13 Quantitative Analysis

Learning Outcomes

By the end of this chapter, learners will:

• Understand basic terminology associated with quantitative analysis,
• Describe the differences between descriptive and inferential statistics.

Introduction

Quantitative data are data represented numerically; this includes anything that can be counted, measured, or given a numerical value. Quantitative data analysis is the process of summarizing the data using statistical analysis. In this sense, quantitative data analysis allows us to describe observations that can be found in a dataset. Quantitative data analysis can:

• Succinctly summarize data
• Highlight trends in quantitative data
• Show a rate of change
• Be used to draw comparisons
• Determine the average, or mean

A lot of OER related data is quantitative. For example, counting the number of courses that have adopted, adapted, or created OER is a type of quantitative data.  Quantitative data analysis can, therefore, shed light on the magnitude of the effect of OER programs or projects, the most pressing needs of an institution or a department when selecting course materials, or how different introductory courses at your institution have adopted OER at different rates. Relevant quantitative data will likely be produced via surveys, questionnaires, and web analytics. Student GPAs and student enrollment rates may also be used to tell a story about OER impact.

Terminology

There are four key terms that are commonly used to describe quantitative data:

The independent variable is the variable that is not influenced by other factors. In this sense, the outcome is independent of other circumstances. For example, if I were running a trial to see if the type of course materials used has an impact on student scores, my independent variable would be the type of course material – I might have some students use a commercial textbook and other students use an OER.

The dependent variable is influenced by other factors. The outcome is dependent on the other circumstances, especially the independent variable. To return to the above example, my dependent variable would be the test scores, since those are influenced by the course materials that are used.

There are also two different types of data. Discrete data is any data that can only be specific values. Most often, discrete data is any data represented by whole numbers. This data cannot be broken down into smaller values. Examples in OER include: the number of courses in a given faculty that use OER or test scores. Most quantitative OER data is discrete.

Continuous data can take any value. Common examples of continuous data include temperature and weight. This type of data is rarer in OER.

Quantitative Analysis

We can draw conclusions about our data and OER program through quantitative data analysis. The two types of quantitative data analysis are descriptive statistics and inferential statistics.

Descriptive Statistics

Descriptive statistics are used to describe and summarize data, such as percentages, averages, measures of central tendency (mode, median, mean), and measures of spread (like range or standard deviation) in a sample or dataset. They might describe the distribution of a single variable (univariate descriptive statistics), the relationship between two variables (bivariate descriptive statistics), or more than two variables (multivariate descriptive statistics).

Many of the questions that we have about our OER programs and their impact can also be answered with basic descriptive statistics. Questions could include:

• Average savings from one OER
• Common perceptions of OER on campus (eg, 50% of individuals surveyed had a positive perception of OER)
• The proportion of courses that use an OER on campus
• The disciplines or courses in which OER are most often used
• The percentage of students who preferred an OER over the commercial text

Many of the survey tools detailed in the Tools and Technologies section (like Google Forms and Qualtrics) can automatically generate certain descriptive statistics for individual survey questions, assuming that the dataset that these tools are working with is clean (see the Data Cleaning section). If you do need to calculate these measures yourself, some of the additional software listed in the Tools and Technologies section (such as Excel, Google Sheets, and SPSS) will be useful.

The specific analysis that you conduct, and the way(s) in which you present results, will depend upon the data and variables that you’re looking at. To learn more about how to conduct analyses and present results, read the Quantitative Data Analysis Section in Practicing and Presenting Social Research.

Inferential Statistics

Inferential statistics allows you to draw inferences from data, by making predictions, deductions, or generalizations based on a sample. By analyzing this sample, it becomes possible to draw larger conclusions about the dataset as a whole. The two main types of inferential statistics are hypothesis testing (a category of these that can be used to determine whether there is a relationship between two variables)  and regression analysis (quantifying how one variable will change with respect to another variable). In the case of OER work, inferential statistics may be used to determine:

• Whether use of OER had an impact on student retention,
• Whether use of OER had an impact on student performance in a course.

Again specific analysis that you conduct, and the way(s) in which you present results, will depend upon the data and variables that you’re looking at. To learn more, read the Quantitative Data Analysis Section in Practicing and Presenting Social Research.

Selecting Appropriate Statistical Analyses

There are several factors to consider when selecting appropriate statistical analyses.

First, consider the data that you have collected. Certain analyses can only be conducted with certain types of data. For example, correlation and linear regression analyses can only be conducted with continuous data from a dependent variable. Knowing which data comes from independent variables and which comes from dependent variables is key.

Second, it is important to consider the goals of your data collection and analyses. Different goals, and different data, will best suit different types of analyses. For example, if I were analyzing the success of an OER Grant program, I could include the following statistical analyses:

• Determine the average amount of money that a grant-funded OER saves students in a single year.
• Send a survey to students who used a grant-funded OER. Use statements such as “The interactive elements of the OER positively impacted my learning” that students must either agree or disagree with using a likert scale. Determine the percentage of students who respond positively to these different statements.

The analyses that can be conducted will also depend on the data that was collected to achieve your goals. This is why survey design and methodology design are important.

Conclusion

Many of the quantitative analyses that you’ll want to run on data from your OER program can likely be generated automatically through survey tools, or with some basic calculations in spreadsheet software. However, if you need to run more in-depth analyses to get the answers you’re looking for, the resources list below will be helpful.

Resources

• Quantitative Data Analysis in Practicing and Presenting Social Research
• Social Data Analysis

References

Gallant, J. (2022). 21. Data Collection and Strategies for OER Programs. In A. K. Elder, S. Buck, J. Gallant, M. Seiferle-Valencia, & A. Ashok, The OER Starter Kit for Program Managers. Rebus Community. https://press.rebus.community/oerstarterkitpm/chapter/chapter-21-data-collection-and-strategies-for-oer-programs/

Robinson, O. & Wilson, A. (2022). Quantitative Data Analysis. In O. Robinson and A. Wilson, Practicing and Presenting Social Research. UBC Library. https://pressbooks.bccampus.ca/undergradresearch/part/quantitative-data-analysis/