Exclude extreme estimates above a threshold parameter sd

Usage

exclude_extreme_estimates(
  data,
  outcome_variable,
  outcome_SE,
  sd_threshold = numeric(1L),
  param_table,
  .fn = ...,
  ...
)

Arguments

data: A dataframe of analyst estimates
outcome_variable: the name of the variable in data containing the analyst estimates
outcome_SE: variable in data containing analyst SE estimates
sd_threshold: A numeric threshold multiplyer see details
param_table: A dataframe containing population parameters mean and sd for each variable in a given dataset
.fn: An optional function that will transform parameter estimates to the same scale as outcome_variable in data
...: Arguments supplied to .fn

Value

A dataframe of analyst estimates with extreme estimates excluded

Details

This function is used to exclude extreme estimates from a dataset. The function calculates a threshold for exclusion based on the mean and standard deviation of the population parameter estimates in param_table. The threshold is calculated as the mean of the population parameter plus sd_threshold times the standard deviation of the population parameter. Estimates in data that are greater than this threshold are excluded from the output.

If the user chooses to supply .fn and ... arguments, the function will transform the population parameter estimates in param_table to the same scale as the outcome_variable in data using .fn, before calculating the threshold for exclusion.

Examples

# example code
data <-   ManyEcoEvo_yi %>% 
mutate(data = 
         map_if(data, 
                ~ filter(.x, 
                         stringr::str_detect(response_variable_name, 
                                             "average.proportion.of.plots.containing",
                                             negate = TRUE)),
                .p = dataset == "eucalyptus")) %>%   
  mutate(
    diversity_data =
      map2(
        .x = diversity_data,
        .y = data,
        .f = ~ semi_join(.x, .y, join_by(id_col)) %>% 
          distinct()
      )
  ) %>% 
  prepare_response_variables(
    estimate_type = "yi",
    param_table = 
      ManyEcoEvo:::analysis_data_param_tables, 
    dataset_standardise = "blue tit",
    dataset_log_transform = "eucalyptus") %>%
  generate_yi_subsets() %>% #TODO: must be run after prepare_response_variables??
  apply_VZ_exclusions(
    VZ_colname = list("eucalyptus" = "se_log", 
                      "blue tit" = "VZ"), 
    VZ_cutoff = 3) %>% 
  filter(dataset == "eucalyptus", estimate_type == "y25")  %>% 
  pluck("data", 1)
#> Error in pluck(., "data", 1): could not find function "pluck"
sd_threshold = 3
param_table <- ManyEcoEvo:::analysis_data_param_tables
exclude_extreme_estimates(data, "mean_log", "se_log", 3, param_table, log_transform, estimate = mean, std.error = sd)
#> 
#> ── Transforming `param_table` using `.fn`: 
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> ✔ Log-transformed out-of-sample predictions, using 10000 simulations.
#> 
#> ── Excluding extreme estimates from data: 
#> Error in UseMethod("left_join"): no applicable method for 'left_join' applied to an object of class "function"