Different Types of Variables
Every object of study has its unique attributes that are called variables in statistical research. While composing an experimental design, be attentive about the choice of which variables to measure for obtaining the best results.
You want to research the influence of excess moisture on the development of different species of plants. Every species has its variables, such as the growth speed, or wilting. You need to identify those variables that will show the influence of water. They should also answer the questions of what type of useful data they provide and what part of the experiment the variable is used for.
Quantitative and Categorical Types of Variables
A variable is measured in data - the special value that is usually recorded. The data can be of two categories - quantitative one that shows amounts and categorical one that shows the signs for grouping.
- If a variable has quantitative data in it, it is a quantitative variable.
- If it represents categorical data, it is called categorical one.
These types are further broken into some subgroups.
Types of Quantitative Variables
Quantitative data is represented in numbers that can be added, multiplied, or subtracted. Quantitative variables can be of two types - continuous and discrete.
- Continuous (or ratio) variables represent the measurements of non-finite values, for example, volumes, distances, or ages.
- Discrete (integer) variables stand for counts of values or individual items, for example, the number of tree species or students in class.
Types of Categorical Variables
They are used for representing groups. Rarely, they can be recorded as numbers but most often, they represent categories but not the amounts. You can use three types of categorical variables - nominal, ordinal, and binary ones.
- Nominal variables represent groups without any ranking or ordering within the group, for example, names of species, brands, or colors.
- Ordinal variables stand for the groups that are ranked inside in a specifically defined order, for example, rating scales in surveys or the starting and finishing places in any race.
- Binary variables represent the presence or absence of the outcomes, for example, victories and losses in some game.
Remember!
Sometimes, a variable can behave as a combination of two types. For example, in numeric scales, an ordinal variable can be considered a quantitative one as well. If you need to assess the star rating in customer reviews of a certain product, the number of stars from 1 to 5 is ordinal, though the average result we can see is quantitative.
Independent and Dependent Variables for Different Parts of an Experiment
You conduct experiments to define the connection between variables or the effects they have on each other.
- Independent variables are those which, as you think, can be a cause.
- Dependent variables are those which, as you think, can be an effect.
Independent variables are also called treatment ones. You manipulate them to influence the results of an experiment, for example adding different amounts of water to the pots with different plants. Experiments are usually designed to find out what effect one variable has on another – in our example, the effect of salt addition on plant growth. However, the outcomes of the experiment are represented by dependent or response variables, for example, any measurements of the plants received under the influence of water.
You can also have constant or control variables that are kept unchangeable within the entire course of the experiment. For instance, they are the temperature and lighting in the room where the plants grow or the volume of fertilizers given to each plant.
Other Types of Variables
To choose the right statistical test, you may need some other variables apart from independent and dependent ones or categorical and quantitative. Let’s consider such types of variables as confounding, latent, and composite ones.
Confounding variables can hide the actual influence of some other variable in your experiment. It can do it if it is closely connected with the variable you are researching but you have not paid much attention to it during your experiment. For example, if we speak about the influence of moisture on plants, we can see that the size of pots and the type of soil can also affect the growth. You need to control these variables as well to get reliable results.
Latent variables can be represented via a proxy but you cannot exactly measure them. The tolerance to excess moisture cannot be measured in plants but it is sure to affect the plant growth.
Composite variables are a combination of all the variables used within the experiment. They appear when you start the data analysis but not when you measure something. If you investigate the influence of excess moisture on three different species, you can combine the received scores and results to present your generalized findings.
Final Thoughts
Now you can see that the correct choice of variables, which are the attributes of the sample or individual objects, can have a crucial effect on the results of the experiment. To make these results more distinct and reliable, you need to measure and analyze different types of variables according to the needs and stage of the experiment. In this case, the research will be successful and produce convincing results.