Statistical Methods for Organizational Research

Theory and Practice

Chris Dewberry

• Publisher: Routledge
• Year: 2004
• ISBN:
  9781134314331
• Edition: 1
• Topic: Business
• Subtopic: Business General
• Index: Business

This clearly written textbook clarifies the concepts underpinning descriptive and inferential statistics in organizational research. Acting as much more than a theoretical reference tool, step-by-step it guides readers through the various key stages of successful data analysis.Covering everything from introductory descriptive statistics to advanced

Part I
Descriptive and inferential statistics

Chapter 1
Descriptive statistics

There is virtually no limit to the research that can be carried out on organizations. Areas as diverse as culture change, the factors associated with job satisfaction, team performance, marketing, organizational effectiveness, fairness, accounting practices and strategic management can be examined with a variety of research perspectives and using a broad range of research techniques. Very often this research involves measurement and, with it, the use of numbers. For example, it is possible to measure the job satisfaction of nurses and doctors by asking them how many aspects of the job they are happy with, or to measure the performance of hospitals by measuring how long patients take to recover from operations. With numerical measurement it is possible to collect information about large and representative samples of people, teams or organizations, and use this to look for important differences or associations between things, to make valuable predictions and to simplify complex relationships. By analysing and interpreting such numerical data, ideas can be confirmed or refuted, unexpected relationships can be discovered, theories can be developed and tested, and practical recommendations can be made. Numerical analysis is, therefore, of central importance to organizational research, and at the heart of numerical analysis is the discipline of statistics.

Statistics can be divided into two areas:

descriptive;

inferential.

Descriptive statistics is concerned with describing numbers and the relationships between them. Very often the intention is to capture the essence of the numbers, to summarize them in such a way as to render them as easy as possible to understand and digest. The second area, inferential statistics, is concerned with analysing numbers and drawing conclusions from them. Basically, this involves collecting a relatively small set of numbers (for example, those referring to the job commitment of 30 employees in a car manufacturing plant) and using them to make guesses about a larger set of numbers that the researcher is interested in (for example, the job commitment of everyone in the car manufacturing plant). Let’s begin by examining fundamental elements of descriptive statistics.

The numbers collected in quantitative research are referred to as data. So if you collect the ages of five people:

– Ruth: 35 years

– John: 42 years

– Julie: 57 years

– Andy: 27 years

– Jill: 33 years

the ages of these five people constitute your data. Each particular number, such as the 35 years of Ruth, is referred to as a data point. So here you have one set of data and five data points: 35, 42, 57, 27 and 33. Data can be collected from anything – rivers, shirts, carpets, clocks, mice – anything at all can provide us with numerical information. In organizational research, data are usually collected about individuals working in organizations (e.g. the appraisal ratings of 20 different employees), about collections of people (e.g. the number of units sold by 20 different sales teams), about organizations themselves (e.g. the share price of 100 companies), and about their products (e.g. the number of defective radios produced by an electrical manufacturer). In this book, for the sake of consistency, the focus is on individuals working in organizations. In other words, the unit of analysis is people, and it is assumed to be individual people that data are collected on. However, it should be borne in mind that the ideas and principles to be explained can, equally, be applied to other units of analysis such as teams, organizations and the goods and services that organizations produce.

As already explained, the data we collect about people can concern a variety of different things, such as their well-being, their job performance, their age and so on. Each of these is referred to as a variable. A variable is simply something that is measurable and that varies. So age is a variable because we can measure it and not everyone is exactly the same age.

Data come in one of two different types: categorical and continuous.

CATEGORICAL DATA

Using categorical data, as the name suggests, involves placing things in a limited number of categories. Usually in organizational research these ‘things’ are people, and typical categories include gender and level of seniority. So if we say that we have 34 women and 10 men in an organization, we are referring to categorical data because people are being placed in either one of two categories.

In principle, many different types of categories can be used to classify people. Categories are used for the colour of eyes (green-eyed people, blue-eyed people, brown-eyed people etc.), for the region of the country where people live, or all sorts of other things. In organizational research, categories such as gender and level of seniority tend to be selected because these are thought to be the most important. Sometimes the categories exist before the research begins, and gender and level of seniority are examples of this. At other times categories are created during the course of the research. For example, a questionnaire might ask managers to indicate whether they agree or disagree that the organization should be

Table 1.1 Age and gender of five people, with gender coded 0 and 1

re-structured. Those agreeing and disagreeing can be viewed as two different categories of people. Whatever categories are chosen, they are normally qualitatively different from each other: men are qualitatively different from women because we cannot reduce all the differences between them to a small number of dimensions such as how tall they are, how socially perceptive they are, or how extroverted they are.

Because, nowadays, computers are usually used to analyse data, and computer programs designed for statistical analysis are much better at dealing with numbers than with words, 4categorical data are usually represented numerically. So, instead of telling a computer that someone is ‘male’ or ‘female’, we may tell it that the person has a value of 0 or a value of 1 for gender, and that a 0 for gender indicates that they are male whereas a 1 indicates that they are female. You could represent the information about the five people mentioned earlier in the way shown in Table 1.1.

In Table 1.1, we have assigned the number 1 to females and 0 to males. The coding is arbitrary – the fact that 1 is greater than 0 does not mean that females are greater than males – we might just as well have coded males as 1 and females as 0. The important thing is that we have a numerical code that indicates which category people are in: in this case the female category or the male category. It is because these numbers simply indicate which category someone is in, and do not imply that one category is somehow greater than another, that we know we are dealing with categorical data.

Another example would be the classification of employees according to the part of the country they work in: North, South, East or West. If you counted the number of employees working in these regions and found the following:

– North: 156

– South: 27

– East: 43

– West: 92

you would have categorical data. As with the gender data considered earlier you know that they are categorical because the information you have about each person does not suggest that they are greater than someone else in some way, but mer...

Table of contents

• Contents
• Illustrations
• Foreword
• Acknowledgements
• Part I Descriptive and inferential statistics
• Part II Methods of statistical analysis
• Answers
• Appendix
• References
• Index