Introduction

This introduction will discuss the basic concepts of all statistics. The intent of the introduction is to sensitize the reader to the importance of taking statistics into consideration in the design and planning of experiments. Unless the experimenter plans a study appropriately, accounts for certain issues that are inherent in any study, and understands what is needed for a successful experiment, all will be for naught.

WHAT ARE DATA?

Everything we do is based on data. So, the question quite often is: should the word be datum or data? Grammatically speaking, the singular word is datum and the plural is data. However, because generally speaking we have more than one, the convention is that we use data. In common usage, data are any materials that serve as a basis for drawing conclusions. (Notice that the word we use is "materials." That is because materials may be quantifiable or numerical and measurable or on the other hand may be attribute or qualitative. In either case they can be used for drawing conclusions.) Drawing conclusions from data is an activity in which everyone engages ” bankers, scholars, politicians , doctors , and corporate presidents . In theory, we base our foreign policy, methods of treating diseases, corporate marketing strategies, and process efficiency and quality on "data."

Data come from many sources. We can conduct our own surveys or experiments, look at information from surveys other people have conducted , or examine data from all sorts of existing records ” such as stock transactions, election tallies, or inspection records. But acquiring data is not enough. We must determine what conclusions are justified based on the data. That is known as "data analysis." People and organizations deal with data in many different ways. Some people accumulate data but do not bother to evaluate it objectively. They think that they know the answers before they start. Others want to examine the data but do not know where to begin. Sometimes people carefully analyze data, but the data are inappropriate for the conclusions that they want to draw. Unless the data are correctly analyzed , the "conclusions" based on them may be in error. A superior treatment for a disease may be dismissed as ineffectual; you may purchase stocks that do not perform well and lose your life's savings; you may target your marketing campaign to the wrong audience, costing your company millions of dollars; or you may adjust the wrong item in a process, and as a consequence, you may affect the response of the customer in a very unexpected way. The consequences of bad data analysis can be severe and far-reaching. That is why you need to know how to analyze data well.

You can analyze data in many different ways. Sometimes all you need to do is describe the data. For example, how many people say they are going to buy a new product you are introducing? What proportion of them are men and what proportion are women? What is their average income? What product characteristic is the customer delighted with? In other situations, you want to draw more far-reaching conclusions based on the data you have at hand. You want to know whether your candidate stands a chance of winning an election, whether a new drug is better than the one usually used, or how to improve the design of a product so that the customer will be really excited about it. You do not have all of the information you would like to have. You have data from some people or samples, but you would like to draw conclusions about a much larger audience or population.

At this juncture your answer may be, "I do not have to worry about all this because the computer will do it for me." That is not an absolute truth. Computers simplify many tasks , including data analysis. By using a computer to analyze your data, you greatly reduce both the possibility of error and the time required. Learning about computers and preparing data for analysis by computer do require time, but in the long run they substantially decrease the time and effort required. Using a computer also makes learning about data analysis much easier. You do not have to spend time learning formulas. The computer can do the calculating for you. Instead, your effort can go into the more interesting components of data analysis ” generating ideas, choosing analyses, and interpreting their results.

Because calculations are the computer's job, not yours, this volume does not emphasize formulas. It emphasizes understanding the concepts underlying data analysis. The computer can be used to calculate results. You need to learn how to interpret them.