effect: Consider the effect of particular places upon the ID

In MLID: Multilevel Index of Dissimilarity

Description

Evaluates the effect on the index of the named places under three different scenarios.

Usage

effect(object, places)

Arguments

object	an object of class index: a multilevel index created using function id
places	a character vector containing the names of the places in any of the higher-level geographies for which the evaluation will be made

Details

The three different scenarios considered are:

1. if the effects (the estimated residuals from the multilevel model) are set to zero for the named higher-level places;

2. if the shares of the two population groups are equal everywhere except within the named places; and

3. if all but the neighbourhoods within the named places are omitted from the data and the index then recalculated using only those that remain.

The evaluation also includes:

• the impact of the chosen places upon the overall ID, where a value over 100 indicates a group of neighbourhoods with a disproportionately high (in the same way that impacts calculates it)

• an R-squared value. This is the proportion of the total variation in (Y - X) that is due to the chosen places, where (Y - X) are the differences between the share of population group Y and the share of population group X that are observed in each neighbourhood.

See vignette("MLID") for further details

Value

an object, primarily a list containing the evaluated values

See Also

id

Harris R (2017) Fitting a multilevel index of segregation in R: using the MLID package http://rpubs.com/profrichharris/MLID

Examples

## Not run: 
data(aggdata)
index <- id(aggdata, vars = c("Bangladeshi", "WhiteBrit"),
levels = c("MSOA","LAD","RGN"))
ci <- confint(index)
catplot(ci)
# Note Tower Hamlets and Newham. Obtain the predictions for them:
effect(index, "Tower Hamlets")
effect(index, "Newham")
effect(index, c("Tower Hamlets","Newham"))

## End(Not run)

MLID documentation built on May 2, 2019, 11:05 a.m.