Plot proportion or count of responses

Usage

plot_responses(
  df,
  x,
  y,
  ymin = NULL,
  ymax = NULL,
  ribbon_ymin = NULL,
  ribbon_ymax = NULL,
  color = NULL,
  group = NULL,
  cutpoints = NULL,
  fill_interval = c("gray90", "gray85")
)

Arguments

df: Table with data
x: Name of the response level variable
y: Name of variable with statistic, either proportion P or counts N
ymin: Name of variable with lower limit for error bar, e.g., LowerCI. Optional.
ymax: Name of variable with lower limit for error bar, e.g., UpperCI. Optional.
ribbon_ymin: Name of variable with lower limit for ribbon, e.g., PosteriorLowerCI. Optional.
ribbon_ymax: Name of variable with upper limit for ribbon, e.g., PosteriorUpperCI. Optional.
color: Name of variable that maps color. Optional.
group: Name of variable that maps grouping. Optional.
cutpoints: Optional vector of estimated cut points. In case of the matrix, average as computed per column
fill_interval: Fill colors for stripes that indicate intervals.

Value

ggplot

Examples

data(aiq)
aiq_avg <- bootstrap_response_counts(aiq, Response ~ Group + Question, R = 100)
#>  ■■■■■                             13% |  ETA: 11s
#>  ■■■■■■■■■■■■■                     39% |  ETA:  7s
#>  ■■■■■■■■■■■■■■■■■■■■■             65% |  ETA:  4s
#>  ■■■■■■■■■■■■■■■■■■■■■■■■■■■■      90% |  ETA:  1s
aiq_q1 <- dplyr::filter(aiq_avg, Question == Question[1])
plot_responses(aiq_q1, x = Response, y = P, color = Group)

#>            x      xend PANEL group alpha   fill
#> 1 0.03703704 0.1693122     1    -1     1 gray90
#> 2 0.16931217 0.3015873     1    -1     1 gray85
#> 3 0.30158730 0.4338624     1    -1     1 gray90
#> 4 0.43386243 0.5661376     1    -1     1 gray85
#> 5 0.56613757 0.6984127     1    -1     1 gray90
#> 6 0.69841270 0.8306878     1    -1     1 gray85
#> 7 0.83068783 0.9629630     1    -1     1 gray90

# with errorbars
plot_responses(aiq_q1, x = Response, y = P, ymin = LowerCI, ymax = UpperCI, color = Group)

#>            x      xend PANEL group alpha   fill
#> 1 0.03703704 0.1693122     1    -1     1 gray90
#> 2 0.16931217 0.3015873     1    -1     1 gray85
#> 3 0.30158730 0.4338624     1    -1     1 gray90
#> 4 0.43386243 0.5661376     1    -1     1 gray85
#> 5 0.56613757 0.6984127     1    -1     1 gray90
#> 6 0.69841270 0.8306878     1    -1     1 gray85
#> 7 0.83068783 0.9629630     1    -1     1 gray90

# with ribbon intead of error bars
plot_responses(aiq_q1, x = Response, y = P,
               ribbon_ymin = LowerCI, ribbon_ymax = UpperCI, color = Group)

#>            x      xend PANEL group alpha   fill
#> 1 0.03703704 0.1693122     1    -1     1 gray90
#> 2 0.16931217 0.3015873     1    -1     1 gray85
#> 3 0.30158730 0.4338624     1    -1     1 gray90
#> 4 0.43386243 0.5661376     1    -1     1 gray85
#> 5 0.56613757 0.6984127     1    -1     1 gray90
#> 6 0.69841270 0.8306878     1    -1     1 gray85
#> 7 0.83068783 0.9629630     1    -1     1 gray90

# with cutpoints from the model fit
data(aiq_draws)
cutpoints_draws <- extract_stan_posterior_matrix(aiq_draws, cutpoints)
plot_responses(aiq_q1, x = Response, y = P, ymin = LowerCI, ymax = UpperCI,
               color = Group, cutpoints = inv_logit(colMeans(cutpoints_draws)))

#>            x       xend PANEL group alpha   fill
#> 1 0.03703704 0.04360385     1    -1     1 gray90
#> 2 0.04360385 0.09866389     1    -1     1 gray85
#> 3 0.09866389 0.33707025     1    -1     1 gray90
#> 4 0.33707025 0.61361725     1    -1     1 gray85
#> 5 0.61361725 0.92193144     1    -1     1 gray90
#> 6 0.92193144 0.96101913     1    -1     1 gray85
#> 7 0.96101913 0.96296296     1    -1     1 gray90