Bullet Phase test

Determine a p-value for the strength of similarity between two bullets based on scores for each land-to-land comparison.

Usage

phase_test(land1, land2, score, sigma_0 = NA, alpha = 0.05)

Arguments

land1

vector with land ids of the LEAs in bullet 1

land2

vector with land ids of the LEAs in bullet 2

score

real-valued vector consisting of land-to-land comparisons. Higher values are assumed to describe higher similarity.

sigma_0

real valued, standard deviation of the different source distribution. If NA (default), the standard deviation of the scores corresponding to non-matches (after selecting the highest phase) is taken. Real-values are interpreted as fixed standard deviation. The user can provide their own function of the form function (data), where data is a data frame with land1, land2, score, and phase. The phase vector consists of integer values 1, ..., k, where k is the maximum of the number of unique lands in land1 and land2. The values are ordered such that the highest value k of phase corresponds to the elements in score with the highest average score.

alpha

significance level for a rejection, acceptable Type 1 error.

Value

phase.test object - special case of an htest (hypothesis test). Additionally includes the data used.

Examples

score <- rnorm(36)
phase_test(land1 = rep(1:6, 6), land2 = rep(1:6, each = 6), score = score)
#> 
#> 	Phase Test
#> 
#> data:  
#> T = 0.71524, sigma_0 = 0.73631, df = 6, p-value = 0.7481
#> alternative hypothesis: true difference in means between group SS and group DS is greater than 0 (indicating same-source)
#> sample estimates:
#>  SS (mean)  DS (mean) 
#>  0.4073304 -0.3079061 
#>