“Ann, should my bar charts be horizontal or vertical?” The vertical-column-chart-or-horizontal-bar-chart question is one of the most common questions I receive about bar charts. My answer: It depends on what type of variable you’re graphing. If you took a research methods or statistics class back in college, then you might remember learning about terms like *nominal, ordinal, interval,* or *ratio* variables. Some of these variables are better suited to vertical column charts while other variables are better suited to horizontal bar charts. Let’s look at a few examples.

# Use Horizontal Bar Charts When You’re Graphing Nominal Variables

*Nominal variables*—favorite ice cream flavors, types of organizations where conference attendees are employed—can be arranged in any order. I arrange nominal variables in a list from top to bottom, which means their bars are horizontal. You can sort nominal variables from greatest to least or from least to greatest. Both are correct. My advice: Sort the data so that the item that warrants attention is displayed first—a high number deserving a celebration, or a lower-than-hoped-for number that needs to be turned around.

Use horizontal bar charts to display nominal variables like favorite ice cream flavors or employment settings.

# Use Vertical Column Charts When You’re Graphing Ordinal Variables

*Ordinal variables* follow a natural progression—an order. Another name for ordinal variables is *sequential variables* because the subcategories have a natural sequence. I arrange ordinal categories from left to right so my viewers can view the sequence across the page, which means their bars are vertical.

Use vertical column charts to display ordinal variables like age ranges, salary ranges, and even cohorts or graduating classes (e.g., the percentage of students from each graduating class who achieved *x* outcome).

# Real-Life Examples

Okay, you’ve seen some fictional, generic examples. Let’s look at a few examples of horizontal bar charts and vertical column charts from real projects.

## Compare Age Ranges with Vertical Histograms

While working with a museum, we wanted to display some key demographic characteristics about people who had responded to the museum’s survey. Age ranges are ordinal, so we used vertical column charts to visualize how many people fell into each age bracket.

While working with public health researchers, we needed to visualize how many males and females in each age range were diagnosed with x disease. We built stacked column charts with vertical columns. We also used a population pyramid to compare the distributions of males and females.

## Compare Cohorts of Survey Respondents with Vertical Column Charts

In the *State of Grantseeking* project, an organization called GrantStation sends out surveys a couple times a year, in the spring and fall. We wanted to compare how different iterations or cohorts of survey takers responded. In other words, we wanted to see whether people who responded to the spring 2016 survey were any different from people who responded to the fall 2016 survey or the spring 2017 survey. Cohorts are ordinal, so we used vertical stacked column charts to depict the proportion of survey respondents who are employed in nonprofits.

## Compare Patterns Over Time with Vertical Charts

While working with a hospital’s analytics team, we needed to display the proportion of hospital procedures (“Product A” and “Product B”) that were performed by this hospital (“the ABC Org”). We wanted to compare two points in time, 2012 and 2016. Time is ordinal, so we used vertical columns.

## Visualize Agree-Disagree Scales with Horizontal Stacked Bars

While working with a museum, we needed to display how many people agreed or disagreed with each statement on a survey. Agree/disagree scales are ordinal. (Well, technically, agree/disagree scales are a special type of ordinal variable called a *diverging variable*.) We wanted our readers to see the natural progression from *agree* to *disagree*, so we used horizontal bars to show that progression from left to right across the page.

## Compare Before-After Survey Responses with Vertical Column Charts

In this example, we needed to make a one-page handout that summarized whether program participants were more knowledgeable about New England history, Martha’s Vineyard, or whaling after completing an educational program at a museum. The museum administered the survey twice, at the beginning of the program (pre) and at the end of the program (post). Pre-post comparisons are comparisons over time… and time is an ordinal variable… so we selected vertical columns. But wait! There’s more! We were graphing two sets of ordinal variables: 1) the timeframe (pre or post survey administration) and 2) the scaled survey responses (very, moderate, somewhat, slightly, or not at all). We chose to give the timeframe more weight than the scaled survey responses. In other words, we decided that the pre-post timeframe was the most important, and used vertical columns first and foremost. Then, within those columns, we displayed the ordinal survey responses.

These are guidelines, not rules. If you can explain the logic for going against this guidance, then your graph is probably going to be alright. My goal is to develop critical thinking masterminds, not robots.

# Bonus: Download the Materials

Download the Excel spreadsheet that I used to make the generic bar and column charts at the top of this article.