Subjective Probability Interval Estimate

Subjective Probability Interval Estimate (also known as SPIES) is a decision making method shown to produce more precise estimates within a desired probablity threshold.

Step 1

Enter the name of the estimates value:

Configure your range's minimum and maximum possible values, and select the size of the bins to which we will divide this range.

Minimum Maximum Bin Size

SPIES stands for Subjective Probability Interval EStimates. It is a forecast elicitation method developed by Uriel Haran (Ben-Gurion University of the Negev), Don Moore (University of California, Berkeley) and Carey Morewedge (Carnegie Mellon University). This method was designed to help forecasters use their meta-knowledge in order to make more accurate and more informative forecasts.

The foundation of the SPIES method consists of four principles: 1) Forecasting should be as simple as possible; 2) The forecaster should take into account all possible outcomes; 3) Likelihood judgments should be generated by the forecaster rather than built into the question; 4) The likelihoods of the possible outcomes should be weighed relative to one another. Based on these principles, the SPIES method's graphical interface presents the full range of possible outcomes, divided into a number of bins. The forecaster estimates the likelihood that the true outcome will fall inside each of these bins, using points. The more probable the forecaster thinks a bin is to include the true outcome, the more points she assigns it.

After rating the bins, SPIES produces a confidence interval estimate. By combining the estimated probabilities of the bins, SPIES can help the decision maker construct any confidence interval he or she may find useful. The decision maker can adjust the interval's width or confidence level to receive exactly the information he or she needs. All of this can be done without asking the forecaster to redo the forecast.

SPIES forecasts outperform intervals produced in the traditional method in a wide range of estimate types and contexts. The SPIES method produces more inclusive intervals, which reduce the human tendency toward overprecision in judgment. Consequently, the intervals produced by SPIES are better calibrated than confidence intervals produced by other methods.

Subjective Probability Interval Estimates (also known as SPIES) consists of three steps:

In Step 1 you define the estimate setting. You will indicate the object of the estimate, and determine the boundaries of the range of possible values. This range will be divided equally into a number of bins. You will determine how big these bins should be.

Step 1

Enter the name of the estimates value:

Configure your range's minimum and maximum possible values, and select the size of the bins to which we will divide this range.

Minimum Maximum Bin Size

In Step 2, you will rate the relative likelihood of each bin to include the correct outcome. You will do this by assigning points to each bin. The more point you assign a bin, the more likely you think it is that the actual outcome will fall within this bin. Of course, you can assign a bin zero points if you think there is no possibility that this bin will include the actual outcome.

Step 2

Set the sliders below to assign points to each bin. The higher the likelihood that the actual value will fall within a bin, the more points you should assign it.

In Step 3, you will define the confidence interval's desired width or confidence level. Increasing confidence widens the interval, such that it includes more possible values. Making the interval more focused, or precise, results in a lower confidence level.

Step 3

Your confidence interval for 95% is 36 - 54
Confidence
95%
Width
18