Sentiment analysis matters. That’s why last time we looked at three good reasons why, now we’ll look at the final reason and sum it up.
You can find out why sentiment analysis matters, otherwise read on for the final reason and summary.
4. Get an idea of the likelihood of acceptance
Let’s make an assumption – that, at heart, most writers of persuasive writing want you to accept what they are saying. It goes with the name, you wish to “persuade”.
No one likes criticism, or other forms of disagreement – which may be why so many dictators lock up their critics.
I’ll be honest, one of my aims is that, due to the supplied evidence, you will agree with me – ideally whole-heartedly, or at least largely. If we assume – and this is a big assumption – that newspapers are a form of persuasive writing on the assumption then at heart they want to persuade you to keep reading.
If the theory of the filter bubble is true, that we tend to seek out opinion we agree with, then papers are not only going to want to commission work you agree with, but present it in a way you agree with. They want to be part of your filter bubble and secondary to writing content you want to read it is leaving you in an emotional state.
Films are excellent ways of getting an emotional kick – the highs and lows Asa film progresses gives you an emotional fix.
As I wrote recently, Gravity is a recent good example of a film that takes you high as safety seems reached, and low again as the next peril strikes Sandra Bullock. Other writing is like this. Books too tend to have this, but what about shorter pieces?
Newspaper articles, speeches, blogs, all may be short but that is no reason why they can’t pack an emotional punch, and, better still, have those highs and lows that films have.
I will be looking at different forms over the next few months, but this is where word and sentiment analysis can come in. Orations may be deemed a ‘good speech’ and we can pick out some phrases that explain why.
Martin Luther King’s solid imagery of “Stony Mountain” in his “I have a dream speech” is one example. Or Shakespeare’s Henry V who appeals to the few, his band of brothers.
But how about the overall sentiment and the flow? I believe that while picking out phrases is useful, or a line-by-line breakdown, sentiment analysis can give a quick snapshot and help us know where to look for key phrases and their juxtaposition.
Combined with the other points I mentioned, particularly how the audience is addressed (is it you or we, us and them), we can work out the likelihood that readers will accept the piece.
We can also use it for other exercises, such as in my comparison of Dawkins’ and Jesus’ words to see if they hit all levels of the Maslow pyramid and perhaps give one reason why 2,000 year old text is still relevant.
Newspapers can – and do – do this in the most obvious way by hiring or commissioning writers who already are known for the association – Polly Toynbee at the Guardian, Richard Littlejohn at the Daily Mail – then go by topics. The Mail’s take on Mark Duggan, whose death sparked the 2011 London riots, was that of ‘The ‘gangsta’ gunman whose death sparked riots’ while the Guardian went for a farewell from a close friend of him Maslow pyramid.
Sentiment analysis is another tool to add to analyse of a text rather than a replacement for traditional methods. However on a larger scale we can make some generalisations about type of content or publisher.
Much like height measurements can be added to a map to give us contour lines and so add to our understanding, and also allow us to make assumptions about larger areas, so sentiment and word analysis be used in the same way.
The main point for writers is – will this help my writing be remembered? Yes, and it does not mean compromising your style, rather being aware of how you write.
According to the research I’ve looked at, you can be remembered for being negative or positive, write in the first-, second-, third-person – or not at all- as much as you wish.
Sylvia Plath’s later writing showed she was suicidal and this hasn’t stopped her fame, though it is doubtful a suicidal person was after fame. What we do want to know are the nuances – with Plath is there a a limit or optimum amount she could write about herself, or on negativity before it affects her readership?
My research into figures at the office (which means I can’t share them, much as I’d like) suggest there is. For example, addressing readers as “you” increases readership receptiveness to a point and then declines as it becomes overused.
Likewise is there a pattern? I have looked at newspapers and seen it and have studied some great speeches from fiction and real life and analysed the ebb and flow of pronouns and emotion and there are some surprising results.
I will publish these over the coming weeks but for now is there anything you can do? There is. The most important thing is to know your audience. One thing consistent we the papers I have read and with research into sites is that it comes down to knowing who you are writing for.
All the papers make a point of saying what the situation is where the words are taken from – there is always some relationship between the author and the speaker and their audience and how this changes changes their style.
Look at your own writing. The primary edits must be for spelling, structure and so on, but sentiment analysis can help you get a feel of how readers will receive it. If you always talk about yourself and you have a large audience you are fine, but what if you talk about others more, how will they react to a piece that is all about you?
As this clip from the Office shows that ultimately your audience sees you as being there for them, they don’t particularly care about you but how you can benefit them. As this Cracked article also shows.
Next time – the tools available.