Media Research Information and Insights

The sentiment analysis of a simple red card

Written by Chrystal Glassman | Nov 7, 2024 1:44:35 AM

When you think of ‘sentiment analysis’, there are two words that probably come to mind – “good” and “bad”, and somewhere in between. But there is so much more to giving something a thumbs up or thumbs down.

Sentiment analysis can be used in an array of ways and is an important tool for organisations to help gauge the media’s and/or people’s feelings, opinions or attitudes towards them, their brand or their people. Organisations use sentiment analysis not only to understand how they are perceived but also to evaluate the performance of their media engagement, their competitors' positioning and issues management.

At Medianet Insights, our analysts specialise in delivering sentiment analysis for a plethora of clients.

When analysing sentiment, Medianet's award-winning team of expert (human) analysts approach each media mention by putting themselves in the shoes of you, the ‘reasonable reader, listener or viewer’.

By assuming the perspective, they will objectively read, listen and view each piece of content, and can conclude if something is Favourable, Neutral or Unfavourable.

To help contextualise this, I want to take you back to the 1998 Football World Cup in France (Bonjour!). The atmosphere is electric as you settle in to watch the infamous Argentina vs. England match, a clash that has become a staple of football lore. Both teams have two goals apiece, and the tension in the stadium is palpable, with fans on the edge of their seats, knowing that the game is set to go to extra time.

Suddenly, the narrative takes a dramatic turn.

Diego Simeone fouls David Beckham, who dramatically falls to the ground. The crowd erupts, a mix of gasps and cheers echoing through the stands as a kerfuffle between Beckham and Simeone ensues. Tempers flare, and before anyone can fully comprehend the chaos, the referee, Kim Milton Nielsen, raises the dreaded red card. Beckham walks off the pitch, leaving his team at a disadvantage and sending shockwaves through fans and players alike.

This pivotal moment not only changed the course of the match but also etched itself into the annals of World Cup history. In that moment - or even 26 years later - how did it make you feel then—and how does it make you feel now? And the biggest question is, how do you, the ‘reasonable reader’, think you would rate the overall sentiment of this?

If the Medianet analysts were going to analyse a piece of media like this, we would be looking at the tone and attitudes attributed to the story, the language used, as well as spokesperson input and/or active engagement.

By examining these elements, we can gain deeper insights into how the narrative is shaped and how it reflects broader themes within the tournament, such as sportsmanship, rivalry, brand impact, personal impact, national pride etc.

Understanding these nuances allows us to appreciate the complex dynamics at play (excuse the pun).

But what if I told you, there are a couple of other ways sentiment can be represented, specifically through entity-level sentiment and article-level sentiment?

What are those you ask? Well, let me give you a summary so we can “Bend it like Beckham” (sorry, sorry, couldn’t help it! Will stop now).

Entity-level vs. Article-level sentiment

Entity-level sentiment focuses on specific entities—such as individuals, organisations, or places—mentioned within a media piece. For example, in our World Cup story, we could look at the sentiment associated with David Beckham, Diego Simeone, and the World Cup itself. Each entity could elicit a different response, depending on their portrayal.

So let’s head back to the World Cup quickly and look at the different entities that are mentioned in the media. 

Do you feel the same way about David Beckham as you do about Diego Simeone?

Do you feel the same way about the coach as you do the players?

How do you feel about the 1998 Football World Cup in general?

You are now framing the questions in a different way, which might result in a different sentiment outcome. On the other hand, the Article-level sentiment is when you want to understand the general perceptions around a topic or event, to understand trends in the landscape over time. So instead of assessing sentiment based on a specific organisation, person or place, you are assessing it in its entirety.

Revisiting the World Cup one last time (I know what you must be thinking, the air miles we would be accumulating going back and forth to France would be insane), and considering all the facets of article-level sentiment - the game’s atmosphere, the foul, the players, the red card incident, the referee and the fan reactions;

How do you feel about the media commentary as a whole now?

Do you feel the same way you did before, and would you give it the same sentiment outcome?

In some cases, when looking at entity-level and article-level sentiment together the sentiment might be the same, but when you break it down into smaller, bite-sized pieces, you get a completely different outcome. 

So, when you next think of sentiment analysis, know there’s a lot more complexity to it than simply handing someone a coloured card.

Automating entity-level sentiment analysis: a Medianet first

 

With so much complexity involved in analysing entity sentiment, could this ever be automated?

 

While article-level sentiment analysis has been automated for some time, historically, entity-level sentiment analysis requires human analysts due to the nuance and expertise needed to determine the true sentiment of specific entities within an article. In a groundbreaking move, Medianet has automated this process within its media monitoring platform, thanks to rigorous AI model training and oversight from expert human analysts.

 

This world-first approach by Medianet lets you understand the nature and quality of monitored media mentions, gauge the overall sentiment of a story, and gain deeper insights into how specific entities are portrayed—all automatically and at your fingertips. Thus empowering communications professionals with deeper, automated insights to drive more informed and impactful media strategies.

 

Book a demo to learn more about this feature.

 

Want to see Medianet media monitoring and its features in action, register your interest for a free trial here.

Master media monitoring with our comprehensive guide. Learn key tips, best practices, and actionable insights to enhance your communication strategy and stay ahead of the competition.

For more insights, subscribe to our newsletter.