Skip to main content

Grammar analysis in Relative Insight

Learn how to interpret the grammatical analysis in Relative Insight and explore common applications of grammar insights.

Trish Pencarska avatar
Written by Trish Pencarska
Updated over 2 weeks ago

Relative Insight’s comparative text analysis software surfaces differences and similarities in topics, phrases, words, emotion and grammar. To understand grammar, Relative Insight analyzes the context and word endings in order to classify each word into pre-defined categories.

While topical, phrase and word analysis helps you understand what is being said, putting in the effort to understand emotions and grammar can provide additional insight into how people are saying it. This can reveal crucial details about how people think and feel about a particular topic or brand.

But even if you have an English degree, grammatical analysis can be intimidating. This article explores how grammar analysis can be helpful and the best way to approach it in the platform.

Understanding grammar categories

Grammar insights are easiest to understand once you are already familiar with the comparison you are analyzing. Before diving into grammar, make sure you have examined the other categories of linguistic features (topics, words, phrases, emotion) and have a strong sense of the key insights.

While the importance of a particular word, phrase or topic might be obvious, fully comprehending grammar insights often requires an openness to thinking abstractly. Here are a few of our top tips for making the most of grammar analysis:

  • Have an idea of what you are trying to get out of the grammar insights before you start (e.g. improving brand tone of voice)

  • Expand the grammar categories (by clicking on the name) to reveal the words that fall into that category. This can help overcome the challenge of complicated-sounding categories like ‘singular determiner’

  • Understand both sides of the comparison - for example, you might be hard-pressed to derive insight from knowing that one audience group uses a lot of first-person pronouns, but if another group is more likely to use collective pronouns that helps contextualise the discovery

  • Reflect on the possible reasons why the group of words might be used more or less frequently – there are no wrong answers!

Grammar category descriptions

Easy definitions for each of the grammar categories analyzed in Relative Insight.

1st person singular personal pronouns - e.g. I, me

This indicates the language of your data set is self-referential or even self-absorbed if from a personal perspective.

1st person plural personal pronouns - e.g. we, us

This indicates that the people captured within a data set present themselves as part of a unit with others.

2nd person personal pronouns - e.g. you

This indicates that the language in your dataset directly addresses others.

3rd person pronouns - e.g. she, his, they

This means that your dataset is talking about other people and is often related to politics or making judgments on the actions or behavior of others.

3rd person singular neuter personal pronoun - e.g. it

This shows that the data set is trying to disassociate itself with the subject of the text and may be trying to increase our psychological distance from the subject to make it seem more abstract.

Adjectives - e.g. old, older, oldest

This shows the data set tends to go into greater detail about the subject.

Comparative adverbs - e.g. more, less

Superlative adverbs - e.g. most, least

These categories signal that the data set is evaluating or comparing something.

Comparative after-determiner - e.g. more, less, fewer

This language is focused on quantities or amounts of something which can change or be altered, and this group might be calling for a change in this regard.

Degree adverb - e.g. very, so, too

This suggests that the data set tends to exaggerate or use hyperbole about the subject.

Interjection - e.g. oh, um, ah
This may show that the language in the set is more colloquial or informal.

Locative adverbs - e.g. there, here

This means that the language is trying to make the concept less abstract by anchoring the subject in a place or time. This reduces our psychological distance from the subject to make it seem more tangible and relatable.

Modal auxiliary - e.g. will, would, must

This shows that the data set tends to be more forward-thinking and may be looking for a change.

Numbers - e.g. 1, 1770-1827, quarter

This shows that the data set is more detail-oriented and will provide specifics instead of speculative or abstract suggestions. Can be further classified as singular cardinal numbers, hyphenated numbers, fractions or numeral numbers.

Possessive pronoun, pre-nominal - e.g. my, your, our

This shows that the data set is taking ownership of action, issue or organization, which potentially reveals a high level of engagement.

Quasi-nominal adverb of time - e.g. now, tonight, tomorrow

The language group is trying to anchor itself in a particular timeframe, perhaps stating they are socially current.

Reflexive indefinite pronoun - e.g. oneself

This shows that the data set tends to be more formal.

Superlative adjective - e.g. strongest, biggest

The group may tend to exaggerate or use hyperbole.

Unclassified word - e.g. N/A, kg/m2

This is when the data set uses words that we are not able to classify into a part of speech and may show that they are using more technical or codified terms.

Grammar analysis example use cases

Tone of voice analysis

Content writing is an integral part of any marketing function and ensuring a consistent tone of voice is foundational to establishing a coherent brand voice. Do you address your customers as you or we? Do you use superlative or fact-based language to describe your product? These questions may seem trivial, but they are essential to creating a consistent voice. Comparing new pieces of content against a collection of previous pieces can help to ensure consistency across all touchpoints of your brand.

How people experience your brand

Understanding the use of pronouns in voice of the customer data sources such as reviews, social mentions and surveys can help brands understand how people experience their products. For example, high usage of collective pronouns such as we and us may indicate that people consume your products in a group setting. This could inform brand messaging, focusing on the experience of sharing the product. Conversely, high usage of personal pronouns such as my and I could indicate that people consume your product privately.

Distinguishing facts from opinions

Grammar analysis can also provide insight into whether a body of text is rooted in fact or opinion. Heavy use of maximisers (totally, virtually, completely etc.) can be an indicator of highly persuasive language being used to make a point. On the other hand, a high prevalence of the ‘Unit of measurement’ category can indicate that a particular message is well researched and supported with stats.

This kind of analysis can be useful for both assessing how competitors position their products and understanding how to appeal to your audiences. For example, if a competitor relies heavily on maximisers, perhaps there is an opportunity to respond to that with fact-based messaging. Similarly, analyzing reviews may reveal that older customers cite hard facts while younger people are more prone to sensationalised comments – this can inform how to target advertising to each segment.

Comparative vocabulary

The use of comparative language (more, less, better, worse etc.) by consumers can serve as an indicator that an audience has a high degree of familiarity with other options available to them. This can also be an indicator that the data is a good source of competitor intelligence. Equipped with this information, marketers may decide to incorporate explicit comparative elements in their marketing communications.

Did this answer your question?