All research of a scientific nature is supported and based on a set of data duly analysed and interpreted.To reach a point where we can extract relationships of causality or correlation, it is necessary to observe multiple observations so that the existence of the same relationship in different cases or in the same subject over time can be falsified and verified. And once these observations have been made, aspects such as the frequency, average, fashion or dispersion of the data obtained must be taken into account.

In order to facilitate understanding and analysis both by the researchers themselves and in order to show the variability of the data and where the conclusions come from to the rest of the world, it is very useful to use visual elements that are easy to interpret: graphs or charts.

Depending on what we want to show, we can use different types of graphics. In this article we’ll look at different types of graphs that are used in research using statistics.

The graph

On a statistical and mathematical level, called graph a that visual representation from which can be represented and interpreted generally numerical values. Among the multiple information that can be extracted from the observation of the graph, we can find the existence of a relationship between variables and the degree to which it occurs, the frequencies or the proportion of appearance of certain values.

This visual representation supports the display and understanding of the data collected during the research in a synthesized way, so that both the researchers who carry out the analysis and others can understand the results and it is easy to use it as a reference , as information to be taken into account or as a point of contrast when carrying out new research and meta-analysis.

Types of charts

There are many different types of graphs, generally applying one or another depending on what you want to represent or simply the author’s preferences. Below we indicate some of the best known and most common.

1. Bar chart

The best known and most used of all types of charts is the bar chart. In this one, the data is presented in the form of bars contained in two Cartesian axes (coordinate and abscissa) that indicate the different values. The visual aspect that indicates the data is the length of these bars , not being important their thickness.

It is generally used to represent the frequency of different conditions or discrete variables (for example, the frequency of different iris colors in a given sample, which can only be specific values). Only one variable is observed in the abscissa, and the frequencies in the coordinates.

2. Pie chart or sectorial chart

The also very usual graph in the form of a “cheese”, in this case the representation of the data is carried out by dividing a circle into as many parts as values of the variable being investigated and each part having a size proportional to its frequency within the total of the data . Each sector will represent a value of the variable being worked with.

This type of graph or diagram is usual when the proportion of cases within the total is being shown, using percentage values (the percentage of each value) to represent it.

3. Histogram

Although at first glance very similar to the bar graph, the histogram is one of the most important and reliable types of graph on a statistical level. On this occasion, bars are also used to indicate through Cartesian axes the frequency of certain values, but instead of simply establishing the frequency of a specific value of the evaluated variable, it reflects a whole interval. A range of values is therefore observed, which could also reflect intervals of different lengths .

This allows us to observe not only the frequency but also the dispersion of a continuum of values, which in turn can help to infer probability. It is generally used in the presence of continuous variables, such as time.

4. Line graph

In this type of graph, lines are used to delimit the value of a dependent variable from an independent variable . It can also be used to compare the values of the same variable or of different investigations using the same graph (using different lines). It is usually used to observe the evolution of a variable over time.

A clear example of this type of graph is the frequency polygons. Their operation is practically identical to that of histograms, although they use points instead of bars, with the exception that they allow the slope between two of these points to be established and the comparison between different variables related to the independent one or between the results of different experiments with the same variables, such as the measurements of an investigation with respect to the effects of a treatment, by observing the data of a pre-treatment and post-treatment variable .

.

8. Scatter plot

The scatter plot or xy plot is a type of graph in which all the data obtained by observation is represented in the form of points on the Cartesian axes. The x and y axes each show the values of a dependent variable and an independent variable or two variables of the one being observed if they are related in any way.

The points represent the value reflected in each observation, which will visually show a cloud of points through which we can observe the level of dispersion of the data.

You can see whether or not there is a relationship between the variables by means of the calculation. This is the procedure that is usually used, for example, to establish the existence of linear regression lines to determine whether there is a relationship between variables and even the type of relationship that exists.

9. Box and Whisker Chart

Box plots are one of the types of graphs that tend to be used to look at the dispersion of data and how the data are grouped. The starting point is the calculation of the quartiles, which are the values that p allow the data to be divided into four equal parts . Thus, we can find a total of three quartiles (the second of which would correspond to the median of the data) that will configure the “box” in question. The so-called whiskers would be the graphic representation of the extreme values.

This graph is useful when evaluating intervals , as well as observing the level of dispersion of the data from quartile values and extreme values.

10. Area Chart

In this type of graph, the relationship between dependent and independent variables can be seen in a similar way to what happens with line graphs. Initially a line is made joining the points that mark the different values of the measured variable , but everything below is also included: this type of graph allows us to see the accumulation (a given point includes those located below).

Through it you can measure and compare the values of different samples (for example, compare the results obtained by two people, companies, countries, by two records of the same value…). The different results can be stacked, and the differences between the various samples can be easily observed.

11. Pictogram

A pictogram is a graph in which, instead of representing data from abstract elements such as bars or circles , elements specific to the topic under investigation are used . This makes it more visual. However, its operation is similar to that of the bar graph, representing frequencies in the same way

12. Cartography

This graph is useful in the field of epidemiology, indicating the zones or geographical areas in which a certain value of a variable appears with greater or lesser frequency. The frequencies or frequency ranges are indicated by the use of color (a legend is required to understand them) or size.

Bibliographic references:

  • Martínez-González, M.A.; Faulin, F.J. and Sánchez, A. (2006). Bioestadística amigable, 2ª ed. Diaz de Santos, Madrid.