By Chantelle Brule, Data Analyst, MediaMiser
So you want to know how your company was reported on in the news over the course of a year? Then you likely also want to know how to tag these news articles based on trends, themes, issues or other elements that are important to you.
Without tagging (otherwise known as coding or categorization), a collection of articles is simply a gigantic mass of content that’s difficult to manage or summarize. Tagging helps one understand content and makes it easier to discover key trends.
Tagging can be defined as a process of labeling categories, themes and ideas within content. These tags can then be utilized by researchers using media analysis software like MediaMiser Enterprise to uncover findings about a topic. Examples of things you can discover through tagging and data analysis include:
- The top ideas or themes within a topic
- The top stakeholders within a topic (if you’re a law firm, for example, you could track which practice area gets the most coverage)
- The tone of articles on a given subject or about a specific company (ie. positive/negative)
- Key spokespeople and their level of influence
Depending on your software and topic, there are a number of ways to tag content. Here are some best practices to ensure your findings will always be accurate:
1. Get to know your topic
If you’re interested in looking into a topic, chances are you already know something about it. Perhaps the topic is your company itself or a topic of interest within your industry. This existing knowledge may help you create your categories, but make sure to do some additional preliminary research. Regardless of your level of expertise, it’s possible to miss out on a key theme or category.
Say, for instance, another industry is also concerned with your topic but uses different keywords when describing it. If these terms aren’t included when both gathering content and setting up categories, you may miss relevant content or risk miscategorizing an article. A quick search on a reputable search engine is likely all you need to make sure you’re on track.
2. Narrow your field
This is more of a media monitoring issue, which we have covered in the past, but due to the abundance of online content any general topic search will likely give you thousands of results. It’s unlikely you’ll view and tag all this content. Instead, you need to know what information will be of most significance to you and your research. You can then narrow your search to these results. Here are some questions to ask yourself that may help narrow your search:
- Is there a specific question you want answered?
- Do you only need content from a specific date range (ie. after a big event or over the course of one year)?
- Are you only concerned with content from a specific country?
- What publications do you want to look at?
- If your research is related to your company are you concerned about certain competitors and not others
3. Create a tagging system
Once you have a clear idea what you’re looking for, you can now create your categories.
How you go about creating these from a technical perspective is largely defined by the system you’re using. MediaMiser Enterprise allows for the automated tagging of content based on a set of keywords assigned to each category you’ve created. Because not every category is conducive to this sort of automation, however, Enterprise also allows for the manual tagging of articles as well.
Other data analysis software (SPSS for instance) requires the manual tagging of data. To do this you must first assign number values to your keywords, because the computer software cannot recognize words. You then have to enter these numbers manually in place of your keywords.
Here is an example:
Say I’m researching the Rio Olympics and one of my categories is sports. Within this category, I want to further sub-categorize or separate my coverage based on the specific sport mentioned. To do this, you’d create a coding scheme like the following:
- 1 = Swimming
- 2 = Track3 = Gymnastics
- 3 = Gymnastics
When I later run my data, the system will determine how many times I tagged something 1, and I can infer from this how many times swimming was mentioned.
4. Get on the same page
Once you’ve determined your tagging system (the categories and keywords you’ll use) you must ensure that everyone involved with the project understands the coding system you’ve devised. Large research projects with heavy manual tagging will often employ a number of people to tag content, and it’s absolutely crucial to ensure that all involved are tagging content using the same methodology. This is sometimes referred to as intercoder reliability.
This is particularly important for categories that are not necessarily clearly defined, or may have some grey area. When tagging for tone, for instance, some researchers may tone an article positive only if it explicitly states something good about the topic. Others, though, may tag as positive anything that mentions a company and not attributing anything negative to it.
Inconsistency in regards to this will almost certainly make your findings inaccurate.
Providing a thorough tagging system, including definitions for each category or keyword, should be the first step. You should also consider running a test with a sample of your articles, which involves giving coders a small sample of content to categorize. The categorization will then be compared, and if this test goes well it should provide those involved with a better understanding of the tagging system, as well as the changes they have to make in their tagging of the content.
5. Get down to business
After the previous steps are carried out it is now time to tag the content. Depending on your specific project or timeline this may happen at different intervals. Some companies or monitors tag their content on an ongoing basis and then run their reports at the end of specific time periods. Others may collect their data all at once and tag it then. Whatever the timeline it is important to realize that tagging is time consuming and at times tedious. Make sure to set realistic goals and to split up the work in a way that ensures these goals can be reasonably met.
The exact method for carrying out each of these steps is dependent on what a researcher or organization wants to find out from their research. But for those who want accurate data, it is essential to follow these general rules for tagging.