Skip to contents

Confidence interval for a mean absolute prediction error

Source: R/statpsych2.R
ci.mape.Rd

Computes a confidence interval for a population mean absolute prediction error (MAPE) in a general linear model. The MAPE is a more robust alternative to the residual standard deviation. This function requires a vector of estimated residuals from a general linear model. This confidence interval does not assume zero excess kurtosis but does assume symmetry of the population prediction errors.

For more details, see Section 1.16 of Bonett (2021, Volume 2)

Usage

ci.mape(alpha, res, s)

Arguments

alpha

alpha level for 1-alpha confidence

res

vector of residuals

s

number of predictor variables in model

Value

Returns a 1-row matrix. The columns are:

  • Estimate - estimated mean absolute prediction error

  • SE - standard error

  • LL - lower limit of the confidence interval

  • UL - upper limit of the confidence interval

References

Bonett DG (2021). Statistical Methods for Psychologists https://dgbonett.sites.ucsc.edu/.

Examples

res <- c(-2.70, -2.69, -1.32, 1.02, 1.23, -1.46, 2.21, -2.10, 2.56,
      -3.02, -1.55, 1.46, 4.02, 2.34)
ci.mape(.05, res, 1)
#>  Estimate        SE       LL       UL
#>    2.3744 0.3314752 1.806022 3.121654

# Should return:
# Estimate        SE       LL       UL
#   2.3744 0.3314752 1.751678 3.218499