. We used a single indicator of
social isolation: “I feel isolated from other people”
(0 = Not at all to 4 = Almost always). Single-item
measures can provide highly valuable information.69 This indicator is comparable to items commonly included in national surveys that capture
a global measure of perceived social isolation.69,70
Participants were not provided a standard definition of social isolation, which allowed for the item
to capture individuals’ subjective experience and
feelings about social isolation (eg, their perceived
adequacy of companionship and social support).71
Risk and Protective Factors
Non-modifiable factors. Participants answered
a battery of sociodemographic questions, including age, sex, race/ethnicity, employment status,
relationship status, number of children in household, and geographic residence. For analytic purposes, we created dichotomous dummy variables
for all variables except for the number of children
in the household, which was a continuous variable.
To capture differences between 3 developmental
stages72,73 with distinct features: we used emerging [ages 18-29], middle [ages 30-59], and late
[age 60+] adulthood). We recoded sex (male and
female; “other” was recoded as missing for 3 participants), race/ethnicity (White, Black, and other;
sample size for Asian and Hispanic/Latino were
48
small), employment (working full/part time and
other), and relationship status (married/partnered
vs other); for geographic residence, we created
dummy variables for each category (rural, smalltown/mid-size city, suburban, or urban).
Modifiable factors. Self-rated health was measured by asking: “In general, would you say your
health is:” with a 5-point response scale (0 = Poor
to 5 = Excellent). This single-item measure is commonly used as a reliable and valid measure of selfrated health.69
Resilience was measured using a 6-item scale
with a 5-point response scale (1 = strongly disagree
to 5 = strongly agree), with higher scores indicating
higher resilience.74 Cronbach’s alpha (α) was .918.
Depression symptoms were measured using the
Center for Epidemiological Studies Depression
Short Form (CES-D-10) with a 4-point response
scale (0 = rarely or none of the time to 3 = all of the
time).75 The total score was calculated as the sum of
the 10 items, where scores ≥ 10 indicates the presence of significant depressive symptoms (α = .866).
Participants completed the Multidimensional
Scale of Perceived Social Support,76 asking about
support from a significant other, family, and
friends. Responses (1 = strongly disagree to 5 =
strongly agree) to the 12-item scale were summed
into a total score (α = .940).
To assess subjective wellbeing, participants were
asked: “Taking all things together, how satisfied are
you with your life as a whole these days?” and responded on a 4-point response scale (1 = not at all
satisfied to 4 = very satisfied).77
Due to the abrupt disruption to household income experienced since COVID-19 we used food
security and personal financial wellbeing as proxyindicators for financial security. Food security
(binary) was based on a validated 2-item food insecurity screening tool 78 that captured participants’
responses of “sometimes” or “often” true to at least
one of the items. Using the 8-item Personal Financial Wellbeing, participants indicated their level of
financial distress; an average score was calculated
with higher scores indicating higher financial wellbeing 79 Reliability was high (α = .922).
Data Analysis
First, we conducted univariate, bivariate, and de-
Weaver et al
Table 1
Characteristics of the Representative
Sample (N = 362)
Non-modifiable Factors
M (SD)
N (%)
Age Categories
Table 1 (continued)
Characteristics of the Representative
Sample (N = 362)
Non-modifiable Factors
M (SD)
N (%)
Income/Year in 2019 a
Emerging Adulthood (ages 18-29)
23.9 (3.4)
71 (21.5)
< $35,000 108 (29.8) Middle Adulthood (ages 30-59) 45.4 (9.4) 189 (52.2) $35,000 – $51,999 80 (22.1) Late Adulthood (ages 60+) 65.7 (4.8) 94 (26.0) $52,000 – $73,999 66 (18.2) $74,000 – $99,999 52 (14.4) > $100,000
49 (13.5)
Sex
Male
173 (47.8)
Female
186 (51.4)
Race/Ethnicity
Modifiable Factors
M (SD)
Self-rated Health (0-3)
1.3 (0.9)
N (%)
White, non-Hispanic
249 (68.8)
Black
42 (11.9)
Poor/Fair
61 (16.9)
Asian
27 (7.5)
Good
151 (41.7)
Hispanic/Latino
20 (5.5)
Very Good
118 (32.6)
Other
22 (6.1)
Marital Status
Excellent
32 (8.8)
Resiliency (1-5)
3.4 (0.9)
Depressive Symptoms
11.1(6.3)
Single
125 (34.5)
Married/Partnered
185 (51.1)
Not at Risk of Depression b
150 (41.4)
Divorced/Separated
49 (13.6)
Depression Risk
210 (58.0)
Other
3 (0.8)
People in the Household (1-11)
2.6 (1.4)
Number of Children (0-5)
0.3 (0.8)
Children (binary)
No
296 (81.8)
Yes
66 (19.2)
Geographic Location
Rural
71 (19.6)
Mid-size city/town
36 (9.9)
Suburban
174 (48.1)
Urban
81 (22.4)
# Chronic Condition a
1.2 (1.4)
Education level a
41 (11.3)
Some college
97 (26.8)
Associate Degree
47 (13.5)
4-year Degree
122 (33.7)
Graduate Degree
53 (14.6)
Current Employment
Unemployed
79 (21.8)
Employed (Full/Part Time)
187 (51.7)
89 (24.6)
(continued on next column)
Am J Health Behav.™ 2021;45(1):44-61
Social Support (12-60)
45.0 (11.0)
Subjective Wellbeing, Prior to COVID-19 (1-4)
2.9 (0.8)
Not At All Satisfied
13 (3.6)
Not Very Satisfied
94 (26.0)
Satisfied
177 (48.9)
Very Satisfied
78 (21.5)
Subjective Wellbeing, after COVID-19 restrictions (1-4)
2.6 (0.8)
Not At All Satisfied
28 (7.2)
Not Very Satisfied
130 (35.9)
Satisfied
172 (47.5)
Very Satisfied
32 (8.8)
Food Security, Prior to COVID-19
≤ High School/GED
Not Seeking Work
b
Food Secure
228 (63.0)
Food Insecure
87 (24.0)
Food Security, after COVID-19
restrictions
Food Secure
221 (61.0)
Food Insecure
140 (38.7)
Personal Financial Wellbeing (1-10)
5.6 (2.1)
Note.
Total N (%) may not add up due to missing data
a
Included for descriptive purposes only
b
Based on CES-D Short cutoff
DOI: doi.org/10.5993/AJHB.45.1.4
49
Health Behaviors at the Onset of the COVID-19 Pandemic
Table 2
Evaluating Health Behavior Changes
Outcome
Pre COVID-19
M(SD)
Current
M(SD)
t-test
p value
Effect
size
Total METs (per week)
2205.0 (3342.7)
1616.3 (2176.6)
4.828
< .001 0.27 Sitting (minutes per day) 460.9 (281.6) 494.5 (211.5) -7.178 < .001 -0.46 Vegetables/Fruits (cups) 2.4 (0.7) 2.4 (0.7) 0.555 .580 0.03 Added Sugar (tsp) 16.4 (6.7) 16.0 (6.5) 1.258 .209 0.07 Social Isolation (0-4) 1.1 (1.1) 1.6 (1.3) -7.692 < .001 -0.40 scriptive statistics to characterize the study sample; t-tests were used to assess changes in food security status and subjective wellbeing after COVID-19 restrictions were implemented/recommended. Then, we ran t-tests to assess perceived change (RQ1) in physical activity, diet quality, and social isolation variables from pre-COVID-19 to current. Finally, data were analyzed using hierarchical linear regression (RQ2-4) to identify risk/protective factors for physical activity (total METs, sitting), diet quality (vegetables/fruits intake, added sugar intake), and social isolation. Each model contained 2 steps: (1) a step containing relevant non-modifiable, and (2) a step that added modifiable factors. To determine if each step improved the model fit, we used the F statistic assessing significance of R2 change. As this was an exploratory study, we included different factors in each model, based on theoretical reasons identified in the literature, statistically significant correlations (see Online Resource: Correlation Matrices), and factors likely to be relevant to circumstances related to COVID-19 (ie, age category, self-rated health, resilience, depressive symptoms, social support, and subjective wellbeing). We used pairwise deletion so all cases with available data were included in the model, regardless of missing values for other variables. Other than planned missingness on the diet quality measure, missingness was not an issue. Data were analyzed using SPSS version 26, with a p-value ≤ .05 to indicate statistical significance. RESULTS Sample Characteristics Table 1 shows descriptive characteristics of the study participants. On average, emerging adults 50 were 23.9 years old, middle-aged adults were 45.4 years old, and older adults were 65.7 years old. There were near equal number of men and women who completed the survey, with the majority identifying as non-Hispanic White. More than half of the sample was married/partnered and 19.2% had at least one child in the household. There was an average of 2.6 people per household. The majority of participants were employed full or part-time, though 20% were seeking employment. Table 1 provides additional descriptive statistics about non-modifiable and modifiable factors. As for modifiable factors, overall the sample self-rated their health as good, very good, or excellent, although one in 6 people reported poor/fair health. Resilience scores were slightly lower than average74 and personal financial wellbeing aligned with national average norms;79 however, more than half the sample exhibited significant depressive symptoms (ie, at risk of depression). More than one-third of participants indicated they were food insecure, which increased significantly after COVID-19 restrictions were implemented/recommended, t(315) = -4.07, p < .01. Although descriptive findings for current subjective wellbeing suggested more than half the sample was satisfied or very satisfied, subjective wellbeing decreased after COVID-19 restrictions were implemented/recommended, t(361) = 8. 698, p = .041. As for COVID-19 mitigation efforts, nearly all participants indicated at least one restriction was in place, with more than 90% indicating recommendations to avoid social gatherings, remain physically distant (ie, at least 6 feet apart) from persons outside the household, and leave the house only for essential reasons; most participants indicated that Weaver et al Table 3 Risk and Protective Factors Associated with Physical Activity Modifiable Factors Non-Modifiable Factors R2 Model 1: Total METs (N = 259) ∆ R2 F change R2 Model 2: Sitting (N = 298) ∆ R2 F change Step 1 a .059 .033 2.244* .036 .023 2.771* Step 2 .152 .100 3.360** .124 .091 4.116** Beta p value 95% CI -.011 .847 -60.82 − 48.60 -.165 .005 -176.96 – -32.62 Beta p value 95% CI Ages 18-29 b .121 .120 -168.19 − 1448.81 Ages 30-59 b .111 .152 -179.37 − 1144.72 Female b -.009 .889 -582.74 − 505.33 Black b -.156 .013 -1813.52 – -216.22 Working Full/Part Time b .031 .634 -419.00 − 686.53 # Children in Household .131 .033 29.84 − 719.04 -.038 .514 -41.99 − 21.08 Rural .057 .347 -391.59 − 963.31 -.060 .283 -90.80 − 26.61 Self-rated Health .124 .059 -11.40 − 639.42 -.116 .051 -57.33 – -0.11 Resilience .046 .543 -258.06 − 498.10 -.004 .950 -34.24 − 32.11 Depressive Symptoms -.115 .142 -91.80 − 13.21 .038 .592 -3.41 − 5.96 Social Support -.078 .244 -41.14 – 10.53 -.060 .328 -3.45 – 1.16 Subjective Wellbeing .011 .883 -391.76 − 455.24 -.090 .194 -62.72 – 12.78 Food Secure b .089 .158 -155.02 − 946.01 -.116 .045 -99.42 – -1.06 Total METs − − − -.154 .009 -0.03 − -0.00 Sitting -.149 .018 -2.80 − -0.26 − − − Fruits/Vegetables Intake .115 .073 -35.51− 801.15 Ages 60+ b b *p ≤ .05 **p ≤ .001 Note. Factors were included based on theoretical and statistical reasoning and differ among models. a Only non-modifiable factors were entered on Step 1. For parsimony, specific beta coefficients are presented only for Step 2. b This variable is a dichotomous dummy variable compared to the reference group (eg, “Ages 18-29” compared to “all other ages” and “Female” compared to “male”). restaurants were serving take-out only. Less than half of participants (42.0%) adopted restrictions when they were announced, yet more than onethird (35.1%) had adopted the recommendations prior to implementation. At the time of data collection, 63.5% of participants had been following recommendations/restrictions for 5+ weeks. in health behaviors from pre-COVID to current behaviors. Specifically, total METs decreased (small effect size). Sitting and isolation increased (moderate effect size). The 2 diet quality variables did not change (Table 2). Perceived Health Behavior Change There were several significant perceived changes Risk/Protective Factors Associated with Health Behaviors For each model, adding Step 2 significantly increased the percentage of variance accounted for Am J Health Behav.™ 2021;45(1):44-61 DOI: doi.org/10.5993/AJHB.45.1.4 51 Health Behaviors at the Onset of the COVID-19 Pandemic Table 4 Risk and Protective Factors Associated with Diet Quality Modifiable Factors Non-Modifiable Factors Model 3: Fruits/Vegetables Intake (N = 306) Model 4: Added Sugar Intake (N = 306) R2 ∆ R2 F change R2 ∆ R2 F change Step 1a .077 .068 8.410** .084 .072 6.924** Step 2 .165 .134 3.892** .136 .104 2.533* Beta p value 95% CI Beta p value 95% CI -.143 .014 -0.41 – -0.05 .046 .453 -1.18 – 2.63 -.075 .201 -2.82 – 0.59 Ages 18-29 b Ages 60+ b Female b -.239