diff --git a/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC2.ipynb b/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC2.ipynb
index 3bc78f7f..9c2212bf 100644
--- a/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC2.ipynb
+++ b/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC2.ipynb
@@ -2172,7 +2172,7 @@
     "\n",
     "We will be doing this graphically as well, which may seem like an even less objective method. The alternative is to use *Bayesian p-values*. These are still subjective, as the proper cutoff between good and bad is arbitrary. Gelman emphasises that the graphical tests are more illuminating [7] than p-value tests. We agree.\n",
     "\n",
-    "The following graphical test is a novel data-viz approach to logistic regression. The plots are called *separation plots*[8]. For a suite of models we wish to compare, each model is plotted on an individual separation plot. I leave most of the technical details about separation plots to the very accessible [original paper](http://mdwardlab.com/sites/default/files/GreenhillWardSacks.pdf), but I'll summarize their use here.\n",
+    "The following graphical test is a novel data-viz approach to logistic regression. The plots are called *separation plots*[8]. For a suite of models we wish to compare, each model is plotted on an individual separation plot. I leave most of the technical details about separation plots to the very accessible [original paper](https://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2011.00525.x), but I'll summarize their use here.\n",
     "\n",
     "For each model, we calculate the proportion of times the posterior simulation proposed a value of 1 for a particular temperature, i.e. compute $P( \\;\\text{Defect} = 1 | t, \\alpha, \\beta )$ by averaging. This gives us the posterior probability of a defect at each data point in our dataset. For example, for the model we used above:"
    ]
diff --git a/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC3.ipynb b/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC3.ipynb
index 3eae5093..431b43de 100644
--- a/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC3.ipynb
+++ b/Chapter2_MorePyMC/Ch2_MorePyMC_PyMC3.ipynb
@@ -2236,7 +2236,7 @@
     "\n",
     "We will be doing this graphically as well, which may seem like an even less objective method. The alternative is to use *Bayesian p-values*. These are still subjective, as the proper cutoff between good and bad is arbitrary. Gelman emphasises that the graphical tests are more illuminating [7] than p-value tests. We agree.\n",
     "\n",
-    "The following graphical test is a novel data-viz approach to logistic regression. The plots are called *separation plots*[8]. For a suite of models we wish to compare, each model is plotted on an individual separation plot. I leave most of the technical details about separation plots to the very accessible [original paper](http://mdwardlab.com/sites/default/files/GreenhillWardSacks.pdf), but I'll summarize their use here.\n",
+    "The following graphical test is a novel data-viz approach to logistic regression. The plots are called *separation plots*[8]. For a suite of models we wish to compare, each model is plotted on an individual separation plot. I leave most of the technical details about separation plots to the very accessible [original paper](https://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2011.00525.x), but I'll summarize their use here.\n",
     "\n",
     "For each model, we calculate the proportion of times the posterior simulation proposed a value of 1 for a particular temperature, i.e. compute $P( \\;\\text{Defect} = 1 | t, \\alpha, \\beta )$ by averaging. This gives us the posterior probability of a defect at each data point in our dataset. For example, for the model we used above:"
    ]
diff --git a/Chapter2_MorePyMC/Ch2_MorePyMC_TFP.ipynb b/Chapter2_MorePyMC/Ch2_MorePyMC_TFP.ipynb
index 6f04a856..1e8752ee 100644
--- a/Chapter2_MorePyMC/Ch2_MorePyMC_TFP.ipynb
+++ b/Chapter2_MorePyMC/Ch2_MorePyMC_TFP.ipynb
@@ -3884,7 +3884,7 @@
         "\n",
         "We will be doing this graphically as well, which may seem like an even less objective method. The alternative is to use *Bayesian p-values*. These are still subjective, as the proper cutoff between good and bad is arbitrary. Gelman emphasises that the graphical tests are more illuminating [3] than p-value tests. We agree.\n",
         "\n",
-        "The following graphical test is a novel data-viz approach to logistic regression. The plots are called *separation plots*[4]. For a suite of models we wish to compare, each model is plotted on an individual separation plot. I leave most of the technical details about separation plots to the very accessible [original paper](http://mdwardlab.com/sites/default/files/GreenhillWardSacks.pdf), but I'll summarize their use here.\n",
+        "The following graphical test is a novel data-viz approach to logistic regression. The plots are called *separation plots*[4]. For a suite of models we wish to compare, each model is plotted on an individual separation plot. I leave most of the technical details about separation plots to the very accessible [original paper](https://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2011.00525.x), but I'll summarize their use here.\n",
         "\n",
         "For each model, we calculate the proportion of times the posterior simulation proposed a value of 1 for a particular temperature, i.e. compute $P( \\;\\text{Defect} = 1 | t, \\alpha, \\beta )$ by averaging. This gives us the posterior probability of a defect at each data point in our dataset. For example, for the model we used above:"
       ]
@@ -4028,7 +4028,7 @@
         "def separation_plot( p, y, **kwargs ):\n",
         "    \"\"\"\n",
         "    This function creates a separation plot for logistic and probit classification. \n",
-        "    See http://mdwardlab.com/sites/default/files/GreenhillWardSacks.pdf\n",
+        "    See https://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2011.00525.x\n",
         "    \n",
         "    p: The proportions/probabilities, can be a nxM matrix which represents M models.\n",
         "    y: the 0-1 response variables.\n",
diff --git a/Chapter2_MorePyMC/separation_plot.py b/Chapter2_MorePyMC/separation_plot.py
index a5316bed..15433149 100644
--- a/Chapter2_MorePyMC/separation_plot.py
+++ b/Chapter2_MorePyMC/separation_plot.py
@@ -1,6 +1,6 @@
 # separation plot
 # Author: Cameron Davidson-Pilon,2013
-# see http://mdwardlab.com/sites/default/files/GreenhillWardSacks.pdf
+# see https://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2011.00525.x
 
 
 import matplotlib.pyplot as plt
@@ -11,7 +11,7 @@
 def separation_plot( p, y, **kwargs ):
     """
     This function creates a separation plot for logistic and probit classification. 
-    See http://mdwardlab.com/sites/default/files/GreenhillWardSacks.pdf
+    See https://onlinelibrary.wiley.com/doi/10.1111/j.1540-5907.2011.00525.x
     
     p: The proportions/probabilities, can be a nxM matrix which represents M models.
     y: the 0-1 response variables.