Sampling Theory

When conducting a series of trials it is important to cover all the key groupings. Sampling Theory explains how this can be done even when an exhaustive set of trials is impossible. Sampling theory is part of statistics.

In the most straightforward case, such as measuring the heights of children in a class or school it is possible to identify and measure every single member of the population. However, if we want to estimate heights for the people in a city or country it is not feasible to actually perform all the measurements. Given that we measure a certain number of heights, and that our measurements are not systematically biased, sampling theory can tell us how much uncertainty the size of our sample is introducing.

When we have many dimensions to consider, for example if we are measuring height, weight, hair colour, age and tendency to enjoy Opera, it is important that a reasonable cross section of each dimension is obtained.

Relation to Information Architecture

When Gathering Data it is quite normal to have:

A limited budget, so we can't talk extensively to everyone
A variety of different groupings of stakeholders, such as different departments, age profiles or job descriptions
Different levels of participation in the project, for example one department might be motivated to participate while others are not

It is important to keep track of the participants in as much detail as possible. It is common that late in the project you will realise that a crucial audience for the system has, so far, not been adequately represented. At this point you should be able to identify the "missing input" and ensure that these groups are sought out. You may even be able to ask to request an interview with "a cost accountant in the tractor division that deals with France".

Links to this page

The following pages link to here: Card Sorting, Gathering Data