Our »Streuspanne Lexicon« is not a regular episode, but a section to refer to small explanations without having to go on and on in a separate episode. Here you will find terms such as »the estimate« or »the estimator« explained in less than five minutes.
What is the mathematical definition of an estimate? And how can I even define something so »imprecise«?
In everyday life, estimation often means clever guessing, but in statistics »the estimator« is a fixed expression and rule, or a calculation rule of its own.
An example: We flip a coin to determine the unknown probability of heads, pretending we don't know the probability because the coin could be manipulated.
Then we could simply divide the number of observed heads by the number of tosses and we have an estimate for this unknown probability. We are estimating here on the basis of data. Chance is always involved in one way or another when collecting the data, but after that there is no more variation. This is a brief summary of an estimate in statistics.
Theoretically, any number of things can go wrong with an estimate and the estimated value can therefore deviate greatly from the true, unknown value.
First strategy: The sample size is increased. The unknown probability of a coin can be determined more robustly with 100 tosses than with ten tosses.
Second strategy: A confidence interval is determined. You can find an explanation of how this works in the scattering range lexicon entry »C for Confidence Interval«.
In practice, an estimate often cannot be calculated with a simple division – as in our example. Our proprietary Jurojin software can efficiently calculate such complex estimates based on reliability data, for example.
Have you discovered a curious statistic in the media and would like us to cover it in a regular Streuspanne episode? Or do you have a statistical expression that you would like us to explain in the lexicon? Then get in touch with us at presse@itwm.fraunhofer.de
Have you just jumped here from another part of the Streuspanne world? Then quickly return to the long episode!
Find the podcast on all major portals including Spotify, Podcasters, Amazon Music, Apple Podcasts, Deezer, Google , RadioPublic , Pocket Casts or directly here on our website.
Fraunhofer Institute for Industrial Mathematics ITWM