Compare the Statistical Differences between Classifications Results

The `rict_compare` function allows the user to assess whether there is a real difference in EQR values and/or status class between a pair of results and/or sites and/or time periods.

Usage

rict_compare(results_a = NULL, results_b = NULL)

Arguments

results_a: Dataframe output from `rict(store_eqrs = T)`. The `results_b` and `results_b` input dataframes must have the same number of rows
results_b: Dataframe output from `rict(store_eqrs = T)`. The `results_b` and `results_b` input dataframes must have the same number of rows

Value

Dataframe of compare results with 44 variables see `technical specifications` for details. Returns output of results the same length as the input files.

EQR metric compared: EQR metric from Result A which is being compared ("Vs") the metric in Result B for example SUM_ASPT Vs SUM_ASPT - this row is comparing Summer ASPT results A to Summer ASPT in results B. Note, if only one season is calculated, the "AVG_"... EQR metric (average) will have the same value as the season EQR value
Result A: Concatennated results name of Year and Site
Result B: Concatennated results name of Year and Site
Average EQR for Result A: Average EQR
Average EQR for Result B: Average EQR
Average EQR: Difference (B - A): Difference in EQR
Standard Deviation of Difference: Standard Deviation of Difference
Lower 95% (L95) of Difference: Lower 95% (L95) of Difference
Upper 95% (U95) of Difference: Upper 95% (U95) of Difference
2-sided test probability p of No difference in EQR: p of probability test
Most Probable Class for Result A: Most Probable Class for Result A
Probability of most likely class for Result A: Probability of most likely class for Result A
Most Probable Class for Result B: Most Probable Class for Result B
Probability of most likely class for Result B: Probability of most likely class for Result B
Probability B more than one class Worse than A: Probability B more than one class Worse than A
Probability B one class Worse than A: Probability B one class Worse than A
Probability B same Class as A: Probability B same Class as A
Probability B one class Better than A: Probability B one class Better than A
Probability B more than one class Better than A: Probability B more than one class Better than A
Probability Result A in High & Result B in High: Probability...
Probability Result A in High & Result B in Good: Probability...
Probability Result A in High & Result B in Moderate: Probability...
Probability Result A in High & Result B in Poor: Probability...
Probability Result A in High & Result B in Bad: Probability...
Probability Result A in Good & Result B in High: Probability...
Probability Result A in Good & Result B in Good: ...
Probability Result A in Good & Result B in Moderate: ...
Probability Result A in Good & Result B in Poor: ...
Probability Result A in Good & Result B in Bad: ...
Probability Result A in Moderate & Result B in High: ...
Probability Result A in Moderate & Result B in Good: ...
Probability Result A in Moderate & Result B in Moderate: ...
Probability Result A in Moderate & Result B in Poor: ...
Probability Result A in Moderate & Result B in Bad: ...
Probability Result A in Poor & Result B in High: ...
Probability Result A in Poor & Result B in Good: ...
Probability Result A in Poor & Result B in Moderate: ...
Probability Result A in Poor & Result B in Poor: ...
Probability Result A in Poor & Result B in Bad: ...
Probability Result A in Bad & Result B in High: ...
Probability Result A in Bad & Result B in Good: ...
Probability Result A in Bad & Result B in Moderate: ...
Probability Result A in Bad & Result B in Poor: ...
Probability Result A in Bad & Result B in Bad: ...

Details

The `rict_compare` makes use of two lower level functions: compare_test and compare_probability. Before passing data to these functions it formats the data and removes results for example the 'MINTA' values as these are not applicable for `compare_test()`.

Examples

if (FALSE) {
results_a <- rict(demo_observed_values[1, ], store_eqrs = TRUE, year_type = "single")
results_b <- rict(demo_gis_values_log[2, ], store_eqrs = TRUE, year_type = "single")
compare <- rict_compare(results_a, results_b)
}