updates

Author: msneubauer

_sources/lectures/ModelSelection.ipynb

@@ -536,6 +536,8 @@
     "\n",
     " - larger than 10 is \"strong evidence\" in favor of M1.\n",
     "\n",
+    " - larger (smaller) than one favors M1 over M2 (M2 over M1)\n",
+    "\n",
     " - smaller than 0.1 is \"strong evidence\" in favor of M2.\n",
     "\n",
     " - smaller than 0.01 is \"decisive evidence\" in favor of M2."
@@ -1102,18 +1104,18 @@
    "metadata": {},
    "source": [
     "To summarize in words:\n",
-    " - M1 is \"strongly\" (but not \"decisively\") supported by `Da`.\n",
+    " - Odds ratio = 15.8 $\\to$ M1 is \"strongly\" (but not \"decisively\") supported by `Da`.\n",
     "\n",
-    " - M2 is \"decisively\" supported by `Db`\n",
+    " - Odds ratio = 6.3 $\\times 10^{-19}$ $\\to$ M2 is \"decisively\" supported by `Db`\n",
     "\n",
-    " - M2 is mildy supported by `Dc`, but the evidence is not \"strong\"."
+    " - Odds ratio = 0.2 $\\to$ M2 is mildy supported by `Dc`, but the evidence is not \"strong\"."
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "If we have a prior bias that M1 is 10x more likely than M1, this would mostly affect our assessment for `Dc`, which now slightly prefers M1:"
+    "If we have a prior bias that M1 is 10x more likely than M1, this would mostly affect our assessment for `Dc`, which now slightly prefers M1 since Odds ratio 0.2 $\\to$ 2.0:"
    ]
   },
   {