SetMethods (version 4.0)

QCAfit: Function calculating the parameters of fit

Description

QCAfit is a function calculating parameters of fit useful in QCA and fsQCA that are consistency, coverage, PRI, Haesebrouck's consistency, RoN and PRODUCT. It works with both single and multiple conditions.

Usage

QCAfit(x, y, cond.lab = NULL, necessity = TRUE, neg.out = FALSE, 
              product = FALSE, sol=1, ttrows= c(), consH = FALSE)

Value

It returns a matrix containing the parameters of fit for each condition.

Arguments

x

A vector containing the values of a condition, a matrix with more than one conditions, or an object of class "qca" when necessity is FALSE and when outcome is specifyied as a character string.

y

A vector containing the values of the output or a character string when y is of class "qca".

cond.lab

When inserting a dataframe or a matrix with more than one condition and column names, the function automatically prints the names of the conditions tested. When inputing a vector, hence a single condition (i.e. a single column in a dataframe, the name of the condition tested should be inserted in this option .)

necessity

logical. It indicates if the output should be for sufficient or necessary condition(s). By default, FALSE, the function returns a table of parameters of fit for sufficient condition(s) (Consistency, Coverage, PRI, Haesebrouck's Consistency, and optionally Product). When it set to TRUE the function returns a table of parameters of fit for necessary condition(s) (Consistency, Coverage, Relevance of Necessity).

neg.out

logical. It indicates if the parameters of fit should be computed for the positive or the negative outcome. By default, FALSE, the function returns parameters of fit for the positive outcome .

product

logical. It indicates whether the parameter of fit PRODUCT should be shown. This stands for the product between the consistency sufficiency parameter and the PRI parameter.

sol

A vector where the first number indicates the number of the conservative or parsimonious solution according to the order in the "qca" object. For more complicated structures of model ambiguity, the intermediate solution can also be specified by using a character string of the form "c1p3i2" where c = conservative solution, p = parsimonious solution and i = intermediate solution.

ttrows

A vector specifying the names of the truth table rows for which the function reports parameters of fit. For using this option y must be a "qca" object.

consH

Logical. Print also the Haesebrouck's consistency among the parameters of fit?

Author

Mario Quaranta and Ioana-Elena Oana

References

Haesebrouck, T. (2015) Pitfalls in QCA's consistency measure. Journal of Comparative Politics 2:65-80.

Ragin, C. C. 2006. Set Relations in Social Research: Evaluating Their Consistency and Coverage. Political Analysis 14(3): 291-310.

Schneider, C. Q., Wagemann, C. (2012) Set-Theoretic Methods for the Social Sciences, Cambridge University Press: Cambridge.

See Also

minimize

Examples

# Generate fake data
set.seed(1234)

a <- runif(100, 0, 1)
b <- runif(100, 0, 1)
c <- runif(100, 0, 1)
y <- runif(100, 0, 1)

# Only one condition, for necessity
QCAfit(a, y, cond.lab = "A")

# With three conditions and their negation, for necessity
QCAfit(cbind(a, b, c), y)

# Only one condition, for sufficiency
QCAfit(a, y, cond.lab = "A", necessity = FALSE)

# With three conditions, their negation and negated output, for necessity
QCAfit(cbind(a, b, c), y, neg.out = TRUE)

# Load the Schneider data:

data(SCHF)

# Get parameters of fit for condition EMP as necessary for outcome EXPORT:

QCAfit(SCHF$EMP, SCHF$EXPORT, cond.lab = "EMP")

# Get parameters of fit for condition ~EMP as necessary for outcome ~EXPORT:

QCAfit(1-SCHF$EMP, SCHF$EXPORT, neg.out=TRUE, cond.lab = "~EMP")

# Get parameters of fit for all conditions and their negation as necessary for outcome EXPORT:

QCAfit(SCHF[,1:6], SCHF$EXPORT)

# Obtain the parsimonious solution for outcome "EXPORT":

sol_yp <- minimize(SCHF, outcome = "EXPORT",
                conditions = c("EMP","BARGAIN","UNI","OCCUP","STOCK", "MA"),
                incl.cut = .9, 
                include = "?", 					   
                details = TRUE, show.cases = TRUE)
                
# Get parameters of fit for the parsimonious solution:

QCAfit(x = sol_yp, y = "EXPORT", necessity = FALSE)

# Get parameters of fit for truth table rows 2,8, and 10:

QCAfit(x = sol_yp, y = "EXPORT", ttrows=c("2","8","10"), necessity = FALSE)