Compare reading levels across conditions

Estimate Flesch-Kincaid reading levels and compare their condition means. The function uses quanteda.textstats when it's installed and a syllable heuristic otherwise. The output tells you which method it used.

Usage

check_readability(texts, conditions, threshold = 2)

Arguments

texts

Character vector of materials.

conditions

Condition labels.

threshold

Maximum acceptable difference in mean grade level between any two conditions (default: 2.0).

Value

A list with per_text, by_condition, max_diff, unmeasurable (conditions with no computable grade level, which fail the check), method, n_missing, n_no_condition, and pass.

Details

This is an English-language measure. I'd use a suitable measure for the language of the study before interpreting differences in other texts.

Examples

check_readability(
  c("The policy creates jobs by investing in new infrastructure.",
    "We must act now to protect future generations from harm."),
  c("economic", "moral")
)
#> ! Readability: grade-level spread 4.2 (> 2)
#> $per_text
#> # A tibble: 2 × 3
#>   condition text                                                        fk_grade
#>   <chr>     <chr>                                                          <dbl>
#> 1 economic  The policy creates jobs by investing in new infrastructure.     10.2
#> 2 moral     We must act now to protect future generations from harm.         6  
#> 
#> $by_condition
#> # A tibble: 2 × 4
#>   condition     n mean_grade sd_grade
#>   <chr>     <int>      <dbl>    <dbl>
#> 1 economic      1       10.2       NA
#> 2 moral         1        6         NA
#> 
#> $max_diff
#> [1] 4.2
#> 
#> $unmeasurable
#> character(0)
#> 
#> $method
#> [1] "quanteda.textstats::textstat_readability (Flesch.Kincaid)"
#> 
#> $n_missing
#> [1] 0
#> 
#> $n_no_condition
#> [1] 0
#> 
#> $pass
#> [1] FALSE
#>