R/selftest.regGLM.R

In asbio: A Collection of Statistical Tools for Biologists

selftest.regGLM.tck1 <-function(){
  options(guiToolkit="tcltk")
  w <- gwindow(title = "Generalized linear models")
  size(w) <- c(700, 650)
  g <- ggroup(container=w, horizontal=FALSE, use.scrollwindow = TRUE)

  #------------- Question 1 -----------#
 
  gp1 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp1.1 <- ggroup(container = gp1, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("1) ", container = gp1.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("(T or F) Generalized linear models (GLMs) can be used to analyze data from all four basic types of 
experimental designs.",
         container = gp1.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans1 <- c("(a)  TRUE ",
            "(b)  FALSE")

  f1 <- function(h,....){ 
    if(tail(svalue(r1),1) == ans1[1]){
      gmessage(msg="Correct.  GLMs can be used to analyze cases when Y is quantitative or categorical and X(s) is (are) quantitative or categorical (or both).")
    }
    if(tail(svalue(r1),1)== ans1[2]){
      gmessage(msg="Incorrect")
    }
   svalue(r1) <- character(0)
  }
  
  
  r1 <- gcheckboxgroup(ans1, container = gp1, checked = FALSE, where = "beginning", handler = f1)
  
   
  
  #------------- Question 2 -----------#
  
  gp2 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp2.1 <- ggroup(container = gp2, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("2) ", container = gp2.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("General linear models are not a type of GLM.",
               container = gp2.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans2 <- c("(a)  TRUE.",
            "(b)  FALSE.")
  
  f2 <- function(h,....){ 
    if(tail(svalue(r2),1) == ans2[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r2),1)== ans2[2]){
      gmessage(msg="Correct")
    }
   
    svalue(r2) <- character(0)
  }
  
  
  r2 <- gcheckboxgroup(ans2, container = gp2, checked = FALSE, where = "beginning", handler = f2)
  
  
  #------------- Question 3 -----------#
  
  gp3 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp3.1 <- ggroup(container = gp3, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("3) ", container = gp3.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("What distinguishes a GLM?",
               container = gp3.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans3 <- c("(a)  Simultaneous consideration of BIC and AIC in the model.",
            "(b)  Incorporation of model II regression procedures.",
            "(c)  A user-defined link function that allows the expression of the Y mean function in linear terms,\n and a user-defined error term distribution.",
            "(d)  All of the above.")
  
  f3 <- function(h,....){ 
    if(tail(svalue(r3),1) == ans3[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r3),1)== ans3[2]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r3),1)== ans3[3]){
      gmessage(msg="Correct")
    }
    if(tail(svalue(r3),1)== ans3[4]){
      gmessage(msg="Inorrect")
    }
    
    svalue(r3) <- character(0)
  }
  
  
  r3 <- gcheckboxgroup(ans3, container = gp3, checked = FALSE, where = "beginning", handler = f3)
  
  
  
  #------------- Question 4 -----------#
  
  gp4 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp4.1 <- ggroup(container = gp4, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("4) ", container = gp4.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("We would use a logistic GLM...",
               container = gp4.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans4 <- c("(a)  to account for inequality of BIC and AIC in a model.",
            "(b)  to model bivariate forms of Y.",
            "(c)  to find the optimal power transformation of X, to account for nonconstant variance in residuals.",
            "(d)  to model a dichotomous Y as a function of X(s). ",
            "(e)  to find the optimal power transformation of X, to account for nonlinearity in the relationship of Y and X.")
  
  f4 <- function(h,....){ 
    if(tail(svalue(r4),1) == ans4[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r4),1)== ans4[2]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r4),1)== ans4[3]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r4),1)== ans4[4]){
      gmessage(msg="Correct")
    }
    if(tail(svalue(r4),1)== ans4[5]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    svalue(r4) <- character(0)
  }
  
  
  r4 <- gcheckboxgroup(ans4, container = gp4, checked = FALSE, where = "beginning", handler = f4)  
  
 
  #------------- Question 5 -----------#
  
  gp5 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp5.1 <- ggroup(container = gp5, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("5) ", container = gp5.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("The logit of an outcome A is defined as...",
               container = gp5.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans5 <- c("(a)  P(A) - (1 - P(A)).",
            "(b)  1 - P(A).",
            "(c)  P(A)/(1 - P((A)).",
            "(d)  ln(P(A)/(1 - P((A))).")
  
  f5 <- function(h,....){ 
    if(tail(svalue(r5),1) == ans5[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r5),1)== ans5[2]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r5),1)== ans5[3]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r5),1)== ans5[4]){
      gmessage(msg="Correct")
    }
    
    svalue(r5) <- character(0)
  }
  
  
  r5 <- gcheckboxgroup(ans5, container = gp5, checked = FALSE, where = "beginning", handler = f5)  
  
 
 #------------- Question 6 -----------#
  
  gp6 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp6.1 <- ggroup(container = gp6, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("6) ", container = gp6.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("In the context of GLMs, deviance is...",
               container = gp6.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans6 <- c("(a)  a maladjusted human behaviour.",
            "(b)  a generalized form of model error.",
            "(c)  equivelent to SSE for general linear models.",
            "(d)  b) and c) above.")
  
  f6 <- function(h,....){ 
    if(tail(svalue(r6),1) == ans6[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r6),1)== ans6[2]){
      gmessage(msg="Partially correct",  icon = "error")
    }
    if(tail(svalue(r6),1)== ans6[3]){
      gmessage(msg="Partially correct",  icon = "error")
    }
    if(tail(svalue(r6),1)== ans6[4]){
      gmessage(msg="Correct")
    }
    
    svalue(r6) <- character(0)
  }
  
  
  r6 <- gcheckboxgroup(ans6, container = gp6, checked = FALSE, where = "beginning", handler = f6)  
  
  
  
  #------------- Question 7 -----------#
  
  gp7 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp7.1 <- ggroup(container = gp7, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("7) ", container = gp7.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("In a logistic GLM, units for coefficients from the fitted model will be...",
               container = gp7.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  
  ans7 <- c("(a)  odds.",
            "(b)  log odds.",
            "(c)  probability.",
            "(d)  inverse probability.")
  
  f7 <- function(h,....){ 
    if(tail(svalue(r7),1) == ans7[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r7),1)== ans7[2]){
      gmessage(msg="Correct")
    }
    if(tail(svalue(r7),1)== ans7[3]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r7),1)== ans7[4]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    
    svalue(r7) <- character(0)
  }
  
  
  r7 <- gcheckboxgroup(ans7, container = gp7, checked = FALSE, where = "beginning", handler = f7)  
  
  
  
  #------------- Question 8 -----------#

  gp8 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp8.1 <- ggroup(container = gp8, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("8) ", container = gp8.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("In a Poisson GLM, units for coefficients from the fitted model will be...",
               container = gp8.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)


  ans8 <- c("(a)  odds.",
            "(b)  log counts.",
            "(c)  counts.",
            "(d)  inverse counts.")

  f8 <- function(h,....){
    if(tail(svalue(r8),1) == ans8[1]){
      gmessage(msg="Incorrect.",  icon = "error")
    }
    if(tail(svalue(r8),1)== ans8[2]){
      gmessage(msg="Correct.")
    }
    if(tail(svalue(r8),1)== ans8[3]){
      gmessage(msg="Incorrect.",  icon = "error")
    }
    if(tail(svalue(r8),1)== ans8[4]){
      gmessage(msg="Incorrect.",  icon = "error")
    }

    svalue(r8) <- character(0)
  }


  r8 <- gcheckboxgroup(ans8, container = gp8, checked = FALSE, where = "beginning", handler = f8)



  #------------- Question 9 -----------#

  gp9 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp9.1 <- ggroup(container = gp9, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("9) ", container = gp9.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("Overdispersion in GLMs...",
               container = gp9.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)


  ans9 <- c("(a)  occurs when the observed error variance exceeds the model (theoretical) variance.",
            "(b)  can be estimated as the residual deviance divided by the degrees of freedom error for the model.",
            "(c)  indicates the GLM model is inappropriate.",
            "(d)  a) and b) above.",
            "(e)  All of the above.")

  f9 <- function(h,....){
    if(tail(svalue(r9),1) == ans9[1]){
      gmessage(msg="Partially correct",  icon = "error")
    }
    if(tail(svalue(r9),1)== ans9[2]){
      gmessage(msg="Partially correct",  icon = "error")
    }
    if(tail(svalue(r9),1)== ans9[3]){
      gmessage(msg="Partially correct",  icon = "error")
    }
    if(tail(svalue(r9),1)== ans9[4]){
      gmessage(msg="Partially correct",  icon = "error")
    }
    if(tail(svalue(r9),1)== ans9[5]){
      gmessage(msg="Correct")
    }

    svalue(r9) <- character(0)
  }


  r9 <- gcheckboxgroup(ans9, container = gp9, checked = FALSE, where = "beginning", handler = f9)



  #------------- Question 10 -----------#

  gp10 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  gp10.1 <- ggroup(container = gp10, spacing = 2, pos = 0, horizontal = TRUE)
  q <- glabel("10) ", container = gp10.1, horizontal = TRUE)
  font(q) <- list(weight = "bold")
  qq <- glabel("The area under the curve (AUC) from from a receiver operating characteristic (ROC) curve...",
               container = gp10.1, anchor = c(-1,1))
  font(qq) <- list(family = "cambria", size = 11)
  
  ans10 <- c("(a)  provide a measure of parsimony for Poisson GLMs.",
            "(b)  measure deviance in logistic GLMs.",
            "(c)  provide a general measure of the predictive effectiveness of a logistic GLM.",
            "(d)  All of the above.")

  f10 <- function(h,....){
    if(tail(svalue(r10),1) == ans10[1]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r10),1)== ans10[2]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    if(tail(svalue(r10),1)== ans10[3]){
      gmessage(msg="Correct")
    }
    if(tail(svalue(r10),1)== ans10[4]){
      gmessage(msg="Incorrect",  icon = "error")
    }
    svalue(r10) <- character(0)
  }


  r10 <- gcheckboxgroup(ans10, container = gp10, checked = FALSE, where = "beginning", handler = f10)




  # #------------- Question 11 -----------#
  # 
  # gp11 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp11.1 <- ggroup(container = gp11, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("11) ", container = gp11.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("Consider the matrix expression of a general linear model:",
  #              container = gp11.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # qq1 <- glabel("Y = \u03b2X + \u03b5.",
  #               container = gp11.1, anchor = c(-1,1))
  # font(qq1) <- list(family = "cambria", size = 11, weight = "bold")
  # qq2 <- glabel("X",
  #               container = gp11.1, anchor = c(-1,1))
  # font(qq2) <- list(family = "cambria", size = 11, weight = "bold")
  # qq3 <- glabel("has dimension...",
  #               container = gp11.1, anchor = c(-1,1))
  # font(qq3) <- list(family = "cambria", size = 11)
  # 
  # ans11 <- c("(a)  (n x p).",
  #            "(b)  (n x 1).",
  #            "(c)  (p x 1).",
  #            "(d)  (p x n).")
  # 
  # f11 <- function(h,....){ 
  #   if(tail(svalue(r11),1) == ans11[1]){
  #     gmessage(msg="Correct")
  #   }
  #   if(tail(svalue(r11),1)== ans11[2]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r11),1)== ans11[3]){
  #     gmessage(msg="Inorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r11),1)== ans11[4]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   svalue(r11) <- character(0)
  # }
  # 
  # 
  # r11 <- gcheckboxgroup(ans11, container = gp11, checked = FALSE, where = "beginning", handler = f11)  
  # 
  # 
  # 
  # #------------- Question 12 -----------#
  # 
  # gp12 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp12.1 <- ggroup(container = gp12, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("12) ", container = gp12.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("The coefficient of determinination, R\u00B2...",
  #              container = gp12.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # 
  # 
  # ans12 <- c("(a)  will be in the range [0, 1].",
  #            "(b)  is calculated as 1 - SSE/SSTO.",
  #            "(c)  represents the proportion of variability in Y explained by the model.",
  #            "(d)  is equivalent to r\u00B2 for a simple linear regression model.",
  #            "(e)  (a) and (c) above.",
  #            "(f)  All of the above.")
  # 
  # f12 <- function(h,....){ 
  #   if(tail(svalue(r12),1) == ans12[1]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r12),1)== ans12[2]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r12),1)== ans12[3]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r12),1)== ans12[4]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r12),1)== ans12[5]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r12),1)== ans12[6]){
  #     gmessage(msg="Correct")
  #   }
  #   svalue(r12) <- character(0)
  # }
  # 
  # 
  # r12 <- gcheckboxgroup(ans12, container = gp12, checked = FALSE, where = "beginning", handler = f12)  
  # 
  # 
  # #------------- Question 13 -----------#
  # 
  # gp13 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp13.1 <- ggroup(container = gp13, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("13) ", container = gp13.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("The adjusted R\u00B2...",
  #              container = gp13.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # 
  # 
  # ans13 <- c("(a)  will be identical to R\u00B2 in most situations.",
  #            "(b)  will occasionally be larger than R\u00B2.",
  #            "(c)  adjusts for the number of parameters in the model.",
  #            "(d)  is equivalent to r\u00B2 for a simple linear regression model.",
  #            "(e)  (a) and (c) above.",
  #            "(f)  All of the above.")
  # 
  # f13 <- function(h,....){ 
  #   if(tail(svalue(r13),1) == ans13[1]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r13),1)== ans13[2]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r13),1)== ans13[3]){
  #     gmessage(msg="Correct")
  #   }
  #   if(tail(svalue(r13),1)== ans13[4]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r13),1)== ans13[5]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r13),1)== ans13[6]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   svalue(r13) <- character(0)
  # }
  # 
  # 
  # r13 <- gcheckboxgroup(ans13, container = gp13, checked = FALSE, where = "beginning", handler = f13)  
  # 
  # 
  # #------------- Question 14 -----------#
  # 
  # gp14 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp14.1 <- ggroup(container = gp14, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("14) ", container = gp14.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("The partial R\u00B2...",
  #              container = gp14.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # 
  # 
  # ans14 <- c("(a)  measures the reduction in SSE after a variable of interest is added to model.",
  #            "(b)  will occasionally be larger than R\u00B2.",
  #            "(c)  requires the specification of a nested model (that lacks a variable of interest) and a reference (full) model.",
  #            "(d)  is equivalent to r\u00B2 for a simple linear regression model.",
  #            "(e)  (a) and (c) above.",
  #            "(f)  All of the above.")
  # 
  # f14 <- function(h,....){ 
  #   if(tail(svalue(r14),1) == ans14[1]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r14),1)== ans14[2]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r14),1)== ans14[3]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r14),1)== ans14[4]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r14),1)== ans14[5]){
  #     gmessage(msg="Correct")
  #   }
  #   if(tail(svalue(r14),1)== ans14[6]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   svalue(r14) <- character(0)
  # }
  # 
  # 
  # r14 <- gcheckboxgroup(ans14, container = gp14, checked = FALSE, where = "beginning", handler = f14)  
  # 
  # 
  # 
  # 
  # #------------- Question 15 -----------#
  # 
  # gp15 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp15.1 <- ggroup(container = gp15, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("15) ", container = gp15.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("An ANOVA approach for assessing the affect of individual predictors in multiple regression\n is problematic because...",
  #              container = gp15.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # 
  # 
  # ans15 <- c("(a)  Sums of squares and p-values will vary depending on the order predictors are specified in the linear model.",
  #            "(b)  Linear model algorithms typically use type I sums of squares which will be untrustworthy for assessing\n the affect of individual predictors.",
  #            "(c)  Type II sums of squares, which will provide consistent results, require the specification of a nested model\n (that lacks a variable of interest) and a reference (full) model.",
  #            "(d)  All of the above.")
  # 
  # f15 <- function(h,....){ 
  #   if(tail(svalue(r15),1) == ans15[1]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r15),1)== ans15[2]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r15),1)== ans15[3]){
  #     gmessage(msg="Partially correct",  icon = "error")
  #   }
  #   if(tail(svalue(r15),1)== ans15[4]){
  #     gmessage(msg="Correct")
  #   }
  #     svalue(r15) <- character(0)
  # }
  # 
  # 
  # r15 <- gcheckboxgroup(ans15, container = gp15, checked = FALSE, where = "beginning", handler = f15)  
  # 
  # 
  # 
  # #------------- Question 16 -----------#
  # 
  # gp16 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp16.1 <- ggroup(container = gp16, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("16) ", container = gp16.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("In multiple regression we can assess the null hypothesis: H\u2080: \u03B2\u2081 = \u03B2\u2082 = , ... , = \u03B2\u209A\u02CD\u2081.  \nWhat is the null distribution for this test?",
  #              container = gp16.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # 
  # 
  # ans16 <- c("(a)  N(0, 1).",
  #            "(b)  t(n - p).",
  #            "(c)  t(p - n).",
  #            "(d)  F(p-1, n-p).",
  #            "(e)  F(n-p, p-1).")
  # 
  # f16 <- function(h,....){ 
  #   if(tail(svalue(r16),1) == ans16[1]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r16),1)== ans16[2]){
  #     gmessage(msg="Incorrect.  This is the null distribution for considering the effect of individual predictors.",  icon = "error")
  #   }
  #   if(tail(svalue(r16),1)== ans16[3]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r16),1)== ans16[4]){
  #     gmessage(msg="Correct")
  #   }
  #   if(tail(svalue(r16),1)== ans16[5]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   svalue(r16) <- character(0)
  # }
  # 
  # 
  # r16 <- gcheckboxgroup(ans16, container = gp16, checked = FALSE, where = "beginning", handler = f16)  
  # 
  # 
  # 
  # #------------- Question 17 -----------#
  # 
  # gp17 <- gframe(container = g, spacing = 2, pos = 0, horizontal = FALSE)
  # gp17.1 <- ggroup(container = gp17, spacing = 2, pos = 0, horizontal = TRUE)
  # q <- glabel("17) ", container = gp17.1, horizontal = TRUE)
  # font(q) <- list(weight = "bold")
  # qq <- glabel("Why are predicted values a concern in multiple regression?",
  #              container = gp17.1, anchor = c(-1,1))
  # font(qq) <- list(family = "cambria", size = 11)
  # 
  # 
  # ans17 <- c("(a)  It is difficult to determine if new X outcomes from multiple predictors constitute extrapolated values.",
  #            "(b)  It is difficult to predict Y from multiple X variables.",
  #            "(c)  It is difficult to predict multiple Ys from a single X variable.",
  #            "(d)  The null distribution will not exist.")
  # 
  # f17 <- function(h,....){ 
  #   if(tail(svalue(r17),1) == ans17[1]){
  #     gmessage(msg="Correct")
  #   }
  #   if(tail(svalue(r17),1)== ans17[2]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r17),1)== ans17[3]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   if(tail(svalue(r17),1)== ans17[4]){
  #     gmessage(msg="Incorrect",  icon = "error")
  #   }
  #   svalue(r17) <- character(0)
  # }
  # 
  # 
  # r17 <- gcheckboxgroup(ans17, container = gp17, checked = FALSE, where = "beginning", handler = f17)  
  # 
  # 
  # 
  # 
  # 
  # 
  # 
  # 
  # 
   
}