Using Quantitative and Qualitative Variables in Science Experiments

In a scientific experiment there are a number of terms that must always be accounted for. Other types of scientific projects exist, which employ observation and sometimes participation, are better termed case studies and ethnographies. A true experiment must have a few basic items which we should define up front. These are the major items that need to be prominently displayed for a science fair project.

*An hypothesis is a prediction of what will happen to X when Y is done to it. Sometimes this prediction is an extension or test of a theory, but other times it is just a practical question that needs answering.

*An independent variable is the Y, or the treatment received by the experimental group. Examples of Y include adding fertilizer to one plant and not another, giving students less homework, or using essay tests instead of multiple choice. There could be several independent variables, if you were measuring the effects of different kinds of fertilizers on different experimental groups for instance.

*A host of dependent variables always exist and must be accounted for. One or more of these variables will be measured to determine the effect of the independent variable.

*Finally, controls are needed to make sure the dependent variables being measured are only being influenced by the treatment and nothing else. For this reason, control groups are established when possible. A control group receives everything the experimental group(s) receives except for the treatment. An experiment is further controlled through random selection of group members.

Qualitative and Qualitative Measure of Variables

Measurement can be tricky. One example might be an experiment to determine whether adding extra fertilizer (above manufacturer’s suggestions) helps plants grow larger or faster. We will need to measure time, and that is easy because time is standardized. We will also need to measure growth. Generally we think of growth in terms of height and weight, though gardeners also think of it as time to harvest AND maturity. This presents us with choices over the dependent variable to measure.

Height would be problematic in many circumstances. The height of a plant doesn’t tell you whether it is bushy or has just a few leaves compared to the others. Weight would also be problematic, because we would have to kill part of our experimental and control group every time we measured. The time to harvest sounds better, easily managed, but it tells us nothing about the actual harvest (like did the produce have the same nutrients, size, etc.). Also harvest is not usually an isolated event, it continues for several weeks or more in tomatoes. Time to maturity, a purely quantitative variable, might work best, but we need to define maturity as when the first blossoms form or when the first flower-buds appear.

To understand qualitative variables, consider plant maturity. As with measuring time to harvest, noting when the first flowers appear does not tell us how healthy those flowers are or whether there are more or less relative to the other plants. A chemist or biologist might take samples from the plant to analyze their chemical composition in order to quantitatively measure health. Since students most often do not have this kind of knowledge or access to tools, we rely on descriptions of plant health and vigor. Health and vigor then are qualitative variables.

Qualitative variables require as much rigor as quantitative ones. We can even make them quantitative by using a scale of 1-10, if we have some experience with the plants and know what to expect. Otherwise the groups can be described in terms of our definition for health and vigor. Number of flowers, whether flowers form properly, density and color of foliage, and other variables can be observed and reported scientifically without the use of quantitative measures.

Which Is Best?

Laboratory sciences generally rely on quantitative measure of variables almost exclusively. Recording numbers using standardized measuring procedures is preferable, because it removes researcher bias. If you know anything about statistics though, you have an understanding that numbers can and are manipulated by researchers all the time.

Imagine a phone survey. It seems bias free on the face of it. You ask people a series of multiple choice questions and record the numerical responses. There exists however a sampling error in that not everyone has a residential phone. Also there is a potential for bias in the questions. “Do you like chicken nuggets?” is a much different question from “Would you prefer the school serve chicken nuggets once a week?” Quantitative measurement is not a sure-fire way to test the hypothesis reliably.

The social sciences, educational research, and many other endeavors make use of quantitatively measured variables, especially when research questions or the environment do not allow for a proper control group. Say we performed a study to determine whether regular museum visits helped children get better grades. The grades we can measure quantitatively. But then we want to go further and ask, “Why did regular museum visits help children get better grades?” We can makes some predictions, like hands-on experience helped solidify the concepts for students, discussing their work with museum staff kept children excited about the subject, etc. In order to test these, we will need to interview teachers and students and perhaps provide them with a survey, which would include an open response section in case there were reasons we did not think of.

Mixed Method Conclusion

Most studies that contain a qualitative measure will also contain quantitative ones. This is called mixed method research, and it is common in the social sciences. The museum study above would be a good example. Science really is an art, because researchers often have to get creative in order to find meaningful answers to their questions.

As you can see, qualitative and quantitative refer to measurement and not variables. A variable such as creativity seems purely qualitative, and we could certainly measure it with descriptions of a child’s activity and use of play time. However, creativity is a construct, meaning we need to have some kind of definition for it to be testable. How we define a construct, which is just another name for a dependent variable, will lead us toward the appropriate qualitative and/or quantitative measures to employ.