Skip to contents

Possible GRIM inconsistencies

Source: R/grim-stats.R
grim-stats.Rd

Even without GRIM-testing, means / proportions and sample sizes of granular distributions entail some key data:

  • grim_total() returns the absolute number of GRIM-inconsistencies that are possible given the mean or percentage's number of decimal places (D) and the corresponding sample size.

  • grim_ratio() returns a proportion that is normalized by 10^D, and therefore comparable across mean or percentage values reported to varying D.

  • grim_ratio_upper() returns the upper bound of grim_ratio() for a given D.

For discussion, see vignette("grim"), section GRIM statistics.

Usage

grim_total(x, n, items = 1, percent = FALSE)

grim_ratio(x, n, items = 1, percent = FALSE)

grim_ratio_upper(x, percent = FALSE)

Arguments

x

String or numeric (length 1). Mean or percentage value computed from data with integer units (e.g., mean scores on a Likert scale or percentage of study participants in some condition). Note: Numeric inputs don't include trailing zeros, but these are important for GRIM functions. See documentation for grim().

n

Integer. Sample size corresponding to x.

items

Integer. Number of items composing the mean or percentage value in question. Default is 1.

percent

Logical. Set percent to TRUE if x is expressed as a proportion of 100 rather than 1. The functions will then account for this fact through increasing the decimal count by 2. Default is FALSE.

Value

Integer or double. The number or proportion of possible GRIM inconsistencies.

References

Brown, N. J. L., & Heathers, J. A. J. (2017). The GRIM Test: A Simple Technique Detects Numerous Anomalies in the Reporting of Results in Psychology. Social Psychological and Personality Science, 8(4), 363–369. https://journals.sagepub.com/doi/10.1177/1948550616673876

See also

grim() for the GRIM test itself; as well as grim_map() for applying it to many cases at once.

Examples

# Many value sets are inconsistent here:
grim_total(x = "83.29", n = 21)
#> [1] 79
grim_ratio(x = "83.29", n = 21)
#> [1] 0.79

# No sets are inconsistent in this case...
grim_total(x = "5.14", n = 83)
#> [1] 17
grim_ratio(x = "5.14", n = 83)
#> [1] 0.17

# ... but most would be if `x` was a percentage:
grim_total(x = "5.14", n = 83, percent = TRUE)
#> [1] 9917
grim_ratio(x = "5.14", n = 83, percent = TRUE)
#> [1] 0.9917