Unit 2 Research Process and Strategy

Methods of Data Collection

1

A necessary step in applying statistical methods is to collect data. Data can be considered the raw numbers, which are then analyzed and turned into information. Recommendations are then made based on the information.

There are three basic ways to collect data.

The first way to collect data is formal research, usually in the form of a survey. This, along with the next collection method, is one of the most common ways to collect data. Not everybody has designed a survey, but most people have at least taken one. It is important to give the survey to a large enough and representative sample of the people whose opinions you want to gather. Surveys are used to gather data about the following:

• Attitudes (How do you feel about this?) • Preferences (Which do you prefer? Which is better or worse?) • Stated behavior (How often do you do this? How likely would you be to

purchase this product?)

Many surveys contain some unstructured, unforced-response questions (to which respondents can answer anything they want), but the structured, forced- response questions (that only allow the respondent to choose from among predetermined answers) provide the vast majority of data that can be further analyzed using statistical methods.

Surveys can be conducted in a variety of ways, each having pros and cons:

• By telephone • In person • On paper • Over the Internet

The second way to collect data is as a by-product of day-to-day operations. Data are often automatically collected as a result of processes that occur during the operation of the business. This is the other common way of collecting data. Examples include the following:

• Sales figures • Machine output • Web site usage statistics • Hours worked by workers clocking in • Bank deposits and withdrawals

Methods of Data Collection

2

The third way to collect data is by observations made directly by people. This is used when it is not possible to conduct a survey, and the data are not otherwise available. Examples include the following:

• Sampling every nth product as it comes off the assembly line and observing whether there are any defects

• Counting the number of empty spaces in a parking lot every hour • Observing whether it rained on a particular day

Keep in mind that many characteristics about people (or other things) are not naturally represented by numbers. A good example of this is gender— male and female. This question appears on virtually all consumer surveys and can be analyzed statistically by arbitrarily assigning the value 0 for male and 1 for female (or vice versa). This technique is also used to analyze survey questions that have a yes-or-no response. Yes is usually coded as 1, whereas no is coded as 0.