Research Methods and Data Analysis for Business Decisions

Table of Contents

1. Frontmatter

2. Designing the Study

   1. Frontmatter

   2. 1. Research Methods and Philosophy of Science

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      Big data, SoLoMo (social, local, and mobile), Screen Scraping, ClickStream, Passive Data Collection, Algorithms, Network Intelligence, Real Time, WebCrawling, Data Mining, Dashboards, Blog Mining, Neuromarketing—never before have data analytics been subjected to such intense creativity and innovation. Never before has it been more important to return to the roots, to the scientific foundation of how to interpret, understand, and rely on data analysis. Making flawed decisions based on huge amounts of data through automated data processing can be more damaging than not making decisions at all.

   3. 2. The Research Process and Problem Formulation

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      What insights are needed for making decisions? With the advent of digitalization and big data, do managers have perfect information? Making a decision is easy. However, making the right decision requires insight and knowledge. Knowing the weaknesses in information is critical for making wise decisions. Digitalization and big data do not directly translate to quality decisions.

   4. 3. Research Design

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      Good market analysis is largely about empathy, often expressed as the ability to put yourself in another person’s shoes. The challenge comes when the people being studied will not or cannot express their opinions, or they literally do not know what is good or bad for them. How, then, can we develop products, services, and companies that create value? Innovation across industries is a major trend in entrepreneurship. Perhaps one has to look to other markets to gain knowledge of one’s own market.

3. Data Collection

   1. Frontmatter

   2. 4. Secondary Data and Observation

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      In this chapter, we discuss secondary data and collecting data by observation. Secondary data already exists, having been collected by someone else for another purpose. In contrast, primary data is collected to address specific research questions. Chapters 5 and 6 describe primary data collection. Observation is a structured way of collecting data about someone or something without direct communication. Observation may be collected as primary or secondary data. However, with the advent of the Internet and digitalization, the amount of secondary observational data being collected is staggering. It is revolutionizing data collection and analysis.

   3. 5. Qualitative Methods

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      Perhaps a little oversimplified, we can say that qualitative methods delve deeply into a subject to find whether something exists and to understand how it works. With quantitative methods, the focus is on measuring how much exists. Though the methods may seem incompatible. In fact, they are complementary and can be conceptualized as the two anchors on a scale. It is the research question that determines which method should be used in a given context, and in many situations, both are applicable. Using different methods along the qualitative spectrum on a specific research question is called method triangulation. It provides different perspectives on the same problem, and when the same conclusions are reached, it increases the validity of the findings. Applying qualitative and quantitative methods in harmony increases the robustness of the research. Both methods are used in social science research and business research because most social phenomena have both qualitative and quantitative aspects. Strictly speaking, it is the format of the data that determines whether it is characterized as either qualitative or quantitative. Quantitative data is expressed as numbers, while other data is referred to as qualitative.

   4. 6. Questionnaire Surveys

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      As authors, our goal with this book is to provide the reader with research tools that improve decision skills. One way to measure how well the book is working could be to measure how engaged the reader is in the subject. We could ask the following questions (Fig. 6.1):

   5. 7. Sampling

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      When was the last time you thoroughly and conscientiously answered a survey? Now you find yourself sitting on the other side of the table wondering how to get enough of the right respondents to thoroughly and conscientiously answer your survey. Whether doing qualitative or quantitative research, you need to decide where to get data. The procedure is important for quantitative research because you will run statistical analyses and may want to generalize from the results. It is equally important for qualitative research because you need to identify research objects that can provide rich information on the phenomenon being studied. In this chapter, we will go through different types of samples and how to determine sample size. This includes identifying respondents for qualitative methods like depth interviews and focus groups. We discuss probability and non-probability sampling, and sources of sampling error.

4. Quantitative Data Analysis

   1. Frontmatter

   2. 8. Simple Analysis Techniques

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      A relatively recent data analysis trend is to examine clickstreams on the Internet: which pages you visit, how long, what you click on, the frequency of clicks, which pages you came from, and which pages you go to. Clickstream analyses look at correlations and patterns in online behavior. Take note, clickstreams do not say anything about the motives behind the behavior or who the person is. Nevertheless, the data is so extensive that the patterns still provide valuable information. In this section of the book, we will learn about the most common quantitative analysis methods. Our examples are done in IBM SPSS Statistics software. Basic statistics can be done in Excel, although it has the disadvantage of not being a dedicated statistics software so it is, relatively speaking, quite limited. Other popular software are JMP, SAS, and Stata. Each has its own advantages and disadvantages. Often, different scientific domains or even geographic areas favor different software. It is wise to consider where and with whom you will be working before deciding on which software to use. One important distinguishing characteristic is whether a software has a windows-based point and click interface or whether it requires syntax-based programming. While point and click may be intuitively easier, syntax-based programming has the advantage of leaving a clear trail of how the analysis was conducted. One does not always remember what choices were made in the point and click environment. SPSS is a windows-based point and click software, though it can be programmed with syntax. To leave a trail of what choices were made, when in an SPSS dialogue box, clicking on “paste” will generate a syntax file of the requested analysis. On the book’s website, you can get the data for the examples in each software format.

   3. 9. Hypothesis Testing

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      If you approach your career as a business model, that is, how you create value to make money, then consider the vital link between analytical expertise and business knowledge. Unlocking the value in big data requires analytical skills coupled with an understanding of how to apply it in business. Data alone does not create value. You, along with a computer, process and interpret data to provide valuable insights. Negligent misuse of data often causes more harm than good.

   4. 10. Regression Analysis

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      Imagine that someone has a lot of data about you: where you work, go to school, what you like, what you do not like, who your friends are, and so on. Then, based on this data, with great accuracy they can predict your behaviors and choices. What car you will purchase, where you will go on holiday, what you will buy your mother for her birthday. They can do this because most human behaviors are rational in so far as they follow consistent patterns; we are creatures of habit. And, they can predict your behavior because they know how to use regression analysis.

   5. 11. Cluster Analysis and Segmentation

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      About 10 years ago, the International English School opened a campus in Uppsala. At the time, the Uppsala International School staff were very worried that they would lose a lot of their students when the new school opened. They were afraid that the two schools had the same offering and that students might change due to convenience or because the English School had new facilities. To investigate their fears, a professional marketing consultant along with one of the authors of this book asked the staff of the International School to describe what was unique with their educational offering. Their answer: “Uppsala International School does everything!”

   6. 12. Factor Analysis

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      Would you let someone measure your intelligence with one question? What if you are good at algebra, but bad at geometry? Perhaps you are gifted with language and poor at mathematics? How can a single question capture all the dimensions of human intelligence to arrive at a reasonable conclusion? In research method terms, can human intelligence be validly measured with a single question? Our answer is no.

5. Reporting

   1. Frontmatter

   2. 13. Reporting Findings

      James E. Sallis, Geir Gripsrud, Ulf Henning Olsson, Ragnhild Silkoset

      Abstract

      From a decision-making perspective, we have discussed both qualitative and quantitative aspects of the research process. Our mission with this book is to promote a data-driven scientific approach to decision-making. Our world is drowning in data. Data that is grist for the mill of data analysis. When properly analyzed, data provides a wealth of value to decision-makers. Given the nature of digital data, quantitative methods dominate the pages of this book. Despite this, we advocate for a pragmatic approach, fluctuating between inductively developing theory and deductively testing it. Qualitative and quantitative methods are innately related in a scientific process that leads to better theories and better decisions.

6. Backmatter