Test whether a confidence interval excludes a given value

Tests whether a confidence interval excldes a specified reference value. This will generally be the null value for the relevant test, such that excluding this value indicates the test is statistically significant.

Usage

ci_sig(
  ll,
  ul,
  reference = 1,
  return = c("logical", "asterisks"),
  marks = c("*", NA_character_)
)

Arguments

ll

a numeric vector containing confidence interval lower limits.

ul

a numeric vector containing corresponding upper limits.

reference

the value to check against. This will generally be the null value for the relevant test (e.g., 1 for odds ratios, 0 for beta coefficients).

return

if "logical", will return a logical vector indicating whether each confidence interval excludes reference. If asterisks, will return a character vector, using the characters passed to marks.

marks

a length-2 vector specifying strings to mark significant and non-significant results when return = "asterisks".

Value

• If return = \"logical\" (default), a logical vector.

• If return = \"asterisks\", a character vector.

Examples

beta_CIs <- glm(
    Survived ~ Sex * Age,
    family = binomial,
    weights = Freq,
    data = as.data.frame(Titanic)
  ) %>%
  confint()
#> Waiting for profiling to be done...
OR_CIs <- tibble::as_tibble(exp(beta_CIs), rownames = "term")
beta_CIs <- tibble::as_tibble(beta_CIs, rownames = "term")

beta_CIs %>%
  dplyr::mutate(sig = ci_sig(`2.5 %`, `97.5 %`))
#> # A tibble: 4 × 4
#>   term               `2.5 %` `97.5 %` sig  
#>   <chr>                <dbl>    <dbl> <lgl>
#> 1 (Intercept)        -0.687     0.303 TRUE 
#> 2 SexFemale          -0.0834    1.48  FALSE
#> 3 AgeAdult           -1.69     -0.669 TRUE 
#> 4 SexFemale:AgeAdult  0.918     2.56  FALSE

beta_CIs %>%
  dplyr::mutate(sig = ci_sig(`2.5 %`, `97.5 %`, return = "asterisks"))
#> # A tibble: 4 × 4
#>   term               `2.5 %` `97.5 %` sig  
#>   <chr>                <dbl>    <dbl> <chr>
#> 1 (Intercept)        -0.687     0.303 *    
#> 2 SexFemale          -0.0834    1.48  NA   
#> 3 AgeAdult           -1.69     -0.669 *    
#> 4 SexFemale:AgeAdult  0.918     2.56  NA   

beta_CIs %>%
  dplyr::mutate(sig = ci_sig(`2.5 %`, `97.5 %`, return = "asterisks", marks = c("*", "ns")))
#> # A tibble: 4 × 4
#>   term               `2.5 %` `97.5 %` sig  
#>   <chr>                <dbl>    <dbl> <chr>
#> 1 (Intercept)        -0.687     0.303 *    
#> 2 SexFemale          -0.0834    1.48  ns   
#> 3 AgeAdult           -1.69     -0.669 *    
#> 4 SexFemale:AgeAdult  0.918     2.56  ns   

OR_CIs %>%
  dplyr::mutate(sig = ci_sig(`2.5 %`, `97.5 %`, reference = 1, return = "asterisks"))
#> # A tibble: 4 × 4
#>   term               `2.5 %` `97.5 %` sig  
#>   <chr>                <dbl>    <dbl> <chr>
#> 1 (Intercept)          0.503    1.35  NA   
#> 2 SexFemale            0.920    4.39  NA   
#> 3 AgeAdult             0.185    0.512 *    
#> 4 SexFemale:AgeAdult   2.50    12.9   *