Skip to content

Instantly share code, notes, and snippets.

Show Gist options
  • Save AustinRochford/8165834 to your computer and use it in GitHub Desktop.
Save AustinRochford/8165834 to your computer and use it in GitHub Desktop.
Blog post draft
Display the source blob
Display the rendered blob
Raw
{
"metadata": {
"name": "The Importance of Sequential Testing"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": "As the world becomes ever more data-driven, the basic theory of hypothesis testing is being used by more people and in more contexts than ever before. This epansion has, however, come with a cost. The dominant Neyman-Pearson hypothesis testing framework is subtle and easy to unknowingly misuse. In this post, we'll explore the common scenario where we would like to monitor the status of an ongoing experiment and stop the experiment early if an effect becomes apparent.\n\nThere are many situations in which it is advantageous to monitor the status of an experiment and terminate it early if the conclusion seems apparent. In business, experiments cost money, both in terms of the actual cost of data collection and in terms of the opportunity cost of waiting for an experiment to reach a set number of samples before acting on its outcome, which may have been apparent much earlier. In medicine, it may be unethical to continue an experimental treatment which appears to have a detrimental effect, or to deny the obviously better experimental treatment to the control group until the predetermined sample size is reached.\n\nWhile these reasons for continuous monitoring and early termination of certain experiments are quite compelling, if this method is applied naively, it can lead to wildly incorrect analyses. Below, we illustrate the perils of the naive approach to sequential testing (as this sort of procedure is known) and show how to perform a correct analysis of a fairly simple, yet illustrative introductory sequential experiment.\n\nA brief historical digression may be informative. The frequentist approach to hypothesis testing was pioneered just after the turn of the 20th century in England in order to analyze agricultural experiments. According to Armitage (in [1]), one of the pioneers of sequential experiment design,\n\n> [t]he classical theory of experimental design deals predominantly with experiments of predetermined size, presumably because the pioneers of the subject, particularly R. A. Fisher, worked in agricultural research, where the outcome of a field trial is not available until long after the experiment has been designed and started. It is interesting to speculate how differently statistical theory might have evolved if Fisher had been employed in medical or industrial research.\n\nIn this post, we will analyze the following hypothesis test. Assume the data are drawn from the normal distribution, $N(\\theta, 1)$ with unknown mean $\\theta$ and known variance $\\sigma^2 = 1$. We wish to test the simple null hypothesis that $\\theta = 0$ versus the simple alternative hypothesis that $\\theta = 1$ at the $\\alpha = 0.05$ level. By the [Neyman-Pearson lemma](http://en.wikipedia.org/wiki/Neyman%E2%80%93Pearson_lemma), the most powerful test of this hypothesis rejects the null hypothesis when the [likelihood ratio](http://en.wikipedia.org/wiki/Likelihood-ratio_test) exceeds some critical value. In our normal case, it is well-known that this criterion is equivalent to $\\sum_{i = 1}^n X_i > \\sqrt{n} z_{0.05}$ where $z_{0.05} \\approx 1.64$ is the $1 - 0.05 = 0.95$ quantile of the standard normal distribution.\n\nTo model the ongoing monitoring of this experiment, we define a random variable $N = \\min \\{n \\geq 1 | \\sum_{i = 1}^n X_i > \\sqrt{n} z_{0.05}\\}$, called the [stopping time](http://en.wikipedia.org/wiki/Stopping_time). The random variable $N$ is the first time that the test statistic exceeds the critical value, and the naive approach to sequential testing would reject the null hypothesis after $N$ samples (when $N < \\infty$, of course). At first, this procedure may seem reasonable, because when the alternative hypothesis that $\\theta = 1$ is true, $N < \\infty$ [almost surely](http://en.wikipedia.org/wiki/Almost_surely) by the [strong law of large numbers](http://en.wikipedia.org/wiki/Law_of_large_numbers#Strong_law). The first inkling that something is amiss with this procedure is the surprising fact that it is also true that $N < \\infty$ almost surely under the null hypothesis that $\\theta = 0$. (See [2] for a proof.) More informally, if we sample long enough, this procedure will almost always reject the null hypothesis, even when it is true.\n\nTo illustrate this problem more quantitatively, consider the following simulation. Suppose we decide ahead of time that we will stop the experiment when the critical value for the current sample size is exceeded, in which case we reject the null hypothesis, or we have collected one thousand samples without exceeding the critical value at any point, in which case we accept the null hypothesis.\n\nHere we use a [Monte Carlo method](http://en.wikipedia.org/wiki/Monte_Carlo_method) to approximate the level of this naive sequential test. First we generate ten thousand simulations of such an experiment, assuming the null hypothesis that the data are $N(0, 1)$ is true."
},
{
"cell_type": "code",
"collapsed": false,
"input": "from __future__ import division\n\nimport numpy as np\nfrom scipy import stats\n\nn = 100\nN = 10000\n\nsamples = stats.norm.rvs(size=(N, n))",
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 1
},
{
"cell_type": "markdown",
"metadata": {},
"source": "Now we calculate the proportion of these simulated experiments that would have been stopped before one thousand samples, incorrectly rejecting the null hypothesis."
},
{
"cell_type": "code",
"collapsed": false,
"input": "alpha = 0.05\nz_alpha = stats.norm.isf(alpha)\n\ncumsums = samples.cumsum(axis=1)\nns = np.arange(1, n + 1)\n\nnp.any(cumsums > np.sqrt(ns) * z_alpha, axis=1).sum() / N",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 2,
"text": "0.30909999999999999"
}
],
"prompt_number": 2
},
{
"cell_type": "markdown",
"metadata": {},
"source": "We see that the actual level of this test is an order of magnitude larger than the desired level of $\\alpha = 0.05$. To check that our method is reasonable, we see that if we always collect one thousand samples, we achieve a simulated level quite close to $\\alpha = 0.05$."
},
{
"cell_type": "code",
"collapsed": false,
"input": "(cumsums[:, -1] > np.sqrt(n) * z_alpha).sum() / N",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 3,
"text": "0.051700000000000003"
}
],
"prompt_number": 3
},
{
"cell_type": "markdown",
"metadata": {},
"source": "This simulation is compelling evidence that the naive approach to sequential testing is not correct.\n\nFortunately, the basic framework for the correct analysis of sequential experiments was worked out during World War II to more efficiently test lots of ammunition (among other applications). In 1947, Wald introduced the [sequential probability ratio test](http://en.wikipedia.org/wiki/Sequential_probability_ratio_test) (SPRT), which produces a correct analysis of the experiment we have been considering.\n\nLet\n\n$$\n\\begin{align*}\n\\Lambda (x_1, \\ldots, x_n)\n & = \\frac{L(1; x_1, \\ldots, x_n)}{L(0; x_1, \\ldots, x_n)}\n\\end{align*}\n$$\n\nbe the likelihood ratio corresponding to our two hypotheses. The SPRT uses two thresholds, $0 < a < 1 < b$, and continues sampling whenever $a < \\Lambda (x_1, \\ldots, x_n) < b$. When $\\Lambda (x_1, \\ldots, x_n) \\leq a$, we accept the null hypothesis, and when $b \\leq \\Lambda (x_1, \\ldots, x_n)$, we reject the null hypothesis. We choose $A$ and $B$ by fixing the approximate level of the test, $\\alpha$, and the approximate power of the test, $1 - \\beta$. With these quantities chosen, we use\n\n$$\n\\begin{align*}\na\n & = \\frac{\\beta}{1 - \\alpha}, \\textrm{and} \\\\\nb\n & = \\frac{1 - \\beta}{\\alpha}.\n\\end{align*}\n$$\n\nFor our hypothesis test $\\alpha = 0.05$. The power of the naive test after $n$ samples is"
},
{
"cell_type": "code",
"collapsed": false,
"input": "power = 1 - stats.norm.sf(z_alpha - 1 / np.sqrt(n))\nbeta = 1 - power\n\npower",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 4,
"text": "0.93880916378665569"
}
],
"prompt_number": 4
},
{
"cell_type": "markdown",
"metadata": {},
"source": "Which gives the following values for $a$ and $b$:"
},
{
"cell_type": "code",
"collapsed": false,
"input": "a = beta / (1 - alpha)\nb = (1 - beta) / alpha\n\na, b",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 5,
"text": "(0.06441140654036244, 18.776183275733114)"
}
],
"prompt_number": 5
},
{
"cell_type": "markdown",
"metadata": {},
"source": "For our example, it will be benificial to rewrite the SPRT in terms of the log-likelihood ratio,\n\n$$\n\\begin{align*}\n\\log a\n & < \\log \\Lambda (x_1, \\ldots, x_n)\n < \\log b.\n\\end{align*}\n$$\n\nIt is easy to show that $\\log \\Lambda (x_1, \\ldots, x_n) = \\frac{n}{2} - \\sum_{i = 1}^n X_i$, so the SPRT in this case reduces to\n\n$$\n\\begin{align*}\n\\frac{n}{2} - \\log b\n & < \\sum_{i = 1}^n X_i\n < \\frac{n}{2} - \\log a.\n\\end{align*}\n$$\n\nThe logarithms of $a$ and $b$ are"
},
{
"cell_type": "code",
"collapsed": false,
"input": "np.log((a, b))",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 6,
"text": "array([-2.74246454, 2.93258922])"
}
],
"prompt_number": 6
},
{
"cell_type": "markdown",
"metadata": {},
"source": "We verify that this test is indeed of approximate level $\\alpha = 0.05$ using the simulations from our previous Monte Carlo analysis."
},
{
"cell_type": "code",
"collapsed": false,
"input": "np.any(cumsums >= ns / 2 - np.log(a), axis=1).sum() / N",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 7,
"text": "0.036299999999999999"
}
],
"prompt_number": 7
},
{
"cell_type": "markdown",
"metadata": {},
"source": "The following plot shows the rejection boundaries for both the naive sequential test and the SPRT along with the density of our Monte Carlo samples."
},
{
"cell_type": "code",
"collapsed": false,
"input": "from matplotlib import pyplot as plt\n\nfig = plt.figure()\nax = fig.add_subplot(111)\n\nplot_ns = np.repeat(ns.reshape((1, n)), N, axis=0).reshape(N * n)\nbins = ax.hexbin(plot_ns, cumsums.reshape(N * n), bins='log', cmap=plt.cm.binary);\n\nax.plot(ns, np.sqrt(ns) * z_alpha, color='b', label='Naive test');\nax.plot(ns, ns / 2 - np.log(a), color='r', label='SPRT');\n\nax.legend(loc=2);\nax.set_xlabel('Number of samples, $n$');\nax.set_ylabel('Test statistic, $\\sum_{i = 1}^n X_i$');\nax.set_title('Rejection boundaries and Monte Carlo samples');\n\nbar = fig.colorbar(bins)\nbar.set_label('log(Density)');",
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAZQAAAEbCAYAAAD9I3KtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noQiCIAiOIApFEARBcIT/AUvANn3RWz8tAAAAAElFTkSuQmCC\n",
"text": "<matplotlib.figure.Figure at 0x38c1450>"
}
],
"prompt_number": 8
},
{
"cell_type": "markdown",
"metadata": {},
"source": "From this diagram we can easily see that a significant number of samples fall above the rejection boundary for the naive test (the blue curve) but below the rejection boundary for the SPRT (the red line). These samples explain the large discrepancy between the desired level of the test ($\\alpha = 0.05$) and the simulated level of the naive test. We also see that very few samples exceed the rejection boundary for the SPRT, leading it to have level smaller than $\\alpha = 0.05$.\n\nIt is important to note that we have barely scratched the surface of the vast field of sequential experiment design and analysis. We have not even attempted to give more than a cursory introduction to the SPRT, one of the most basic ideas in this field. One property of this test that bears mentioning is its optimality, in the following sense. Just as the Neyman-Pearson lemma shows that the likelihood ratio test is the most powerful test of a simple hypothesis at a fixed level, the Wald-Wolfowitz theorem shows that the SPRT is the sequential test that minimizes the expected stopping time under both the null and alternative hypotheses for a fixed level and power. For more details on this theorem, and the general theory of sequential experiments, consult [2]."
},
{
"cell_type": "markdown",
"metadata": {},
"source": "[1]: Armitage, P. (1993). Interim analyses in clinical trials. _In Multiple Comparisons, Selection, and Applications in Biometry_, (Ed., F.M. Hoppe), New York: Marcel Dekker, 391\u2013402.\n\n[2]: Bartroff, Jay; Lai, Tze Leung; Shih, Mei-Chiung, _Sequential experimentation in clinical trials. \nDesign and analysis._ Springer Series in Statistics. Springer, New York, 2013. \n\n[3]: Keener, Robert W., _Theoretical statistics, Topics for a core course._ Springer Texts in Statistics. Springer, New York, 2010."
},
{
"cell_type": "code",
"collapsed": false,
"input": "",
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 8
}
],
"metadata": {}
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment