Quantitative research is a systematic investigation that relies on numerical data to understand patterns, relationships, and causes within educational settings. In education and the social sciences, it is commonly used to measure topics spanning from student achievement, management strategies, and evaluation of program effectiveness.

Quantitative research involves the collection and analysis of numerical data to uncover patterns, relationships, and trends. It employs structured methodologies and statistical techniques to quantify observations and test hypotheses. Unlike qualitative research, which focuses on understanding subjective experiences (verstehen) and perspectives (Creswell, 2015), quantitative research aims to produce objective, replicable findings based on quantifiable measures. However, to navigate the complexities of quantitative research, a critical and reflective approach is essential for addressing challenges and maximizing the impact of research endeavors. To do so, let us first understand the strengths and weaknesses of quantitative research and incorporate essential considerations into research practices. By doing so, you can conduct rigorous and meaningful quantitative research that contributes to your scholarly discourse and real-world applications.

Why Use Quantitative Research?

Quantitative research is not only about numbers. It is what these numbers reveal as they help uncover patterns, test theories, and enable to make predictions with precision and objectivity.

First, a powerful characteristic of this method when conducted properly, is its predictive power which helps us forecast trends, such as how student engagement impacts academic performance or how leadership styles influence institutional success. In addition, having entered into a new era, we continue to experience the rise of large-scale quantitative analysis, machine learning, and AI-driven analytics. These challenging and inspirational opportunities make quantitative research even more powerful. Moreover, as discussed in the previous chapter, quantitative research allows for controlled environments (experimental control), enabling us to isolate variables and test causality. Furthermore, one of its greatest power is that well-designed quantitative studies can be repeated and verified, making findings more robust over time.

Given its aforementioned powerful characteristics, you may prefer to use the quantitative method in your research if 1) you need measurable, generalizable insights, that is, if you are looking to understand trends across large populations (e.g., how international students adapt to online learning); 2) you prefer objectivity and comparability – when testing hypotheses with minimal bias, such as evaluating different teaching methods based on student outcomes; 3) your goal is to establish causation by means of experimental and quasi-experimental designs, helping you determine if one variable directly influences another; and 4) you require statistical evidence for prediction and decision-making since policy makers and administrators often rely on quantitative data to guide resource allocation and institutional improvements.

As discussed in Chapter 6, yes, the quantitative method is often viewed as the golden standard for reliability, because it reduces variability (large sample sizes allow broader generalization) and offers data-driven decision making (reduces subjective interpretation). However, it is fundamental to keep in mind that numbers alone may not explain why a phenomenon happens. As noted in the upcoming chapter, Chapter 8, poorly designed surveys or biased datasets can distort findings. Furthermore, just because something is statistically significant does not mean it has practical significance.

In sum, quantitative research is powerful, but it is not a one-size-fits-all solution. This method is most reliable when properly designed and interpreted with context. In a situation when numbers and their interpretations alone do not give us the whole picture of the phenomenon, we want to dig deeper and explore the why of the phenomenon. That is when we need to complement our research by adding qualitative insights. In this case, the mixed method (blended) research (Tashakkori & Teddlie, 2010) would be our best option, which is fully explained in Chapter 12.

Common Quantitative Research Designs

As in qualitative research, quantitative research method also includes different designs, dependent on the purpose of the research. In this section, we will only cover a few of the common designs to give you an introduction. The goal of these designs is to describe patterns, explain attitudes, determine relationships, and to look into effects. Below are the descriptions of the commonly used designs with examples. The critical aspect of these designs is the language, which is usage certain word choices.

Descriptive Research Design

Descriptive research describes characteristics of a population or phenomenon without manipulating the variables. This design focuses on summarizing data and identifying patterns or trends. In a descriptive study, data can be collected by means of a rating scale (Likert Scale) which would yield numerical data. But what is a Likert scale? The scale was developed by Rensis Likert (1932) who was an American social and organizational psychologist, and is considered a rating scale (5-7 point ordinal scale) with range of answers such as “strongly agree,” “agree,” “neutral,” “disagree,” and “strongly disagree,” allowing respondents to indicate their level of agreement/disagreement regarding statements on attitudes, approaches, trends (Roy, 2020). An example would be to describe the sleeping habits of high school students. Your survey results would yield the description of sleeping habits.

Descriptive Research Design Example

The purpose of this research is describe the sleeping habits of high school students. First, to measure the sleeping habits, we need to have a sample (large) selected from a population of high school students. Then, we need to develop an instrument to do the measuring—in this case it will be a survey (Likert Scale). The survey would include statements regarding sleeping habits. The responses students would be expected to mark would vary from “strongly agree,” “agree,” “neutral,” “disagree,” and “strongly disagree.” Important note, if we develop the survey, we would need to first, pilot it, that is test it to determine the reliability and validity of the survey on a small group of students. Once we know that the survey questions will inquire what it is supposed to inquire (validity) and our measurement will be consistent over time (reliability—if we give the survey again, the survey would yield the same numbers), we can go ahead and distribute the survey to collect data. But how do we collect data? First, in this case, we can randomly select a group of 500 students (a large sample) from X High School and distribute the survey. The responses students would mark would vary from “strongly agree,” “agree,” “neutral,” “disagree,” to “strongly disagree.” With this approach, data analysis will yield percentages, numbers of attitudes on their sleep habits which is what we want.

Correlational Research Design

Correlational research examines the relationship between two or more variables to determine whether changes in one variable are associated with changes in another variable. Correlation does not imply causation but helps identify potential relationships for further investigation. With this design we need to pay attention to two factors: 1) determine the direction of the relationship (positive or a negative), and 2) determine the strength of the relationship (significantly strong or weak). If there is no significant strength (zero), there is no relationship to mention. An example would be quality sleep and academic success. Based on these two variables, our central research question would be: Is there a significant relationship between quality sleep and academic achievement? Here is another aspect regarding our central question—if we are almost 100 percent sure—meaning we have a strong hunch about this relationship, instead of developing a central research question, we can opt for a hypothesis, in this case, an alternative hypothesis. Another hypothesis to mention is null hypothesis but in this situation due to our strong hunch, we select the Alternative Hypothesis: There is a significant relationship between quality sleep and academic achievement. If there is no significant strength, there is no relationship to mention. To determine the relationship a statistical test such as Pearson Product Coefficient will be used.

Correlational Research Design Example

The purpose of this research is to explain the relationship between academic achievement and quality sleep. Once we define “quality sleep” and “academic achievement,” we need to develop or identify an inventory regarding quality sleep. What is “quality sleep” and how many hours is required to get quality sleep? Remember the inventory or test needs to have a high reliability and validity score. If we can identify an inventory which has been successfully used by other researchers, we can easily use that inventory. For the “academic achievement, “we can easily refer to the students’ official academic records. In a correlational design, we would need to determine 1) the direction of the relationship and 2) the strength of the relationship. In addition, assuming we find a significant relationship (in this case-positive) between the two variables, we can also use these findings and make predictions by conducting a Regression Analysis. But we will not cover Regression Analysis in this course. You just need to know that a significant relationship can help us make predictions. So, based on our findings, we can predict that “a student who gets quality sleep will likely be academically successful.”

Experimental Research Design

Experimental research involves manipulating one or more independent variables to observe their effects on the dependent variable(s) while controlling extraneous variables and measuring the outcome. It allows researchers to establish cause-and-effect through random assignment of control groups. According to Creswell (2015), “a ‘true’ experiment involves random assignment of participants, groups, or units. This form of experiment is the most rigorous” (p. 328). Experimental research design may involve one or more than one group of participants. If multiple, the groups would need to be similar in characteristics or attributes.

Experimental Research Design Example

The purpose of this research is to determine the effect of the “flipped classroom technique” on student learning. You want to know if student learn better under the traditional method or under the new method (flipped classroom technique). You can do this in two ways: 1) First, you can identify a school and matching participants or homogeneous samples and randomly assign them to two groups. One group will be your control group and will be taught by the traditional method without any manipulations. You can determine the effect by a test. The second group will be your experimental group and will be taught by the new method. What you want to do is determine the effects of this method on their learning by assessment again. Based on the two methods and two assessment outcomes, you can use a statistical test (t-test) and reach a decision by answering the following questions: Is there a difference in their learning? Is the difference significant? YUour findings will indicate the difference as a significant yes or a significant no.

You could also do this as a pre-experiment with one group. First, you would measure the effects on learning by assessing the students (pre-) prior to the experiment. Then, you can manipulate the method by using the flipped classroom technique and determine the effects on learning again by assessing the students (post). Once you have two outcomes, by looking into the differences, you can determine the significance of the effect.

Quasi-Experimental Research Design

Quasi-experimental research resembles experimental research but lacks random assignment to treatment conditions. Instead, researchers use pre-existing groups or natural variations to study the effects of interventions or treatments. An example would be studying the impact of a new curriculum in one school while using another as a comparison group. In quasi experimental design, you would be using an existing classroom teaching the traditional method and flipped classroom technique. Because your participants would not be randomly selected, your experiment will become quasi experimental, and you may experience threats to your research such certainty.

Longitudinal Research Design

Longitudinal research follows participants over an extended period to study changes or developments in variables over time. It allows researchers to examine patterns of stability, change, and continuity in individuals or groups.

Survey Research Design

Survey research involves collecting data through surveys or interviews (not to be confused with qualitative interviews) to gather numerical data on attitudes, opinions, behaviors, or characteristics of individuals or groups. Selected participants are given a survey based on a scale (Likert scale) to describe their thoughts on trends, etc. The difference between experimental design is that survey design does not involve a treatment (Creswell, 2015). You can use a survey design, for an example, to find out about the trends of social media users among high school students.

Plan to Conduct a Rigorous and Successful Quantitative Study

There are several critical considerations to ensure a rigorous and successful study. Following is a comprehensive list of factors to consider:

1. Significance of our Research Problem (What is a significant problem?)

First, prior to starting our research, we must define a problem which is worth investigating. This is also referred to as the “so what” effect of the research which applies to all research projects. So, how do we identify a research problem? Think about the problems you observe or have to tackle at your schools, classrooms, districts, communities, workplaces! Remember these “problems” need to have to be significant meaning they need to be justified. To justify our research, here is how our thought process works (Mills & Gay, 2019). As short-term thoughts appear and cues bubble up as research problems, according to Weick (1995): “our sensemaking perspective is in operation” (p. 2). These bubbles help us shape those problems we identify. We think about the factual aspect and significance (Is it worthwhile to research this problem? What makes it significant?). Once we identify a significant problem, we can move forward to reviewing past research.

So here is how it works—we need to ask ourselves why this research problem is worth investigating—hence, the significance of research. For an example, in this example I want to determine if there is a relationship between quality sleep and job performance. My two variables would be quality sleep and job performance. But which is an independent variable and which is a dependent variable? In terms of identifying independent and dependent variables, we need to determine which variable affects the other variable(s), that is—which is manipulated and which is measured. Since quality sleep differs and can be manipulated, it is my independent variable. Now, we need to study and operationalize them into measurable and observable constructs.

2. Review of Literature (What previous similar research has been conducted?)

Second, once we have a research problem defined, we should not immediately start conducting research. We need to investigate past research to determine if this particular problem had been investigated. Here are some questions to consider: 1) Why is it critical to give past researchers credit by recognizing their work. 2) Do we want to continue in their footsteps? 3) Do we want to branch off and look into another aspect of their research problem? Once we gather publications on our research problem, or any research that is similar to my research question, we need to make notes on the data collections, findings and implications of those research projects. If we are unable to determine any grounded research on this problem, we can conclude that there is a gap in the literature and be the first to look into this particular research problem.

3. Identify Your Purpose with (a) Central Research Question(s) (What is the purpose of your research plan and what do you want to find out?)

Third, once the literature is scanned, past research is recognized, we need to determine how we want to conduct our research. We need to set a goal, which is identify a purpose, meaning what it is that we want to find out. When building a purpose statement, there is a formula, a certain language that we need. We can’t just come up and say, I want to find out about this problem x. Here is how it works: In quantitative research we describe patterns, explain relationships, and look into trends. Unlike qualitative research, we do not explore phenomena. When building purpose statements, we use action verbs such as “describe,” “explain,” “look into,” “compare,” determine a relationship,”

These action verbs tell us to either test an idea, compare groups, explain the relationship among variables, describe attitudes, opinions, etc. Our samples are large (minimum 30 people—the larger our sample, the more significant our findings are), and we collect numerical data (meaning—numbers using instruments like surveys, etc.). The purpose is to get a significant yes or no.

Once a purpose statement is described, you need (a) central question(s). Examples of research questions start with question words including “how,” “what,” or “why” because we want to describe, compare, or relate (Creswell and Guetterman, 2019). You can also set up a hypothesis (null hypothesis or alternative hypothesis) if you have a strong hunch what you expect to find, ensuring it is specific, measurable, and relevant to your field of study (Creighton, 2001; Sheperis, et al., 2010).

4. Research Design (What quantitative design is appropriate for your research plan?)

Once you know the purpose of your research, selecting an appropriate research method and design that aligns with your research question and objectives becomes easier. Since we are tackling quantitative research, while there are many quantitative research designs, common research ones in quantitative research include experimental (such as causal comparative) design, quasi-experimental design, correlational design, survey design, and observational designs.

Example of how to get started:

Let’s assume that we want to find out if there is a significant relationship between quality sleep and job performance. The two variables to consider include quality sleep and job performance. How do I go about selecting my quantitative research design? Since I intend to look into relationships, I would select the correlational design. Why? Because correlational design looks into relationships from two aspects, one aspect is the directions of the relationship (positive or negative) and the other aspect is the strength of the relationship (strong or weak). My research questions would be: Is there a significant relationship between quality sleep and job performance? Now if I have a strong hunch that quality sleep affects academic achievement in a positive way, I can set up a hypothesis: There is a significant relationship between quality sleep and job performance. If I am not to sure, I would stick with my initial question: Is there a significant relationship between quality sleep and job performance?

5. Data Collection (Who or what is your population and how will you collect data?)

The most fundamental step in data collection is to get consent from the institution and population you identify. Your data collection method depends on the nature of your purpose, research question and research design. What population are you interested in? What will your sample size be? Do you need consent/assent letters? How will you collect data? Will you collect data by means of surveys, experiments, observations, or secondary data analysis?

Population/Sample (How many participants do you need?)

Once we know what design, what central question(s), we need to determine our sampling strategy and sample size based on the population of interest, ensuring it is representative and adequately powered to detect the effects of interest. Quantitative sample numbers are large compared to qualitative sample numbers. While in qualitative research, you can select one or two participants, in quantitative research, your sample needs to be large (minimum 30) to be significant. The larger the sample, the more reliable the findings are. But how do you go about selecting your sample from the population of interest? Here are a few common ways to do it: 1) select your sample based on a purpose (purposeful sampling); 2) select your sample based on convenience (convenient sampling); 3) select your sample randomly (randomized sampling). There are many more sampling methods but we just need to know the common ones at this point.

Let us use the variables we used before—we are interested in quality sleep and job performance in a large company. In this particular company, there are 7,342 employees (population). Based on our research plan, we need to do two things: 1) determine our sample size; and 2) method to select (sampling) our participants (sample) whether purposeful, random, or convenience sampling.

Once we collect our data, we need to decipher these data to make sense, which is all about our findings.

Note: Steps 6 (Data Analysis) and 7 (Reporting) to this process of conducting a quantitative research study are located in the next chapter.