# conTestF: F-bar test for iht In restriktor: Restricted Statistical Estimation and Inference for Linear Models

## Description

`conTestF` tests linear equality and/or inequality restricted hypotheses for linear models by F-tests. It can be used directly and is called by the `conTest` function if `test = "F"`.

## Usage

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17``` ```## S3 method for class 'conLM' conTestF(object, type = "A", neq.alt = 0, boot = "no", R = 9999, p.distr = rnorm, parallel = "no", ncpus = 1L, cl = NULL, seed = 1234, verbose = FALSE, control = NULL, ...) ## S3 method for class 'conRLM' conTestF(object, type = "A", neq.alt = 0, boot = "no", R = 9999, p.distr = rnorm, parallel = "no", ncpus = 1L, cl = NULL, seed = 1234, verbose = FALSE, control = NULL, ...) ## S3 method for class 'conGLM' conTestF(object, type = "A", neq.alt = 0, boot = "no", R = 9999, p.distr = rnorm, parallel = "no", ncpus = 1L, cl = NULL, seed = 1234, verbose = FALSE, control = NULL, ...) ```

## Arguments

 `object` an object of class `conLM`, `conRLM` or `conGLM`. `type` hypothesis test type "A", "B", "C", "global", or "summary" (default). See details for more information. `neq.alt` integer: number of equality constraints that are maintained under the alternative hypothesis (for hypothesis test type "B"), see example 3. `boot` the null-distribution of these test-statistics (except under type "C") takes the form of a mixture of F-distributions. The tail probabilities can be computed directly via bootstrapping; if `"parametric"`, the p-value is computed based on the parametric bootstrap. By default, samples are drawn from a normal distribution with mean zero and varance one. See `p.distr` for other distributional options. If `"model.based"`, a model-based bootstrap method is used. Instead of computing the p-value via simulation, the p-value can also be computed using the chi-bar-square weights. If `"no"`, the p-value is computed based on the weights obtained via simulation (`mix.weights = "boot"`) or using the multivariate normal distribution function (`mix.weights = "pmvnorm"`). Note that, these weights are already available in the restriktor objected and do not need to be estimated again. However, there are two exception for objects of class `conRLM`, namely for computing the p-value for the robust `test` = `"Wald"` and the robust `"score"`. In these cases the weights need to be recalculated. `R` integer; number of bootstrap draws for `boot`. The default value is set to 9999. `p.distr` random generation distribution for the parametric bootstrap. For all available distributions see `?distributions`. For example, if `rnorm`, samples are drawn from the normal distribution (default) with mean zero and variance one. If `rt`, samples are drawn from a t-distribution. If `rchisq`, samples are drawn from a chi-square distribution. The distributional parameters will be passed in via .... `parallel` the type of parallel operation to be used (if any). If missing, the default is set "no". `ncpus` integer: number of processes to be used in parallel operation: typically one would chose this to the number of available CPUs. `cl` an optional parallel or snow cluster for use if parallel = "snow". If not supplied, a cluster on the local machine is created for the duration of the conTest call. `seed` seed value. The default value is set to 1234. `verbose` logical; if TRUE, information is shown at each bootstrap draw. `control` a list of control arguments: `absval` tolerance criterion for convergence (default = sqrt(.Machine\$double.eps)). Only used for model of class lm. `maxit` the maximum number of iterations for the optimizer (default = 10000). Only used for model of class mlm (not yet supported). `tol` numerical tolerance value. Estimates smaller than `tol` are set to 0. `...` additional arguments to be passed to the p.distr function.

## Details

The following hypothesis tests are available:

• Type A: Test H0: all constraints with equalities ("=") active against HA: at least one inequality restriction (">") strictly true.

• Type B: Test H0: all constraints with inequalities (">") (including some equalities ("=")) active against HA: at least one restriction false (some equality constraints may be maintained).

• Type C: Test H0: at least one restriction false ("<") against HA: all constraints strikty true (">"). This test is based on the intersection-union principle (Silvapulle and Sen, 2005, chp 5.3). Note that, this test only makes sense in case of no equality constraints.

• Type global: equal to Type A but H0 contains additional equality constraints. This test is analogue to the global F-test in lm, where all coefficients but the intercept equal 0.

The null-distribution of hypothesis test Type C is based on a t-distribution (one-sided). Its power can be poor in case of many inequalty constraints. Its main role is to prevent wrong conclusions from significant results from hypothesis test Type A.

The exact finite sample distributions of the non-robust F-, score- and LR-test statistics based on restricted OLS estimates and normally distributed errors, are a mixture of F-distributions under the null hypothesis (Wolak, 1987). In agreement with Silvapulle (1992), we found that the results based on these mixtures of F-distributions approximate the tail probabilities of the robust tests better than their asymptotic distributions. Therefore, all p-values for hypothesis test Type `"A"`, `"B"` and `"global"` are computed based on mixtures of F-distributions.

Note that, in case of equality constraints only, the null-distribution of the (robust) F-test statistics is based on an F-distribution. The (robust) Wald- and (robust) score-test statistics are based on chi-square distributions.

## Value

An object of class conTest, for which a print is available. More specifically, it is a list with the following items:

 `CON` a list with useful information about the constraints. `Amat` constraints matrix. `bvec` vector of right-hand side elements. `meq` number of equality constraints. `meq.alt` same as input neq.alt. `iact` number of active constraints. `type` same as input. `test` same as input. `Ts` test-statistic value. `df.residual` the residual degrees of freedom. `pvalue` tail probability for `Ts`. `b.eqrestr` equality restricted regression coefficients. Only available for `type = "A"` and `type = "global"`, else `b.eqrestr = NULL`. `b.unrestr` unrestricted regression coefficients. `b.restr` restricted regression coefficients. `b.restr.alt` restricted regression coefficients under HA if some equality constraints are maintained. Only available for `type = "B"` else `b.restr.alt = NULL`. `Sigma` variance-covariance matrix of unrestricted model. `R2.org` unrestricted R-squared, not available for objects of class `conGLM`. `R2.reduced` restricted R-squared, not available for objects of class `conGLM`. `boot` same as input. `model.org` original model.

## Author(s)

Leonard Vanbrabant and Yves Rosseel

## References

Kudo, A. (1963) A multivariate analogue of the one-sided test. Biometrika, 50, 403–418.

Silvapulle, M. (1992a). Robust tests of inequality constraints and one-sided hypotheses in the linear model. Biometrika, 79, 621–630.

Silvapulle, M. (1996) On an F-type statistic for testing one-sided hypotheses and computation of chi-bar-squared weights. Statistics & probability letters, 28, 137–141.

Silvapulle, M.J. and Sen, P.K. (2005). Constrained Statistical Inference. Wiley, New York

Wolak, F. (1987). An exact test for multiple inequality and equality constraints in the linear regression model. Journal of the American statistical association, 82, 782–793.

## See Also

quadprog, `conTest`

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129``` ```## example 1: # the data consist of ages (in months) at which an # infant starts to walk alone. # prepare data DATA1 <- subset(ZelazoKolb1972, Group != "Control") # fit unrestricted linear model fit1.lm <- lm(Age ~ -1 + Group, data = DATA1) # the variable names can be used to impose constraints on # the corresponding regression parameters. coef(fit1.lm) # constraint syntax: assuming that the walking # exercises would not have a negative effect of increasing the # mean age at which a child starts to walk. myConstraints1 <- ' GroupActive < GroupPassive; GroupPassive < GroupNo ' conTest(fit1.lm, myConstraints1) # another way is to first fit the restricted model fit.restr1 <- restriktor(fit1.lm, constraints = myConstraints1) conTest(fit.restr1) ## Not run: # Or in matrix notation. Amat1 <- rbind(c(-1, 0, 1), c( 0, 1, -1)) myRhs1 <- rep(0L, nrow(Amat1)) myNeq1 <- 0 conTest(fit1.lm, constraints = Amat1, rhs = myRhs1, neq = myNeq1) ## End(Not run) ######################### ## Artificial examples ## ######################### # generate data n <- 10 means <- c(1,2,1,3) nm <- length(means) group <- as.factor(rep(1:nm, each = n)) y <- rnorm(n * nm, rep(means, each = n)) DATA2 <- data.frame(y, group) # fit unrestricted linear model fit2.lm <- lm(y ~ -1 + group, data = DATA2) coef(fit2.lm) ## example 2: increasing means myConstraints2 <- ' group1 < group2 group2 < group3 group3 < group4 ' # compute F-test for hypothesis test Type A and compute the tail # probability based on the parametric bootstrap. We only generate 9 # bootstrap samples in this example; in practice you may wish to # use a much higher number. conTest(fit2.lm, constraints = myConstraints2, type = "A", boot = "parametric", R = 9) # or fit restricted linear model fit2.con <- restriktor(fit2.lm, constraints = myConstraints2) conTest(fit2.con) ## Not run: # increasing means in matrix notation. Amat2 <- rbind(c(-1, 1, 0, 0), c( 0,-1, 1, 0), c( 0, 0,-1, 1)) myRhs2 <- rep(0L, nrow(Amat2)) myNeq2 <- 0 conTest(fit2.con, constraints = Amat2, rhs = myRhs2, neq = myNeq2, type = "A", boot = "parametric", R = 9) ## End(Not run) ## example 3: # combination of equality and inequality constraints. myConstraints3 <- ' group1 == group2 group3 < group4 ' conTest(fit2.lm, constraints = myConstraints3, type = "B", neq.alt = 1) # fit resticted model and compute model-based bootstrapped # standard errors. We only generate 9 bootstrap samples in this # example; in practice you may wish to use a much higher number. # Note that, a warning message may be thrown because the number of # bootstrap samples is too low. fit3.con <- restriktor(fit2.lm, constraints = myConstraints3, se = "boot.model.based", B = 9) conTest(fit3.con, type = "B", neq.alt = 1) ## example 4: # restriktor can also be used to define effects using the := operator # and impose constraints on them. For example, is the # average effect (AVE) larger than zero? # generate data n <- 30 b0 <- 10; b1 = 0.5; b2 = 1; b3 = 1.5 X <- c(rep(c(0), n/2), rep(c(1), n/2)) set.seed(90) Z <- rnorm(n, 16, 5) y <- b0 + b1*X + b2*Z + b3*X*Z + rnorm(n, 0, sd = 10) DATA3 = data.frame(cbind(y, X, Z)) # fit linear model with interaction fit4.lm <- lm(y ~ X*Z, data = DATA3) # constraint syntax myConstraints4 <- ' AVE := X + 16.86137*X.Z; AVE > 0 ' conTest(fit4.lm, constraints = myConstraints4) # or fit4.con <- restriktor(fit4.lm, constraints = ' AVE := X + 16.86137*X.Z; AVE > 0 ') conTest(fit4.con) ```

restriktor documentation built on Feb. 25, 2020, 5:08 p.m.