1 Overview

This is a report comparing respondents who completed all three surveys and respondents who completed the first survey, but then did not complete either the second or third.

A total of 2085 respondents completed the first survey.

Completed sample: 930 respondents who completed all three surveys.
Dropout sample: 1155 respondents who completed the first survey, but then did not complete either the second or third survey.

Compared to completers, dropouts were, on average: younger, reported lower household income, reported having more comorbidities, lower health literacy, lower numeracy, less interested in getting a COVID-19 vaccine, less trusting of healthcare, and had less belief in science.

Overall, dropouts were more likely to be non-Veterans, to live in a rural area, and to have had COVID-19. In addition, they were less likely to be non-Hispanic White, to be from the west (vs. the northeast, midwest or south), and to say they have a health condition that they believe makes them more vulnerable to COVID-19.

The table below shows an overview of these measures and results (specifics found further below):

2 Characteristics

2.1 Age

Q. What is your age?

Response options:

18 to 24
25 to 34
35 to 44
45 to 54
55 to 64
65 to 74
75 to 84
85 or older

Summary statistics for the original age options for the complete (all three surveys) sample and dropout (first survey only) sample are shown below. Initial indication that dropout sample were younger (Median of 6 corresponds to 65-74, 5 corresponds to 55-64)

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Age	1	Dropout	1	1150	4.522609	1.926615	5	4.572826	2.9652	1	8	7	-0.2375713	-1.201741	0.0568128
	2	Completed	1	928	5.720905	1.283332	6	5.920699	0.0000	1	8	7	-1.4859480	2.430853	0.0421274

The table below shows the number and proportion of respondents reported age in the complete and dropout samples.

Age	Dropout		Complete
	5	0.4%	2	0.2%
18 to 24	66	5.7%	10	1.1%
25 to 34	168	14.5%	27	2.9%
35 to 44	181	15.7%	39	4.2%
45 to 54	112	9.7%	48	5.2%
55 to 64	120	10.4%	116	12.5%
65 to 74	338	29.3%	475	51.1%
75 to 84	140	12.1%	198	21.3%
85 or older	25	2.2%	15	1.6%
Total	1155	100.0%	930	100.0%

Mann-Whitney U Test testing for differences between distributions. Significant difference found: dropouts were generally younger

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  age1_num by Complete_Fct
## W = 345912, p-value < 2.2e-16
## alternative hypothesis: true location shift is not equal to 0

effsize	magnitude
0.3136811	moderate

2.2 Veteran status

Q. Are you a U.S. military Veteran?

Response options: No [non-Veteran] (0), Yes [Veteran] (1)

Table below shows the number and proportion of Veterans and non-Veterans in the complete and dropout samples.

Veteran status	Dropout		Complete
Non-Veteran	679	58.8%	346	37.2%
Veteran	476	41.2%	584	62.8%
Total	1155	100.0%	930	100.0%

Looks like the majority of respondents who dropped out were non-Veterans. Supported by χ2 test.

## 
##      Chi-square test of categorical association
## 
## Variables:   Veteran_Fct, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##              Complete_Fct
## Veteran_Fct   Dropout Completed
##   Non-Veteran     679       346
##   Veteran         476       584
## 
## Expected contingency table under the null hypothesis:
##              Complete_Fct
## Veteran_Fct   Dropout Completed
##   Non-Veteran     568       457
##   Veteran         587       473
## 
## Test results: 
##    X-squared statistic:  95.164 
##    degrees of freedom:  1 
##    p-value:  <.001 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.214 
##    Yates' continuity correction has been applied

2.3 Income

Q. What is your household income?

Response options:

Less than $20,000
$20,000 – $29,999
$30,000 - $39,999
$40,000 - $49,999
$50,000 - $59,999
$60,000 - $74,999
$75,000 - $99,999
$100,000 - $149,999
$150,000+
Prefer to not say

Summary statistics for the original income options for the complete (all three surveys) sample and dropout (first survey only) sample are shown below. Initial indication that dropout sample reported lower income (Median of 7 corresponds to USD75,000 - 99,999. 6 corresponds to 60,000 - 74,999)

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Income	1	Dropout	1	1155	5.524675	2.811533	6	5.580540	2.9652	1	10	9	-0.1956106	-1.1827908	0.0827279
	2	Completed	1	930	6.378495	2.345958	7	6.590054	1.4826	1	10	9	-0.6792472	-0.3892811	0.0769270

Age	Dropout		Complete
Less than $20,000	146	12.6%	43	4.6%
$20,000 – $29,999	89	7.7%	38	4.1%
$30,000 - $39,999	87	7.5%	55	5.9%
$40,000 - $49,999	111	9.6%	70	7.5%
$50,000 - $59,999	92	8.0%	73	7.8%
$60,000 - $74,999	130	11.3%	121	13.0%
$75,000 - $99,999	146	12.6%	168	18.1%
$100,000 - $149,999	165	14.3%	194	20.9%
$150,000+	120	10.4%	131	14.1%
Prefer to not say	69	6.0%	37	4.0%
Total	1155	100.0%	930	100.0%

Mann-Whitney U Test testing for differences between distributions. Significant difference found: dropouts generally reported lower income

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  income_Num by Complete_Fct
## W = 383212, p-value = 5.171e-16
## alternative hypothesis: true location shift is not equal to 0

effsize	magnitude
0.1822476	small

2.4 Rural/Urban

Q. How would you best describe the place where you live?

Response options:

Rural
Small city, e.g. less than 100,000 people
Suburban, near a large city
Mid-sized city, 100,000 to 1 million people
Large city, more than 1 million
Other, please specify

Residence	Dropout		Complete
Rural	241	20.9%	151	16.2%
Small (less than 100,000)	200	17.3%	159	17.1%
Suburban near large city	438	37.9%	457	49.1%
Mid sized city (100,000 to 1million)	106	9.2%	90	9.7%
large city more than 1million	163	14.1%	70	7.5%
Other	4	0.3%	3	0.3%
Did not respond	3	0.3%	0	0.0%
Total	1155	100.0%	930	100.0%

Just looking at rural vs urban

Residence	Dropout		Complete
Rural	241	20.9%	151	16.2%
Urban	907	78.5%	776	83.4%
NA	7	0.6%	3	0.3%
Total	1155	100.0%	930	100.0%

Looks like the majority of respondents who dropped out reported being from rural areas. Supported by χ2 test.

## 
##      Chi-square test of categorical association
## 
## Variables:   Urban_Fct, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##          Complete_Fct
## Urban_Fct Dropout Completed
##     Rural     241       151
##     Urban     907       776
## 
## Expected contingency table under the null hypothesis:
##          Complete_Fct
## Urban_Fct Dropout Completed
##     Rural     217       175
##     Urban     931       752
## 
## Test results: 
##    X-squared statistic:  7.102 
##    degrees of freedom:  1 
##    p-value:  0.008 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.059 
##    Yates' continuity correction has been applied

2.5 State regions

Looking at regions based on states.

State	Dropout		Complete
Midwest	243	21.0%	188	20.2%
Northeast	230	19.9%	170	18.3%
South	466	40.3%	354	38.1%
West	202	17.5%	212	22.8%
NA	14	1.2%	6	0.6%
Total	1155	100.0%	930	100.0%

Although it doesn’t look like dropouts differed by region dramatically, χ2 test indicates that respondents who reported being from the west were more likely to complete all three surveys.

## 
##      Chi-square test of categorical association
## 
## Variables:   stateFCT, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##            Complete_Fct
## stateFCT    Dropout Completed
##   West          202       212
##   Northeast     230       170
##   Midwest       243       188
##   South         466       354
## 
## Expected contingency table under the null hypothesis:
##            Complete_Fct
## stateFCT    Dropout Completed
##   West          229       185
##   Northeast     221       179
##   Midwest       238       193
##   South         453       367
## 
## Test results: 
##    X-squared statistic:  8.852 
##    degrees of freedom:  3 
##    p-value:  0.031 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.065

2.6 CCI

Q. As far as you know, do you have any of the following health conditions at the present time?

Asthma, emphysema, or chronic bronchitis, COPD (other lung disease)
Arthritis or rheumatism
Cancer, diagnosed in the past 3 years
Diabetes
Digestive problems (such as ulcer, colitis, or gallbladder disease)
Heart trouble (such as angina, congestive heart failure, or coronary artery disease, having a past heart attack)
HIV illness or AIDS
Kidney disease
Liver problems (such as cirrhosis)
Stroke
High blood pressure (hypertension)
Very overweight or obese

Response options: No, do not have this condition (0), Yes, I have this condition (1).

Summary statistics for the total number of comorbidities for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Total comorbidities	1	Dropout	1	1155	1.614719	1.919423	1	1.277838	1.4826	0	12	12	1.8834414	4.9503583	0.0564781
	2	Completed	1	930	1.474193	1.374800	1	1.310484	1.4826	0	7	7	0.8309065	0.1605565	0.0450815

Welch’s independent samples t-test testing for differences between group means. No significant difference found, but the respondents with most comorbidities all dropped out.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   CCI_ttl 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       1.615     1.474
##    std dev.   1.919     1.375
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  1.945 
##    degrees of freedom:  2056.141 
##    p-value:  0.052 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.001, 0.282] 
##    estimated effect size (Cohen's d):  0.084

2.7 Vulnerable

Q1. Do you have a preexisting health condition—respiratory illness, cancer, heart disease, high blood pressure, etc.—that you believe makes you more vulnerable to coronavirus?

Response options: No (0), Yes (1), Unsure (2).

The table below shows the number and proportion of respondents reported vulnerability to COVID in the complete and dropout samples.

Vulnerable	Dropout		Complete
No	642	55.6%	423	45.5%
Yes	413	35.8%	440	47.3%
Unsure	99	8.6%	67	7.2%
Did not respond	1	0.1%	0	0.0%
Total	1155	100.0%	930	100.0%

Looks like the majority of respondents who dropped out reported having vulnerability to COVID. Supported by χ2 test.

## 
##      Chi-square test of categorical association
## 
## Variables:   cci2_FCT1, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##          Complete_Fct
## cci2_FCT1 Dropout Completed
##    No         642       423
##    Yes        413       440
##    Unsure      99        67
## 
## Expected contingency table under the null hypothesis:
##          Complete_Fct
## cci2_FCT1 Dropout Completed
##    No       589.7     475.3
##    Yes      472.3     380.7
##    Unsure    91.9      74.1
## 
## Test results: 
##    X-squared statistic:  28.307 
##    degrees of freedom:  2 
##    p-value:  <.001 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.117

2.8 Race/Ethnicity

Q1. Are you Hispanic or Latino/a or Latinx?

Response options: No [non-Hispanic] (0), Yes [Hispanic] (1).

	Overall		Veteran
Non-hispanic	1060	91.8%	837	90.0%
Hispanic	93	8.1%	92	9.9%
No response	2	0.2%	1	0.1%
Total	1155	100.0%	930	100.0%

No difference in completion found following χ2 test.

## 
##      Chi-square test of categorical association
## 
## Variables:   Latinx_FCT1, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##               Complete_Fct
## Latinx_FCT1    Dropout Completed
##   Non-hispanic    1060       837
##   Hispanic          93        92
## 
## Expected contingency table under the null hypothesis:
##               Complete_Fct
## Latinx_FCT1    Dropout Completed
##   Non-hispanic    1051     846.5
##   Hispanic         102      82.5
## 
## Test results: 
##    X-squared statistic:  1.924 
##    degrees of freedom:  1 
##    p-value:  0.165 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.03 
##    Yates' continuity correction has been applied

Q2. What is your race? Mark all that apply.

Response options:

American Indian or Alaskan Native
Asian or Asian American
Black or African American
Native Hawaiian or other Pacific Islander
White or European American
Other (please specify)

The table below shows the number and percent of respondents answers to the race and ethnicity questions together. From the table it looks like white (both Hispanic and non-Hispanic) respondents were more likely to complete all three surveys. Black respondents (both Hispanic and non-Hispanic) appeared less likely to complete all surveys.

Hispanic	Race	Dropout: n (%)	Completed: n (%)
Non-hispanic	American Indian or Alaskan Native	3 (0.3%)	4 (0.5%)
Non-hispanic	Asian or Asian American	26 (2.5%)	26 (3.1%)
Non-hispanic	Black or African American	173 (16.3%)	64 (7.6%)
Non-hispanic	Multiple	8 (0.8%)	8 (1.0%)
Non-hispanic	Native Hawaiian or other Pacific Islander	1 (0.1%)	2 (0.2%)
Non-hispanic	No response	2 (0.2%)	NA
Non-hispanic	Other	11 (1.0%)	13 (1.6%)
Non-hispanic	White or European American	836 (78.9%)	720 (86.0%)
Hispanic	American Indian or Alaskan Native	3 (3.2%)	NA
Hispanic	Asian or Asian American	1 (1.1%)	1 (1.1%)
Hispanic	Black or African American	15 (16.1%)	5 (5.4%)
Hispanic	Multiple	3 (3.2%)	2 (2.2%)
Hispanic	Other	7 (7.5%)	4 (4.3%)
Hispanic	White or European American	64 (68.8%)	80 (87.0%)
No response	Black or African American	1 (50.0%)	NA
No response	White or European American	1 (50.0%)	1 (100.0%)

The below table shows respondents Race/Ethnicity as coded for analysis between samples.

Non-Hispanic White	Dropout		Complete
No	319	27.6%	210	22.6%
Yes	836	72.4%	720	77.4%
Total	1155	100.0%	930	100.0%

χ2 test, significant difference: Non-hispanic white respondents were more likely to complete all three surveys

Looks like the majority of respondents who completed all three surveys were Non-hispanic white. Supported by χ2 test.

## 
##      Chi-square test of categorical association
## 
## Variables:   NonHispanicWhite_FCT, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##                     Complete_Fct
## NonHispanicWhite_FCT Dropout Completed
##                  No      319       210
##                  Yes     836       720
## 
## Expected contingency table under the null hypothesis:
##                     Complete_Fct
## NonHispanicWhite_FCT Dropout Completed
##                  No      293       236
##                  Yes     862       694
## 
## Test results: 
##    X-squared statistic:  6.643 
##    degrees of freedom:  1 
##    p-value:  0.01 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.056 
##    Yates' continuity correction has been applied

2.9 Had CV19

Q. Have you been infected with coronavirus?

Response options:

Yes, I currently have coronavirus
Yes, I had coronavirus and I recovered
No, I haven’t had coronavirus

The table below shows the number and proportion of respondents who reported having (or not) been infected with COVID-19 in the complete and dropout samples.

	Dropout		Complete
haven’t had COVID	1071	92.7%	899	96.7%
yes, currently have COVID	29	2.5%	4	0.4%
yes, had COVID and recovered	55	4.8%	27	2.9%
Total	1155	100.0%	930	100.0%

Looks like the majority of respondents who dropped out reported having had COVID. Supported by χ2 test.

## 
##      Chi-square test of categorical association
## 
## Variables:   CV_StatusFCT, Complete_Fct 
## 
## Hypotheses: 
##    null:        variables are independent of one another
##    alternative: some contingency exists between variables
## 
## Observed contingency table:
##                               Complete_Fct
## CV_StatusFCT                   Dropout Completed
##   haven't had COVID               1071       899
##   yes, currently have COVID         29         4
##   yes, had COVID and recovered      55        27
## 
## Expected contingency table under the null hypothesis:
##                               Complete_Fct
## CV_StatusFCT                   Dropout Completed
##   haven't had COVID             1091.3     878.7
##   yes, currently have COVID       18.3      14.7
##   yes, had COVID and recovered    45.4      36.6
## 
## Test results: 
##    X-squared statistic:  19.464 
##    degrees of freedom:  2 
##    p-value:  <.001 
## 
## Other information: 
##    estimated effect size (Cramer's v):  0.097

2.10 Health literacy

Q. How often do you have someone (like a family member, friend, hospital/clinic worker or caregiver) help you read instructions, pamphlets or other written health materials from your doctor or pharmacy?

Response options: Never(1), Rarely(2), Sometimes(3), Often(4), Always(5).

Summary statistics for health literacy for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Health literacy	1	Dropout	1	1155	1.619913	1.0880404	1	1.372973	0	1	5	4	1.742399	2.07433	0.0320150
	2	Completed	1	930	1.238710	0.6811363	1	1.052419	0	1	5	4	3.357355	11.88410	0.0223353

Welch’s independent samples t-test testing for differences between group means. Significant difference found: Respondents who dropped out generally reported lower health literacy.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   pharmacyInstructions 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       1.620     1.239
##    std dev.   1.088     0.681
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  9.765 
##    degrees of freedom:  1970.781 
##    p-value:  <.001 
## 
## Other information: 
##    two-sided 95% confidence interval:  [0.305, 0.458] 
##    estimated effect size (Cohen's d):  0.42

2.11 Numeracy

Q1. How good are you at working with fractions?
Q2. How good are you at figuring out how much a shirt will cost if it is 25% off?

Response options: Not at all good(1), — (2), — (3), — (4), — (5) Extremely good (6).

Q3. How often do you find numerical information to be useful?

Response options: Never(1), — (2), — (3), — (4), — (5) Very often (6).

Summary statistics for numeracy for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Numeracy	1	Dropout	1	1155	4.537951	1.357230	5.000000	4.705586	1.4826	1	6	5	-0.8131532	-0.1549362	0.0399358
	2	Completed	1	930	4.978853	1.133651	5.333333	5.160394	0.9884	1	6	5	-1.2216919	1.0349279	0.0371739

Welch’s independent samples t-test testing for differences between group means. Significant difference found: dropouts generally reported lower numeracy.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   Numeracy_Avg 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       4.538     4.979
##    std dev.   1.357     1.134
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  -8.081 
##    degrees of freedom:  2080.2 
##    p-value:  <.001 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.548, -0.334] 
##    estimated effect size (Cohen's d):  0.353

2.12 COVID-19 worry

Q1. How worried are you about getting COVID-19?

Response options: Not at all worried (1), (2), (3), (4), Very worried (5).

Summary statistics for worry about COVID for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
COVID worry	1	Dropout	1	1155	3.013853	1.392533	3	3.017297	1.4826	1	5	4	0.0253291	-1.231917	0.0409746
	2	Completed	1	930	2.918280	1.287427	3	2.897849	1.4826	1	5	4	0.1945095	-1.045154	0.0422164

Welch’s independent samples t-test testing for differences between group means. No significant difference found on worry about COVID-19.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   worried.self 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       3.014     2.918
##    std dev.   1.393     1.287
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  1.625 
##    degrees of freedom:  2043.697 
##    p-value:  0.104 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.02, 0.211] 
##    estimated effect size (Cohen's d):  0.071

2.13 CV19 risk perception

Q1. In your opinion, how likely is it that you will get COVID-19 during the next month?
Q2. If you were to get COVID-19, how likely do you think it is that you would need to be hospitalized?
Q3. If you were to get COVID-19, how likely do you think it is that you would die?

Response options: Not at all likely(1), — (2), — (3), — (4), Very likely (5).

Summary statistics for COVID-19 risk perceptions for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
COVID risk perceptions	1	Dropout	1	1155	2.522944	1.0794225	2.333333	2.464144	0.9884	1	5	4	0.3087535	-0.6729507	0.0317615
	2	Completed	1	929	2.460172	0.9312925	2.333333	2.428859	0.9884	1	5	4	0.3064475	-0.5219726	0.0305547

Welch’s independent samples t-test testing for differences between group means. No significant difference found on COVID-19 risk perceptions.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   CV19Risk_Avg 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       2.523     2.460
##    std dev.   1.079     0.931
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  1.424 
##    degrees of freedom:  2071.776 
##    p-value:  0.155 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.024, 0.149] 
##    estimated effect size (Cohen's d):  0.062

2.14 CV19 Vaccine intentions

Q1. A coronavirus vaccine will soon become available. How interested are you in getting the vaccine?

Response options: I definitely do NOT want to get the vaccine (1), I do NOT want to get the vaccine (2), Uncertain (3), I WANT to get the vaccine (4), I definitely WANT to get the vaccine (5).

Summary statistics for COVID vaccine interest for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Total comorbidities	1	Dropout	1	1155	3.509091	1.407659	4	3.635676	1.4826	1	5	4	-0.5148776	-0.9646097	0.0414196
	2	Completed	1	930	4.122581	1.200119	5	4.358871	0.0000	1	5	4	-1.2552021	0.5955115	0.0393535

Welch’s independent samples t-test testing for differences between group means. Significant difference found: dropouts were generally less interested in getting vaccine.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   vax1 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       3.509     4.123
##    std dev.   1.408     1.200
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  -10.738 
##    degrees of freedom:  2076.201 
##    p-value:  <.001 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.726, -0.501] 
##    estimated effect size (Cohen's d):  0.469

2.15 Healthcare trust

Please indicate how much you agree or disagree with each statement. There are no right or wrong answers. Please answer in a way that reflects your own personal beliefs:

The Health Care System does its best to make patients health better
The Health Care System covers up its mistakes
Patients receive high quality medical care from the Health Care System
The Health Care System makes too many mistakes
The Health Care System puts making money above patients’ needs
The Health Care System gives excellent medical care
Patients get the same medical treatment from the Health Care System no matter what the patient’s race or ethnicity
The Health Care System lies to make money
The Health Care System experiments on patients without them knowing

Response options: Strongly disagree (1), Disagree (2), Somewhat disagree (3), Neither agree nor disagree (4), Somewhat agree (5), Agree (6), Strongly agree (7).

Summary statistics for healthcare trust for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Healthcare trust	1	Dropout	1	1155	3.985618	1.059841	3.666667	3.933634	0.8236667	0.4444444	6.444444	6.000000	0.3115664	0.2161716	0.0311853
	2	Completed	1	930	4.329570	1.071551	4.222222	4.315113	1.1531333	0.7777778	6.444444	5.666667	0.0073459	-0.3368769	0.0351376

Welch’s independent samples t-test testing for differences between group means. Significant difference found: dropouts generally reported less trust in healthcare.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   HealthcareTrust_Avg 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       3.986     4.330
##    std dev.   1.060     1.072
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  -7.321 
##    degrees of freedom:  1979.971 
##    p-value:  <.001 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.436, -0.252] 
##    estimated effect size (Cohen's d):  0.323

2.16 Belief in science

Please indicate how much you agree or disagree with each statement. There are no right or wrong answers. Please answer in a way that reflects your own personal beliefs:

People trust scientists a lot more than they should
People don’t realize just how flawed a lot of scientific research really is
A lot of scientific theories are dead wrong
Sometimes I think we put too much faith in science
Our society places too much emphasis on science
I am concerned by the amount of influence that scientists have in society

Response options: Strongly disagree (1), Disagree (2), Somewhat disagree (3), Neither agree nor disagree (4), Somewhat agree (5), Agree (6), Strongly agree (7).

Summary statistics for lack of belief in science for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
(lack of) belief in science	1	Dropout	1	1154	3.742634	1.549062	4.000000	3.730339	1.4826	1	7	6	-0.0026563	-0.618267	0.0456001
	2	Completed	1	930	3.265054	1.562983	3.333333	3.193100	1.7297	1	7	6	0.2643233	-0.768540	0.0512522

Welch’s independent samples t-test testing for differences between group means. Significant difference found: dropouts generally reported less belief in science.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   BeliefinScience_Avg 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       3.743     3.265
##    std dev.   1.549     1.563
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  6.962 
##    degrees of freedom:  1981.494 
##    p-value:  <.001 
## 
## Other information: 
##    two-sided 95% confidence interval:  [0.343, 0.612] 
##    estimated effect size (Cohen's d):  0.307

2.17 Maximizing

Q1. Sometimes, medical action is clearly necessary, and sometimes it is clearly NOT necessary. Other times, people differ in their beliefs about whether medical action is needed. In medical situations where it’s not clear, do you tend to lean towards taking action or do you lean towards waiting and seeing if action is needed? Importantly, there is no “right” way to be. Please answer on the 1-6 scale below:?

Response options: I lean toward waiting and seeing (1), — (2), — (3), — (4), — (5), I lean toward taking action (6).

Summary statistics for maximizing for the complete and dropout samples are shown below.

	item	group1	vars	n	mean	sd	median	trimmed	mad	min	max	range	skew	kurtosis	se
Maximizing	1	Dropout	1	1155	3.851948	1.553685	4	3.939459	1.4826	1	6	5	-0.3250303	-0.8511671	0.0457164
	2	Completed	1	930	3.894624	1.457571	4	3.954301	1.4826	1	6	5	-0.2836856	-0.8111278	0.0477956

Welch t test for differences. No significant difference found on maximizing.

## 
##    Welch's independent samples t-test 
## 
## Outcome variable:   medicalAction 
## Grouping variable:  Complete_Fct 
## 
## Descriptive statistics: 
##             Dropout Completed
##    mean       3.852     3.895
##    std dev.   1.554     1.458
## 
## Hypotheses: 
##    null:        population means equal for both groups
##    alternative: different population means in each group
## 
## Test results: 
##    t-statistic:  -0.645 
##    degrees of freedom:  2035.133 
##    p-value:  0.519 
## 
## Other information: 
##    two-sided 95% confidence interval:  [-0.172, 0.087] 
##    estimated effect size (Cohen's d):  0.028

3 Regression

I entered all the above characteristics into a multiple logistic regression model predicting sample (0=dropout, 1=complete). It looks like older age, higher income, fewer comorbidities, vulnerability to COVID, higher health literacy, and greater intentions to get a COVID vaccine were significant predictors of completing all three surveys.

	Predictors of completing all three surveys
Predictors	Odds Ratios	p value
(Intercept)	0.10 (0.04 – 0.23)	<0.001
Age	1.36 (1.25 – 1.49)	<0.001
Veteran status	1.02 (0.80 – 1.31)	0.871
Income	1.08 (1.03 – 1.13)	0.002
Residence	1.05 (0.80 – 1.38)	0.725
Region: Northeast	0.80 (0.58 – 1.11)	0.185
Region: Midwest	0.92 (0.67 – 1.26)	0.608
Region: South	0.97 (0.73 – 1.28)	0.817
Comorbidities	0.85 (0.78 – 0.93)	<0.001
Vulnerable to CoV	1.47 (1.13 – 1.93)	0.005
Race/Ethnicity	1.02 (0.80 – 1.30)	0.875
Had CoV: recovered	0.75 (0.20 – 2.32)	0.640
Had CoV: current	0.99 (0.55 – 1.73)	0.970
Health literacy	0.78 (0.67 – 0.89)	0.001
Numeracy	1.04 (0.95 – 1.15)	0.392
CoV worry	0.99 (0.90 – 1.09)	0.851
CoV risk	0.98 (0.85 – 1.12)	0.735
CoV vaccine intentions	1.16 (1.05 – 1.27)	0.003
Healthcare trust	0.99 (0.89 – 1.11)	0.886
(lack of) Belief in science	0.96 (0.89 – 1.03)	0.274
Maximizing	1.01 (0.94 – 1.09)	0.783
Observations	1798
R² Tjur	0.159

COVID-19 VA Project: Comparing dropouts (did not do all 3 surveys) with completed (did all three surveys)

A Thorpe

17/05/2021

1 Overview

2 Characteristics

2.1 Age

2.2 Veteran status

2.3 Income

2.4 Rural/Urban

2.5 State regions

2.6 CCI

2.7 Vulnerable

2.8 Race/Ethnicity

2.9 Had CV19

2.10 Health literacy

2.11 Numeracy

2.12 COVID-19 worry

2.13 CV19 risk perception

2.14 CV19 Vaccine intentions

2.15 Healthcare trust

2.16 Belief in science

2.17 Maximizing

3 Regression