Microwave Anisotropy Probe

Error Control on MAP

In any experiment, two types of error will inevitably be encountered: systematic errors and random errors. The ability to identify, eliminate, and subtract these types of error is crucial to obtaining reliable data. This page is a description of the terms "systematic error" and "random error." It will also describe some of the methods the MAP scientists are using to minimize experimental and systematic errors.

Random Error

There is some amount of random error in every measurement we take. This is partially due to the sensitivity of our instruments and partially due to human error. For example, if you were to measure the width of your computer screen right now, you might come up with a width of 26.7 centimeters. Now, how can you be sure that 26.7 cm is the exact width of your screen? You cannot. Several sources of error make an exact measurement impossible in this case: you might not have looked at each end of your ruler straight-on, causing errors due to parallax. The ruler might not have been exactly parallel with the top and bottom of the screen, resulting in a slight overmeasurement of the screen. If the width fell between the notches on the ruler, as is the case with 26.7 cm, the extra 0.7 cm measurement is based on human judgment and could possibly be off by one or two millimeters. This list of error sources goes on and on, and this is a fairly simple measurement. The objective in any scientific experiment such as MAP is not only to collect and analyze data, but to collect data will the smallest possible error.

Systematic Error

In the above example, the errors are random. Those errors such as error due to human judgment are random errors. However, other errors concern the instruments themselves: the ruler itself probably is not calibrated exactly to the world SI standard meter. If the ruler is made of metal, and the room temperature is not the same as the temperature at which the ruler was calibrated, each centimeter will be slightly longer or shorter than the calibration. These errors are called systematic errors because every measurement that you take with that ruler will contain the same error. In fact, you would not even know that your ruler was a source of systematic error until you compared it with another ruler calibrated exactly to the SI standard meter. Systematic errors are often much harder to recognize than random errors and can be more harmful to the final result.

Error Control

The scientists and engineers behind MAP have developed several methods that will help eliminate random and systematic errors in MAP's data:

Symmetry
The structure of MAP is highly symmetrical. This reduces the sources of error because it allows MAP to measure the CMB anisotropy using two different reflectors from the same position in space. This reduces both the random and systematic error.

Anisotropy
The fact that MAP measures the CMB anisotropy rather than the absolute temperature values of the CMB is a major factor in reducing error. For a more detailed explanation of why this is the case, see The Educator's Page.

Frequency Measurements
MAP collects data at five different frequencies in the microwave band of the electromagnetic spectrum. These are: 22 GHz, 30 GHz, 40 GHz, 60 GHz, and 90 GHz. This has the effect of allowing MAP scientists to eliminate the galactic foreground signal from the data.

Repeated Observations of the Sky
MAP is designed to observe a large fraction of the sky every day. Over the course of a year, each piece of the sky will be reobserved several thousand times, thus reducing the effects of a one-time random error contaminating several data points.

Spin Angle
The angle between the spacecraft's spin axis and the sun is kept constant in order to maintain a stable illumination of MAP's solar panels, providing a thermally stable environment. This in turn reduces the systematic effects.

Back to the Non-technical Technical Information Page

Back to the MAP Education and Outreach Page


Technical Back to the MAP Technical Information Page


Cosmology Back to the Introduction to Cosmology Page


Home Back to the MAP Home Page


NASA

Please help us make this web site more useful and enjoyable by telling us what you would like to see at this site:

Mail

David N. Spergel / dns@astro.princeton.edu
Gary Hinshaw / hinshaw@stars.gsfc.nasa.gov
Charles L. Bennett / bennett@stars.gsfc.nasa.gov

NASA Privacy Statement
NASA IT Security Banner

Last updated: Friday, 05-21-1999