Data Analysis

The next step in a successful business project is the data analysis process. Different data analysis procedures are used depending on whether the study is quantitative or qualitative. This module will also discuss the importance of validity and reliability during the data analysis process. This module provides a broad overview of the various data analysis techniques, how they should be used, and how to avoid misuse in business research. It then provides links to other resources that go into more detail on the technique of interest.           

Learning Objectives

• Understand reliability and validity
• Understand descriptive analysis and its role in business research 
• Understand the various techniques for analyzing qualitative data
• Understand the various techniques for analyzing quantitative data

Validity and Reliability

Validity and reliability of results determine how well the research study findings are accepted in the academic community and beyond. Validity measures whether a scale measures what it is intended to measure (Zikmund et al., 2013). This includes internal validity and external validity. Internal validity, also known as face validity, considers whether the findings correspond to the research questions. External validity considers whether the results could be applied to other situations (Quinton & Smallbone, 2006) 

Reliability considers whether measures are free from error and yield consistent results (Zikmund et al., 2013). The key components of reliability are stability, equivalence, and internal consistency. Reliability can be measured through a test and retest process, through use of two different questionnaires to check for consistent results, or split half correlations using Kronbach’s alpha statistical test (Quinton & Smallbone, 2006). 

Descriptive Analysis

Descriptive analysis is a common starting point for many projects.  It refers to the transformation of raw data into a form that will make them easy to understand and interpret. This can include rearranging, ordering, and manipulating the data to develop descriptive information (Zikmund et al., 2013). Nominal variables can be used for frequency tables, proportions like percentages, mode, and category percentages. Ordinal variable can also be used to rank order and to calculate the median. Interval variables allow for calculation of the simple arithmetic mean. Ratio variables are required for calculation of index numbers and more detail on the mean. 

Qualitative Analysis

Qualitative analysis is distinct from quantitative analysis in that there are no clear rules and agreed procedures for analyzing quantitative data. Approaches vary based on the nature of the qualitative enquiry, but all approaches have an analytic hierarchy through which qualitative findings are built from raw data. The steps include data management, descriptive accounts, and explanatory accounts. In the data management stage, raw data is reviewed, sorted, and labeled. In the descriptive stage, the researcher uses the ordered data to map the range and diversity of the phenomenon and to develop classifications. In the final stage the research builds explanations about why the data takes the forms as found (Ritchie, Lewis, Nicholls, & Ormston, 2013).  

Quantitative Analysis

Quantitative analysis offers a wider range of analytical tools, and therefore a greater risk of choosing the wrong tool for the analysis required. For descriptive and causal analysis, quantitative measures can be classified by the number of dimensions or variables under study. Univariate analysis is used to generalize from a sample about one variable at a time. Analysis may look at measures of central tendency and measures of dispersion. For nominal variables, the mode can be used to analyze the central tendency, but there is no measure of dispersion. For ordinal variables, the median may be used to analyze central tendency and percentile may be used to measure dispersion. For interval or ratio scale measures, the mean may be used to measure central tendency and standard deviation may be used to analyze dispersion (Kothari, 2004). 

Bivariate analysis involves tests of differences or association between two variables at a time. For interval and ratio scale data, the t-test of z-test are ideal for testing between two independent groups, while the one-way ANOVA can be used for three or more groups. If the data is ordinal scaled, Mann-Whitney U-Test or Wilcoxon test can be used for two groups, and the Kruskai-Wallis test for three or more groups. With nominal scale variables, the z-test and Chi-square test can be used for two groups, and the Chi-square test can also be used for three or more tests. Spearman Rank-Order Correlation offers a means to test for association even when the data is neither interval nor ratio scaled (Zikmund et al., 2013).  

Most business problems are multidimensional, therefore use of multivariate analysis, which allows the investigation of more than two variable at time, is commonly used.  Among the analysis techniques used in this situation are multiple regression, multiple correlation and multi-ANOVA. These techniques can be used to define relationships among dependent and independent variables. All require the interval and ratio variables (Kothari, 2004). 

Scenario: ABC Consulting

Data analysis differs in each of the scenarios, depending on the method. In quantitative methods, the most common method involves a statistical analysis of the data. This may include descriptive statistics that provide general information or comparative statistics that try to quantify the differences in variables. A survey or questionnaire would be analyzed usually in a statically program, such as SPSS.

Qualitative analysis would look at the meaning in the words of the participants. In this case, meaning is often described through themes coming from the words used in interviews and individual words from the participants. Themes in this case are analyzed either by hand or using qualitative software to isolate themes. 

For Further Study

Scales of Measurement Part 1: https://youtu.be/KIBZUk39ncI

Scales of Measurement Part 2: https://youtu.be/yJpiUHbLKLU

Reliability and Validity: https://youtu.be/x-lTanX_nok

Reliability and Validity in Qualitative Research Trustworthiness: https://youtu.be/P0lyl16P_S8

Fundamentals of Qualitative Research Methods: Data Analysis: https://youtu.be/opp5tH4uD-w

Describing a Categorical Variable: https://youtu.be/vrWYw8d2830

Descriptive Statistics, Histograms: https://youtu.be/zC3GaPBJ4c4

Introduction to the Mann-Whitney U Test: https://youtu.be/VIzvWiaTN2I

Introduction to the Wilcoxon Signed-Rank Test: https://youtu.be/wW9brbLfF8E

Introduction to the Kruskal-Wallis H Test: https://youtu.be/BMITawOrxwg

Introduction to T Statistics: https://youtu.be/a2rd4Qy8yNI

Single Sample Hypothesis Z-test Concepts: https://youtu.be/HoqzIR8xj4s

Introduction To the Chi-Square Test: https://youtu.be/SvKv375sacA

When to Use Z or T Statistics In Significance Tests: https://youtu.be/-bhfcBvyWoc

Stats ANOVA: https://youtu.be/EFdlFoHI_0I

Pearson’s r Correlation: https://youtu.be/2B_UW-RweSE

An Introduction to Linear Regression: https://youtu.be/zPG4NjIkCjc

Simple Linear Regression, The Very Basics: https://youtu.be/ZkjP5RJLQF4

Multiple Regression, The Very Basics: https://youtu.be/dQNpSa-bq4

Spearman Correlation: https://youtu.be/YpG2MlulP_o

References   

Bezzina, F., & Saunders, M. (2015). The pervasiveness and implications of statistical misconceptions among academics with a special interest in business research methods. Leading Issues in Leading Issues in Business and Management Research, 2(2), 127.

Kothari, C. R. (2004). Research methodology: Methods and techniques: New Age International.

Quinton, S., & Smallbone, T. (2006). Postgraduate research in business: a critical guide: Sage.

Ritchie, J., Lewis, J., Nicholls, C. M., & Ormston, R. (2013). Qualitative research practice: A guide for social science students and researchers: Sage.

Zikmund, W. G., Babin, B. J., Carr, J. C., & Griffin, M. (2013). Business research methods: Cengage Learning.