Introduction

codeit applies a sentiment score to each of the themes it finds within a verbatim. For example, if a particular portion of a verbatim is talking about "service", codeit will assign that section to the "service" theme but also assign a sentiment score to that portion. Since each verbatim can mention a number of themes and these can be positive, negative or neutral, this results in a lot of individual data points.  


What's the score?

The obvious question then is "what's the overall sentiment score?" - it's tempting to look for a "one number" summary that summarises the overall sentiment expressed in your data.  Although this is possible, simply by calculating an overall average, we need to be cautious about this kind of statistic. For example, an overall average might come out somewhere in the middle, hiding the fact that opinion is polarized into strong positive and negative reactions. Or, a positive average score might disguise a subgroup of extremely dissatisfied people and therefore be misleading.  


The "scores" visualisation provides an overall summary sentiment score, but also a set of additional descriptive statistics that illustrate other dimensions to provide a more rounded picture of the sentiment within the data. 

List of scores available

The summary statistics available on the "scores" visualisation are as follows: 

  • Sentiment Score
    An overall sentiment rating across your entire data set in the range -10 (strong negative) to +10 (strong positive) where 0 is neutral or balanced sentiment

  • Strong Positives
    An overall score across your entire data set from 0 (there is no strong positive sentiment in any of the responses) to 10 (all responses have strong positive sentiment)

  • Strong Negatives

    An overall score across your entire data set from 0 (there is no strong negative sentiment in any of the responses) to 10 (all responses have strong negative sentiment)

  • Polarization
    An overall score across your entire data set from 0 (the respondents are in agreement) to 10 (the respondents are in disagreement)

  • Indifference

    An overall score across your entire data set from 0 (all responses have positive or negative sentiment) to 10 (no responses have sentiment)


  • Extreme Reaction

    An overall score across your entire data set from 0 (there is no strong positive or negative sentiment) to 10 (all responses have strong sentiment)