| cpr_classify_signif {canaper} | R Documentation |
Classify statistical significance
Description
Given the results of cpr_rand_test(), classifies statistical significance
of a biodiversity metric. The null hypothesis is that observed value does not
lie in the extreme of the random values.
Usage
cpr_classify_signif(df, metric, one_sided = FALSE, upper = FALSE)
Arguments
df |
Input data frame. |
metric |
Character vector of length 1; selected metric to classify
significance. May choose from |
one_sided |
Logical vector of length 1; is the null hypothesis
one-sided? If |
upper |
Logical vector of length 1; only applies if |
Details
For metrics like pe, you probably want to consider a one-sided
hypothesis testing values in the upper extreme (i.e., we are interested in
areas that have higher than expected endemism). For this, you would set
one_sided = TRUE, upper = TRUE. For metrics like pd, you probably want
to consider a two-sided hypothesis (i.e., we are interested in areas that
are either more diverse or less than diverse than expected at random). For
this, set one_sided = FALSE.
Value
Object of class data.frame with column added for statistical
significance of the selected metric. The new column name is the name of the
metric with _signif appended. The new column is a character that may
contain the following values, depending on the null hypothesis:
-
< 0.01,< 0.025,> 0.975,> 0.99,not significant(two-sided) -
< 0.01,< 0.05,> 0.99,> 0.95,not significant(one-sided)
Examples
set.seed(12345)
data(phylocom)
rand_test <- cpr_rand_test(
phylocom$comm, phylocom$phy,
null_model = "curveball", metrics = "pd", n_reps = 50
)
cpr_classify_signif(rand_test, "pd")