# Step-by-Step Guide 6:

# Descriptive Statistics (Tables & Plots)

## Overview

What you will learn here:

Obtain descriptive statistics for continuous and categorical data

Generate tables that contain descriptive statistics

Generate a simple Frequency distribution

Generate basic descriptive plots for continuous and categorical data

## Dataset & Variables used for examples

Skoczylis, Joshua, 2021, "Extremism, Life Experiences and the Internet", https://doi.org/10.7910/DVN/ICTI8T, Harvard Dataverse, Version 3.

### Variables:

Share_Nothing_Society - Nominal Data

Political_LeaningRight_Left - Continuous Data

Gender - Nominal Data

## Descriptive Statistics Step-by-Step video

## Descriptive Statistics: Variables and plots

1.

Get Descriptive Statistics

Navigate to Analyses > Explore > Descriptives

Now, select your variable(s) and drag them in the empty Variables box.

If you want results broken down by another categorical variable, select it and drag this into the Split by box.

2.

Select your Descriptive Tests

Select how to display your data tables. You have got the following two options:

Variables across columns

Variables across rows

If your data is categorical select the Frequency Table.

A frequency table will be generated. Note, if you split your ordinal/nominal variable by another it will no longer display percentages (you can get this by using a contingency table).

You can now select the relevant descriptive statistics in the Statistics section

Select the variable you want to get descriptive statistics for and then select the test you want. Note, you can also split your variable by another ordinal or nominal variable, e.g. you might want to see the data split by Gender.

Important: What descriptive statistics you select will depend on the type of data you have. If your data is categorical you should only select the option in Sample Size and the Mode under Central Tendency.

3.

Select your Descriptive Plots

To generate a basic descriptive plot navigate to the Plot section.

The type of plot you generate will depend on the type of data you have. If your data is continuous you can select Histograms, Box Plots and Q-Q Plots.

Important: For ordinal/nominal data you should only select a Bar Plot. The Bar Plot produced here is not great if you have long labels.

More descriptive plots are available for all data types using the surveymv and JJStatsPlot modules

## Continuous Data: Results

1.

Results: Descriptive Tables

Once you have selected your descriptive statistics, a table will appear in the output window on your right. Below we used the variable Political_LeaningRight_Left. Here we displayed it with the Variable in the row and the statistics in the columns.

The table displays all of the relevant descriptive statistics for this variable. The first table gives you the results split by Gender

The second table below gives you the results for everyone. Notice how the mean is different? This is the mean for the entire data rather than broken down by another variable (in this case gender).

3.

Results: Descriptive Plots

As noted above, for continuous data you can select Histograms, Box Plots and Q-Q plots.

Below is a Box Plot for the Political_LeaningRight_Left variable. A boxplot shows you how your data is distributed. Essentially a Box Plot has the following elements.

The maximum and minimum value - this is indicated by the bottom and top of the Whiskers (the horizontal line running through the box.

The box - this indicates the middle 50% of your data. The bottom of the box represents the 25th percentile, the dark line through the box represents the median (50th percentile), and the top of the box represents the 75 percentile.

In the example below we have also added scatter to indicate how the data is distributed.

The little square in the box represents the mean.

This particular plot is split the variable output by gender - hence two box plots.

Using the same variable we have also generated a histogram. A histogram shows you how the date is distributed. The higher the bar the more observations there are. Again, we have two graphs one for each gender.

Finally, we have also generated a density plot. A density plot is essentially a smooth version of the histogram.

## Ordinal/Nominal Data: Results

1.

Results: Frequency Tables

Jamovi generates two easy to interpret tables. The first table provides you with descriptive statistics such as the number of observations (N), how many missing values there are and the Mode (most frequently selected option). When summarising categorical data, these are really the only descriptive statistics that are useful.

The Frequency Distribution tells you how the data is distributed. It provides you with the count, percentages and cumulative percentages.

In the table below we can easily identify how participants responded to the statement 'I share nothing with society'. Frequency tables are easy to interpret and tell you how your categorical variables are distributed. Where possible always use percentages rather than the counts.

Note:

Jamovi will generate a new Frequency table for each categorical variable you select. You can split your categorical variable, however, if you do, Jamovi will no longer show you the percentages. To get this information you will need to generate a contingency table (more on that in later tutorials).

2.

Results: Descriptive Plots

Below is your Bar Plot. These plots provide you with the same information a frequency distribution does, but the output is displayed visually.

Unfortunately, Jamovi does not allow you to change any settings and the bar plots are somewhat basic (they don't even display percentages).

Survey Plots, which we cover in the next tab, provides you with more flexibility and allows you to create more functional plots.

Notice, that the labels all overlap - this does not look great. Hopefully, this is something that Jamovi will fix in the future.