1 Data Quality Checks

We only included participants with a gc score of 1. Those with a gc value equal to 1 designate the Good Completes on the filter supplied by Qualtrics.

For the VA version that is currently 688 participants.

gc n
1 688

For the non-VA version that is currently 388 participants.

gc n
1 388

We will filter out all except the good completes so our final sample is 688 for VA version and 388 for Non-VA version (code shown below).

data <-  data %>%filter(gc == "1")
dataNVA <-  dataNVA %>%filter(gc == "1")

1.1 Timing

Median completion timing for the VA version is 13 minutes and 21 seconds (round to 13 minutes).

## [1] "13M 21S"

Median completion timing for the non-VA version is 11 minutes and 10 seconds (round to 11 minutes).

## [1] "11M 10.5S"

Mean completion timing for VA version is 17 minutes and 5 seconds (round to 17 minutes).

## [1] "17M 3.57000000000005S"

Mean completion timing for non-VA version is 14 minutes and 55 seconds (round to 15 minutes).

## [1] "14M 54.4072S"

Names of all variables included.

Checking the number of respondents vaccinated in each group. 0 = no doses 1 = received one dose 2 = received two doses

## 
##   0   1   2 
## 362 243 471

We will remove all participants who have had one dose or more.

## 
##   0 
## 362

Checking how the number of participants was distributed across experimental groups

## 
## controlcondition   Vaccinefactbox  Vaccinetimeline 
##              120              123              119

2 Primary analyses

2.1 Effect of interventions on vaccine interest

DV = Vaccine interest:

“When a coronavirus vaccine becomes available to you, how interested are you in getting the vaccine?”

  • 1 = I definitely do not want the vaccine; 5 = I definitely want the vaccine

Overall descriptives for interest in getting vaccinated…

##    vars   n mean  sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 362 3.68 1.5      4    3.85 1.48   1   5     4 -0.69       -1 0.08

Checking for assumptions before running ANOVA

Plots do not look too extreme and leven’s test looks good too

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 359 3.587 271.5 2.371 0.09483

ANOVA test of H1. No effect of group assignment on vaccine intentions.

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • ANOVA:

      Effect DFn DFd F p p<.05 ges
    2 XIgroup 2 359 0.9131 0.4022 0.005061

Get results with additional effect sizes and power too

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 4.102 2.051 0.913 0.402 0.005 0.005 0 0 0 0.071 0.209
…2 Residuals 359 806.4 2.246

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 3.833 0.1368 359 3.564 4.102
Vaccinefactbox 3.602 0.1351 359 3.336 3.867
Vaccinetimeline 3.613 0.1374 359 3.343 3.884

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars   n mean  sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 120 3.83 1.4      4    4.04 1.48   1   5     4 -0.89    -0.56 0.13
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 123  3.6 1.56      4    3.75 1.48   1   5     4 -0.6    -1.19 0.14
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 119 3.61 1.53      4    3.75 1.48   1   5     4 -0.58     -1.2 0.14

Plotting means for visual inspection

2.2 Effect of interventions on time to get vaccine

DV = Time to vaccine (“When a coronavirus vaccine becomes available to you, how interested are you in gettingthe vaccine?”)

  • Immediately (1)
  • Less than one month (2)
  • One month to less than 3 months (3)
  • 3 months to less than six months (4)
  • 6 months to less than 1 year (5)
  • 1 year to less than 2 two years (6)
  • I would wait 2 years or more (7)
  • I would never get it (8)
  • Not sure (9)

Overall descriptives for time before getting vaccinated…

##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 310 3.15 2.68      2    2.82 1.48   1   8     7 0.88    -0.81 0.15

Checking for assumptions before running ANOVA. Plots do not look too extreme and levene’s test looks good too. 

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 307 6.188 1553 0.6117 0.5431

ANOVA test of H2. No effect of group assignment on time to get vaccine.

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • ANOVA:

      Effect DFn DFd F p p<.05 ges
    2 XIgroup 2 307 0.7616 0.4678 0.004937

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 10.97 5.486 0.762 0.468 0.005 0.005 -0.002 -0.002 -0.002 0.07 0.18
…2 Residuals 307 2212 7.204

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 2.902 0.2658 307 2.379 3.425
Vaccinefactbox 3.351 0.2548 307 2.85 3.853
Vaccinetimeline 3.196 0.2725 307 2.66 3.732

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars   n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 102  2.9 2.58      1    2.51   0   1   8     7 1.09    -0.34 0.26
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 111 3.35 2.76      2    3.07 1.48   1   8     7 0.72    -1.14 0.26
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean  sd median trimmed  mad min max range skew kurtosis   se
## X1    1 97  3.2 2.7      2     2.9 1.48   1   8     7 0.84    -0.88 0.27

Plotting means for visual inspection

3 Demographics

3.1 Age

Age of sample based on raw coding, grouped by survey labels, collapsed into fewer age groups, younger than 55 or 55+

age n percent
2 0.6%
1 12 3.3%
2 34 9.4%
3 40 11.1%
4 40 11.1%
5 82 22.7%
6 123 34.1%
7 27 7.5%
8 1 0.3%
age1 n percent
2 0.6%
18 to 24 12 3.3%
25 to 34 34 9.4%
35 to 44 40 11.1%
45 to 54 40 11.1%
55 to 64 82 22.7%
65 to 74 123 34.1%
75 to 84 27 7.5%
85 or older 1 0.3%
age2 n percent
2 0.6%
18 to 34 46 12.7%
35 to 54 80 22.2%
55 to 74 205 56.8%
75 or older 28 7.8%
age_FCT1 n percent
Younger than 55 128 35.5%
55 or older 233 64.5%

3.2 Gender

Gender of sample based on raw coding, grouped by survey labels

Gender of sample with different groupings

gender n percent
1 123 34.1%
2 235 65.1%
5 3 0.8%
Gender_CHR n percent
Female 123 34.1%
Male 235 65.1%
Non-binary/third gender 3 0.8%

3.3 Income

Overall descriptives for sample income and then frequencies

##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 361 5.57 2.63      6    5.63 2.97   1  10     9 -0.19    -1.06 0.14
income2a n percent
$0 - $49k 131 36.3%
$50K to $99K 125 34.6%
$100K and more 87 24.1%
Prefer to not say 18 5.0%

3.4 Race/Ethnicity

Race and ethnicity frequencies

LatinxCHR n percent
Hispanic 32 8.9%
No response 1 0.3%
Non-hispanic 328 90.9%
LatinxCHR RaceCHR count freq
Non-hispanic American Indian or Alaskan Native 2 2 (0.6%)
Non-hispanic Asian or Asian American 15 15 (4.2%)
Non-hispanic Black or African American 49 49 (13.6%)
Non-hispanic Multiple 4 4 (1.1%)
Non-hispanic Other 7 7 (1.9%)
Non-hispanic White or European American 251 251 (69.5%)
No response White or European American 1 1 (0.3%)
Hispanic American Indian or Alaskan Native 1 1 (0.3%)
Hispanic Black or African American 3 3 (0.8%)
Hispanic Other 3 3 (0.8%)
Hispanic White or European American 25 25 (6.9%)

3.5 Region

How participants best described the place where they live

ruralUrban1 n percent
Rural 77 21.3%
Small (less than 100,000) 58 16.1%
Suburban near large city 157 43.5%
Mid sized city (100,000 to 1million) 29 8.0%
large city more than 1million 40 11.1%

4 Secondary analyses

4.1 Moderations

4.1.1 By age

Moderation by age for vaccine interest. No effect of moderation.

Anova Table (Type III tests)
  Sum Sq Df F value Pr(>F)
(Intercept) 453.6 1 201.1 1.752e-36
XIgroup 0.2549 2 0.05651 0.9451
age_FCT1 2.69 1 1.193 0.2755
XIgroup:age_FCT1 2.637 2 0.5847 0.5578
Residuals 800.7 355 NA NA

Moderation by age for time before vaccine. No effect of moderation.

Anova Table (Type III tests)
  Sum Sq Df F value Pr(>F)
(Intercept) 326.7 1 44.96 0.00000000009835
XIgroup 0.7458 2 0.05132 0.95
age_FCT1 6.733 1 0.9267 0.3365
XIgroup:age_FCT1 7.856 2 0.5406 0.583
Residuals 2202 303 NA NA

4.1.2 By gender

Moderation by gender for vaccine interest. No effect of moderation.

Anova Table (Type III tests)
  Sum Sq Df F value Pr(>F)
(Intercept) 491.8 1 233.1 9.655e-41
XIgroup 6.172 2 1.463 0.233
Gender_Model 8.522 1 4.039 0.04521
XIgroup:Gender_Model 4.02 2 0.9526 0.3867
Residuals 742.7 352 NA NA

Moderation by gender for time before vaccine. No effect of moderation.

Anova Table (Type III tests)
  Sum Sq Df F value Pr(>F)
(Intercept) 418.3 1 61.22 0.00000000000008717
XIgroup 21.01 2 1.538 0.2166
Gender_Model 16.42 1 2.404 0.1221
XIgroup:Gender_Model 10.76 2 0.7875 0.4559
Residuals 2057 301 NA NA

5 Exploratory analyses

5.1 Why not vax

Looking at the reasons why respondents would not have gotten vaccinated (wave 1)

## The following `from` values were not present in `x`: 9
whynotvax1 n percent
I do not like needles 6 1.7%
I do not think I’ll get COVID-19 even if I don’t get the vaccine 7 1.9%
I do not think I’ll get very sick if I get COVID-19 14 3.9%
I do not think the vaccine will work very well 7 1.9%
I don’t trust big Pharma 9 2.5%
I have already had COVID-19 166 46.0%
I have concerns about the process that was used to develop the vaccine 15 4.2%
I will not have time to get the vaccine 10 2.8%
I’m concerned about side effects from the vaccine 5 1.4%
I’m worried I’ll get COVID-19 from the vaccine 64 17.7%
It is against my religious or philosophical beliefs 11 3.0%
no response 13 3.6%
Other (please specify) 27 7.5%
The COVID-19 outbreak is not as serious as some people say it is 7 1.9% 
##     
##      controlcondition Vaccinefactbox Vaccinetimeline
##   1                57             61              48
##   2                 3              1               1
##   3                17             18              29
##   4                 2              1               4
##   5                 3              6               5
##   6                 2              3               2
##   7                 2              4               1
##   8                 4              4               2
##   10                2              2               2
##   11                3              4               4
##   12                5              1               3
##   13                7              3               5
##   14                9             11               7
##   15                4              3               6

5.2 Side effect worry

Seeing how messages impacted worry about experiencing side effects from COVID-19 vaccine

Seeing how messages impacted views about how safe the COVID-19 vaccine is

Seeing how messages impacted belief whether COVID-19 vaccine is safer than getting COVID

Seeing how messages impacted belief whether COVID-19 vaccine was developed too fast

5.3 By veteran status

5.3.1 Vax interest

Moderation by veteran status for vaccine interest. No effect of moderation

Anova Table (Type III tests)
  Sum Sq Df F value Pr(>F)
(Intercept) 871.4 1 390.9 3.366e-59
XIgroup 4.186 2 0.9389 0.392
Group 4.186 1 1.878 0.1715
XIgroup:Group 9.602 2 2.154 0.1176
Residuals 791.3 355 NA NA

5.3.2 Vax time

Moderation by veteran status for time before vaccine. No effect of moderation.

Anova Table (Type III tests)
  Sum Sq Df F value Pr(>F)
(Intercept) 609.4 1 84.76 5.714e-18
XIgroup 11.44 2 0.7953 0.4524
Group 15.49 1 2.155 0.1432
XIgroup:Group 10.39 2 0.7224 0.4864
Residuals 2178 303 NA NA

5.4 By prior vax intent

5.4.1 Don’t want

First, we can see participants responses to the question about vaccine interest at wave 1.

  • 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)

and subset to look at the effect of messages compared to control among people who were hesitant at wave 1 (responses 1 and 2). This gives us an n of 86.

## 
##   1   2   3   4   5 
##  66  20 103  69 103
## 
##  1  2 
## 66 20

5.4.2 Vax interest (Dont want)

Overall descriptives for interest in getting vaccinated from those who were hesitant at wave 1.

##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 86 1.84 1.09      1    1.67   0   1   5     4 1.23     0.88 0.12

Checking for assumptions before running ANOVA

##     V1         
##  Mode:logical  
##  TRUE:86

Plots do not look too extreme and levene’s test looks good too

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 83 0.5518 56.47 0.4055 0.668

Get anova results

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • ANOVA:

      Effect DFn DFd F p p<.05 ges
    2 XIgroup 2 83 0.3243 0.7239 0.007755

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 0.789 0.394 0.324 0.724 0.008 0.008 -0.016 -0.016 -0.016 0.088 0.102
…2 Residuals 83 100.9 1.216

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 1.913 0.2299 83 1.456 2.37
Vaccinefactbox 1.692 0.2163 83 1.262 2.122
Vaccinetimeline 1.892 0.1813 83 1.531 2.252

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars  n mean sd median trimmed  mad min max range skew kurtosis   se
## X1    1 23 1.91  1      2    1.84 1.48   1   4     3 0.43    -1.39 0.21
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 26 1.69 1.09      1    1.55   0   1   5     4 1.32     0.89 0.21
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 37 1.89 1.17      2    1.68 1.48   1   5     4 1.41     1.26 0.19

Plotting means for visual inspection

5.4.3 Vax time (Dont want)

Overall descriptives for time before getting vaccinated…

##    vars  n mean   sd median trimmed mad min max range  skew kurtosis   se
## X1    1 64 7.02 1.91      8    7.46   0   1   8     7 -1.83     2.16 0.24

Checking for assumptions before running ANOVA. Plots do not look too extreme and levene’s test looks good too 

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 59 1.09 153.2 0.2099 0.8113 
## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • ANOVA:

      Effect DFn DFd F p p<.05 ges
    2 XIgroup 2 59 0.2099 0.8113 0.007064

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 1.09 0.545 0.21 0.811 0.007 0.007 -0.026 -0.026 -0.027 0.084 0.083
…2 Residuals 59 153.2 2.596

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 7.133 0.416 59 6.301 7.966
Vaccinefactbox 7.375 0.3289 59 6.717 8.033
Vaccinetimeline 7.087 0.336 59 6.415 7.759

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 15 7.13 1.81      8    7.38   0   3   8     5 -1.4     0.07 0.47
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed mad min max range  skew kurtosis   se
## X1    1 24 7.38 1.24      8    7.65   0   3   8     5 -2.13     4.09 0.25
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed mad min max range  skew kurtosis   se
## X1    1 23 7.09 1.81      8    7.47   0   2   8     6 -1.67     1.37 0.38

Plotting means for visual inspection

5.4.4 Uncertain

Now looking at the effect of messages compared to control among people who were uncertain at wave 1

Again we subset, this time taking only respondents who answered that they were uncertain (3) about getting a vaccine at wave 1. This leaves us with an n of 103.

## 
##   1   2   3   4   5 
##  66  20 103  69 103
## 
##   3 
## 103

5.4.5 Vax interest (Uncert)

Overall descriptives for interest in getting vaccinated from those who were uncertain at wave 1

##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 103  3.6 1.21      4    3.72 1.48   1   5     4 -0.5    -0.55 0.12

Checking for assumptions before running ANOVA

##     V1         
##  Mode:logical  
##  TRUE:103

Plots do not look too extreme and levene’s test looks good too

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 100 3.505 63.41 2.764 0.06785

Get anova results

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 1.226 0.613 0.416 0.661 0.008 0.008 -0.011 -0.011 -0.012 0.091 0.118
…2 Residuals 100 147.5 1.475

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 7.133 0.416 59 6.301 7.966
Vaccinefactbox 7.375 0.3289 59 6.717 8.033
Vaccinetimeline 7.087 0.336 59 6.415 7.759

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars  n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 35  3.6 1.09      4    3.69 1.48   1   5     4 -0.65    -0.18 0.18
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 39 3.49 1.43      3    3.58 2.97   1   5     4 -0.34    -1.22 0.23
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 29 3.76 1.02      4     3.8 1.48   1   5     4 -0.3    -0.37 0.19

Plotting means for visual inspection

5.4.6 Vax time (Uncert)

Overall descriptives for time before getting vaccinated…

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 78 3.19 2.19      3    2.95 2.97   1   8     7 0.65    -0.68 0.25

Checking for assumptions before running ANOVA. Plots do not look too extreme and levene’s test looks good too 

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 75 1.662 121.6 0.5123 0.6012

Get anova results

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • ANOVA:

      Effect DFn DFd F p p<.05 ges
    2 XIgroup 2 75 0.1839 0.8324 0.00488

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 1.796 0.898 0.184 0.832 0.005 0.005 -0.021 -0.021 -0.022 0.07 0.079
…2 Residuals 75 366.3 4.884

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 3.333 0.4253 75 2.486 4.181
Vaccinefactbox 3.233 0.4035 75 2.43 4.037
Vaccinetimeline 2.952 0.4823 75 1.992 3.913

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 27 3.33 2.17      3    3.13 2.97   1   8     7 0.61    -0.54 0.42
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 30 3.23 2.37      3    2.96 2.97   1   8     7 0.61    -1.02 0.43
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 21 2.95 2.01      3    2.71 2.97   1   7     6 0.59    -0.93 0.44

Plotting means for visual inspection

5.4.7 Pro-vaccine

Last, we can look at the effect of messages compared to control among people who wanted a vaccine at wave 1

Again we subset, this time taking only respondents who answered that they wanted (4) or defintely wanted (5) to get a vaccine at wave 1. This leaves us with an n of 172

## 
##   1   2   3   4   5 
##  66  20 103  69 103
## 
##   4   5 
##  69 103

5.4.8 Vax interest (Pro-vax)

Overall descriptives for interest in getting vaccinated from those who were provax at wave 1

##    vars   n mean   sd median trimmed mad min max range  skew kurtosis   se
## X1    1 172 4.66 0.76      5    4.86   0   1   5     4 -2.63     7.28 0.06

Checking for assumptions before running ANOVA

##      V1         
##  Mode :logical  
##  FALSE:5        
##  TRUE :167

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 126 0.9636 78.91 0.7693 0.4655

Get anova results

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 0.964 0.482 0.769 0.466 0.012 0.012 -0.004 -0.004 -0.004 0.111 0.182
…2 Residuals 126 78.91 0.626

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 4.659 0.1193 126 4.423 4.895
Vaccinefactbox 4.523 0.1193 126 4.287 4.759
Vaccinetimeline 4.732 0.1236 126 4.487 4.976

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars  n mean   sd median trimmed mad min max range  skew kurtosis   se
## X1    1 44 4.66 0.83      5    4.86   0   1   5     4 -2.85     8.22 0.13
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 44 4.52 0.88      5    4.69   0   1   5     4   -2     4.07 0.13
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed mad min max range  skew kurtosis  se
## X1    1 41 4.73 0.63      5    4.88   0   2   5     3 -2.62     7.09 0.1

Plotting means for visual inspection

5.4.9 Vax time (Pro-vax)

Overall descriptives for time before getting vaccinated…

##    vars   n mean   sd median trimmed mad min max range skew kurtosis  se
## X1    1 167 1.66 1.34      1    1.34   0   1   8     7 2.64     7.85 0.1

Checking for assumptions before running ANOVA 

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 160 2.571 154.1 1.335 0.2662

Get anova results

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 2.571 1.285 1.335 0.266 0.016 0.016 0.004 0.004 0.004 0.129 0.29
…2 Residuals 160 154.1 0.963

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 1.351 0.13 160 1.094 1.608
Vaccinefactbox 1.6 0.1323 160 1.339 1.861
Vaccinetimeline 1.627 0.1374 160 1.356 1.899

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars  n mean   sd median trimmed mad min max range skew kurtosis  se
## X1    1 57 1.35 0.77      1    1.17   0   1   4     3 2.13     3.57 0.1
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 55  1.6 1.05      1    1.38   0   1   5     4 1.78     2.41 0.14
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 51 1.63 1.11      1    1.39   0   1   5     4 1.52     0.93 0.16

Plotting means for visual inspection

5.4.10 As covariate

Checking the effect of interventions on vaccine interest by controlling for wave 1 vaccine interest. First creating variable of did not want vaccine (1), uncertain (2), and wanted vaccine (3).

## 
##   1   2   3 
##  86 103 172
## 
## Call:
## lm(formula = vax1.x ~ XIgroup * Vax1W1_grp, data = Dmgxs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.7340 -0.6567  0.1729  0.3433  3.1372 
## 
## Coefficients:
##                                   Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                        0.73119    0.28851   2.534   0.0117 *  
## XIgroupVaccinefactbox             -0.26529    0.39902  -0.665   0.5066    
## XIgroupVaccinetimeline            -0.11076    0.37897  -0.292   0.7703    
## Vax1W1_grp                         1.33426    0.11770  11.336   <2e-16 ***
## XIgroupVaccinefactbox:Vax1W1_grp   0.06269    0.16490   0.380   0.7040    
## XIgroupVaccinetimeline:Vax1W1_grp  0.06798    0.15893   0.428   0.6691    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.001 on 355 degrees of freedom
## Multiple R-squared:  0.5596, Adjusted R-squared:  0.5534 
## F-statistic: 90.21 on 5 and 355 DF,  p-value: < 2.2e-16

Run the same again but looking at effect of interventions on time to get vaccine controling for wave 1 vaccine interest

## 
## Call:
## lm(formula = vax2ipsos ~ XIgroup * Vax1W1_grp, data = Dmgxs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.2285 -0.4976 -0.4453  1.1916  6.5285 
## 
## Coefficients:
##                                   Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                         9.0368     0.6121  14.763   <2e-16 ***
## XIgroupVaccinefactbox               0.4401     0.7998   0.550    0.583    
## XIgroupVaccinetimeline             -0.4167     0.8027  -0.519    0.604    
## Vax1W1_grp                         -2.5131     0.2401 -10.467   <2e-16 ***
## XIgroupVaccinefactbox:Vax1W1_grp   -0.1554     0.3203  -0.485    0.628    
## XIgroupVaccinetimeline:Vax1W1_grp   0.1215     0.3220   0.377    0.706    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.783 on 303 degrees of freedom
##   (52 observations deleted due to missingness)
## Multiple R-squared:  0.5663, Adjusted R-squared:  0.5592 
## F-statistic: 79.14 on 5 and 303 DF,  p-value: < 2.2e-16

5.5 10Secs

Now looking at the same primary analyses, but subset the data to only include participants who spent at least 10 seconds looking at the messages (if they were in the experimental conditions).

This gives us the following distribution of participants across conditions.

## 
## controlcondition   Vaccinefactbox  Vaccinetimeline 
##              120               94               84

5.5.1 Vax interest

Overall descriptives for interest in getting vaccinated from those who spent more than 10seconds on the messages

##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 298 3.65 1.51      4    3.81 1.48   1   5     4 -0.66    -1.04 0.09

Checking for assumptions before running ANOVA.

##     V1         
##  Mode:logical  
##  TRUE:298

Plots do not look too extreme and levene’s test looks good too 

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 295 4.229 231.4 2.696 0.06911

Get anova results

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 6.804 3.402 1.505 0.224 0.01 0.01 0.003 0.003 0.003 0.101 0.322
…2 Residuals 295 666.9 2.261

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 3.833 0.1373 295 3.563 4.103
Vaccinefactbox 3.553 0.1551 295 3.248 3.858
Vaccinetimeline 3.5 0.1641 295 3.177 3.823

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars   n mean  sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 120 3.83 1.4      4    4.04 1.48   1   5     4 -0.89    -0.56 0.13
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 94 3.55 1.58      4    3.68 1.48   1   5     4 -0.55    -1.26 0.16
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 84  3.5 1.56      4    3.62 1.48   1   5     4 -0.45    -1.35 0.17

Plotting means for visual inspection

5.5.2 Vax time

Overall descriptives for time before getting vaccinated…

##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 255 3.21 2.73      2    2.89 1.48   1   8     7 0.85    -0.91 0.17

Checking for assumptions before running ANOVA. Plots do not look too extreme and levene’s test looks good too 

## Warning: Converting "partno" to factor for ANOVA.
## Warning: Data is unbalanced (unequal N per group). Make sure you specified a
## well-considered value for the type argument to ezANOVA().
## Coefficient covariances computed by hccm()

  • Levene’s Test for Homogeneity of Variance:

    DFn DFd SSn SSd F p p<.05
    2 252 9.729 1377 0.8901 0.4119

Get anova results

Get results with additional effect sizes and power

  term df sumsq meansq statistic p.value etasq partial.etasq omegasq partial.omegasq epsilonsq cohens.f power
XIgroup XIgroup 2 16.08 8.042 1.077 0.342 0.008 0.008 0.001 0.001 0.001 0.092 0.24
…2 Residuals 252 1882 7.468

Checking group means

Show least squares means (unadjusted) and CIs around means

XIgroup lsmean SE df lower.CL upper.CL
controlcondition 2.902 0.2706 252 2.369 3.435
Vaccinefactbox 3.442 0.2947 252 2.862 4.022
Vaccinetimeline 3.373 0.3339 252 2.716 4.031

Also generic means across groups and other descriptives (e.g., skew, range, etc.)

## 
##  Descriptive statistics by group 
## group: controlcondition
##    vars   n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 102  2.9 2.58      1    2.51   0   1   8     7 1.09    -0.34 0.26
## ------------------------------------------------------------ 
## group: Vaccinefactbox
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 86 3.44 2.84      2     3.2 1.48   1   8     7 0.67    -1.27 0.31
## ------------------------------------------------------------ 
## group: Vaccinetimeline
##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 67 3.37 2.82      2    3.13 1.48   1   8     7 0.72    -1.17 0.34

Plotting means for visual inspection

6 Plots

Figure 1

Main effect of messages and control on vaccine interest and time before getting vaccinated.

Figure 2

Main effect of messages and control on vaccine interest and time before getting vaccinated with respondents based on their vaccine intentions at wave 1.

## quartz_off_screen 
##                 2

Figure 3

Main effect of messages and control on vaccine interest and time before getting vaccinated with respondents who spent at least 10 seconds on the messages and with all respondents (overall).

Figure 4

Main effect of messages and control on vaccine perceptions.