update

html/04-Auxillary_parameter_recovery.html

@@ -456,32 +456,26 @@ <h3>Run the simulation</h3>
 <pre class="r"><code>res &lt;- runSimulation(Design, replications = 5, parallel = FALSE, packages = &#39;mirt&#39;,
                      generate=Generate, analyse=Analyse, summarise=Summarise)</code></pre>
 <pre><code>## 
-## 
-Design: 1/6;   RAM Used: 190 Mb;   Replications: 5;   Total Time: 0.00s 
+## Design: 1/6;   Replications: 5;   RAM Used: 192.2 Mb;   Total Time: 0.00s 
 ##  Conditions: test_length=10, distribution=normal
 ## 
-## 
-Design: 2/6;   RAM Used: 196.4 Mb;   Replications: 5;   Total Time: 3.64s 
+## Design: 2/6;   Replications: 5;   RAM Used: 198.8 Mb;   Total Time: 3.63s 
 ##  Conditions: test_length=20, distribution=normal
 ## 
-## 
-Design: 3/6;   RAM Used: 196.5 Mb;   Replications: 5;   Total Time: 8.86s 
+## Design: 3/6;   Replications: 5;   RAM Used: 198.9 Mb;   Total Time: 8.97s 
 ##  Conditions: test_length=30, distribution=normal
 ## 
-## 
-Design: 4/6;   RAM Used: 196.5 Mb;   Replications: 5;   Total Time: 14.89s 
+## Design: 4/6;   Replications: 5;   RAM Used: 199 Mb;   Total Time: 15.47s 
 ##  Conditions: test_length=10, distribution=bimodal
 ## 
-## 
-Design: 5/6;   RAM Used: 196.6 Mb;   Replications: 5;   Total Time: 17.97s 
+## Design: 5/6;   Replications: 5;   RAM Used: 199 Mb;   Total Time: 18.63s 
 ##  Conditions: test_length=20, distribution=bimodal
 ## 
-## 
-Design: 6/6;   RAM Used: 196.7 Mb;   Replications: 5;   Total Time: 23.18s 
+## Design: 6/6;   Replications: 5;   RAM Used: 199.1 Mb;   Total Time: 24.24s 
 ##  Conditions: test_length=30, distribution=bimodal
 ## </code></pre>
 <pre><code>## 
-## Simulation complete. Total execution time: 29.49s</code></pre>
+## Simulation complete. Total execution time: 31.09s</code></pre>
 <pre class="r"><code>res</code></pre>
 <pre><code>## # A tibble: 6 × 15
 ##   test_length distribution   bias_EAP   bias_MAP    bias_ML bias_EAP_bimodal

html/07-Coverage.html

@@ -412,15 +412,15 @@ <h3>Define the functions</h3>
 <pre><code>## # A tibble: 9 × 8
 ##      mu sample_size   t_CI   z_CI REPLICATIONS SIM_TIME       SEED COMPLETED    
 ##   &lt;dbl&gt;       &lt;dbl&gt;  &lt;dbl&gt;  &lt;dbl&gt;        &lt;dbl&gt; &lt;chr&gt;         &lt;int&gt; &lt;chr&gt;        
-## 1     0          10 0.9531 0.9221        10000 0.65s     488349520 Thu Aug  1 1…
-## 2     5          10 0.9466 0.9169        10000 0.45s     469344150 Thu Aug  1 1…
-## 3    10          10 0.9503 0.9173        10000 0.52s    1550099436 Thu Aug  1 1…
-## 4     0          20 0.9481 0.9325        10000 0.44s      40778633 Thu Aug  1 1…
-## 5     5          20 0.9527 0.9373        10000 0.48s     713327526 Thu Aug  1 1…
-## 6    10          20 0.9523 0.9375        10000 0.46s     831490948 Thu Aug  1 1…
-## 7     0          30 0.9494 0.9408        10000 0.45s    1214377063 Thu Aug  1 1…
-## 8     5          30 0.9476 0.9381        10000 0.44s    1255528026 Thu Aug  1 1…
-## 9    10          30 0.9488 0.9392        10000 0.39s    1229276239 Thu Aug  1 1…</code></pre>
+## 1     0          10 0.9531 0.9221        10000 0.52s     488349520 Thu Dec 12 2…
+## 2     5          10 0.9466 0.9169        10000 0.48s     469344150 Thu Dec 12 2…
+## 3    10          10 0.9503 0.9173        10000 0.53s    1550099436 Thu Dec 12 2…
+## 4     0          20 0.9481 0.9325        10000 0.41s      40778633 Thu Dec 12 2…
+## 5     5          20 0.9527 0.9373        10000 0.46s     713327526 Thu Dec 12 2…
+## 6    10          20 0.9523 0.9375        10000 0.36s     831490948 Thu Dec 12 2…
+## 7     0          30 0.9494 0.9408        10000 0.40s    1214377063 Thu Dec 12 2…
+## 8     5          30 0.9476 0.9381        10000 0.43s    1255528026 Thu Dec 12 2…
+## 9    10          30 0.9488 0.9392        10000 0.40s    1229276239 Thu Dec 12 2…</code></pre>
 <p>From these results it’s clear to see that the t-distribution gives very good coverage rates across all conditions. However, the z-distribution CI is too liberal in the smaller sample sizes, and gives an interval that does not contain the population mean as often as desired. Around a sample size of 30 we can see that the two distributions behave reasonably well.</p>
 <p>Note that the <code>ECR()</code> function was used within the <code>Analyse()</code> function rather than in <code>Summarise()</code>. When only the three values are supplied (CIs + population value) then <code>ECR()</code> will return either a 1 if the population value was within the CI or 0 if it was not. Hence, in the <code>Summarise()</code> function all that is required to obtain the coverage estimate is to find the mean of these 1’s and 0’s. The next section demonstrates how to use <code>ECR()</code> in the <code>Summarise()</code> instead, and it’s really a matter of preference which style you prefer.</p>
 </div>