sjstats (version 0.17.5)

chisq_gof: Compute model quality

Description

For logistic regression models, performs a Chi-squared goodness-of-fit-test.

Usage

chisq_gof(x, prob = NULL, weights = NULL)

Arguments

x

A numeric vector or a glm-object.

prob

Vector of probabilities (indicating the population probabilities) of the same length as x's amount of categories / factor levels. Use nrow(table(x)) to determine the amount of necessary values for prob. Only used, when x is a vector, and not a glm-object.

weights

Vector with weights, used to weight x.

...

More fitted model objects, to compute multiple coefficients of variation at once.

Value

For vectors, returns the object of the computed chisq.test. For glm-objects, an object of class chisq_gof with following values: p.value, the p-value for the goodness-of-fit test; z.score, the standardized z-score for the goodness-of-fit test; rss, the residual sums of squares term and chisq, the pearson chi-squared statistic.

Details

For vectors, this function is a convenient function for the chisq.test(), performing goodness-of-fit test. For glm-objects, this function performs a goodness-of-fit test. A well-fitting model shows no significant difference between the model and the observed data, i.e. the reported p-values should be greater than 0.05.

References

Hosmer, D. W., & Lemeshow, S. (2000). Applied Logistic Regression. Hoboken, NJ, USA: John Wiley & Sons, Inc. 10.1002/0471722146

Examples

# NOT RUN {
data(efc)
efc$neg_c_7d <- ifelse(efc$neg_c_7 < median(efc$neg_c_7, na.rm = TRUE), 0, 1)
m <- glm(
  neg_c_7d ~ c161sex + barthtot + c172code,
  data = efc,
  family = binomial(link = "logit")
)

# goodness-of-fit test for logistic regression
chisq_gof(m)

# goodness-of-fit test for vectors against probabilities
# differing from population
chisq_gof(efc$e42dep, c(0.3,0.2,0.22,0.28))

# equal to population
chisq_gof(efc$e42dep, prop.table(table(efc$e42dep)))

# }