After completing the analysis of a comparison, Relative Insight will generate an output of the statistically significant findings. Users are presented with the results in the comparison explorer, from which the most relevant discoveries can be bookmarked onto insight cards.
The findings of the analysis are split across three primary categories - differences, frequencies and similarities.
Differences are linguistic features (topic, grammar, phrase, word, emotion) that are statistically more prevalent in one data set compared to another. They are particularly useful for their ability to help you understand what makes an audience group, product or brand unique from others.
Differences are expressed using the relative difference metric. Considering the example shown below, the 36.5x relative difference would be interpreted as "British gamers are 36.5x more likely to use the phrase 'limited edition' when discussing video games".
Frequency is a measure of how common a particular linguistic feature is within a data set. It is expressed as a percentage of the total word count. Frequencies are displayed for all linguistic features present in a data set.
Frequencies can help you understand how often a particular linguistic feature of interest is used. For example, when conducting a product comparison of Bluetooth headphones you might be interested to know the frequency of the words 'battery' or 'comfort'.
Similarities are the linguistic features that occur with a similar prevalence in the data sets being analysed. They are indicative of commonalities between data sets and can help you understand where products, brands or audience groups share similar characteristics.
Similarities are identified where a relative difference between 0.9 and 1.1 is determined and it does not meet the threshold for classification as a difference. 'Function words' such as if, the, and, but etc. are removed given they occur with a high frequency in every data set.