When compiling a written analysis of data, it is important to frame explanations in the correct way. Here are a few tips that may help.

  • If you are referring to an age group, make sure that you word it in such a way so as people can’t misinterpret it. For instance, if your age group is 50+ year olds, avoid saying “A high incidence of those aged over 50 years said that they like carrots”. Framing it like this could be misinterpreted to mean the statement applies to those aged 51 years or higher (because they are over 50 years). Instead, it should say “A high incidence of those aged 50 years or over said that they like carrots”.
  • When talking about the results of a factual yes/no question, for instance, Do you have a health care card?, be careful that you are not making assumptions about attitudes. Saying that health care cards are ‘most likely’ to be held by those aged 50 years or over sounds as though if you asked people aged 50 years or over if they wanted one they would be more inclined to say yes than those under the age of 50. It sounds like it is a choice/attitudinal decision. Instead it should say that a health care card is held by a higher proportion of residents aged 50 years or over.
  • Another common mistake is assuming that if the respondent says they have done something within a certain period, then this is a common action. For instance, a question may ask In the last 3 months, have you seen a doctor because you were turning orange from eating too many carrots?. Often this will be reported as “Half of people aged 50 years or over visit the doctor because they think they are turning orange from eating too many carrots”. It can’t be assumed that just because it happened in the last 3 months it will be a regular thing. Instead the analysis should be “Half of people aged 50 years or over visited the doctor in the three months prior to interview because they thought they were turning orange from eating too many carrots”.
  • If you are comparing two different groups, be careful about how you are framing the analysis. Lets assume that the data has respondents from across two suburbs, A and B. In the sample there are 100 respondents in suburb A and 500 respondents in suburb B. All respondents are asked if they like carrots and it was found that in suburb A 50 per cent said that they like carrots and in suburb B 20 per cent said that they like carrots. The tenancy is to report this as “More people in suburb A like carrots”. This is correct proportionally (50 per cent versus 20 per cent) but not numerically (50 people in suburb A versus 100 people in suburb B). In order to report this correctly you would need to say “a higher proportion of those in suburb A said that they like carrots”.
  • Avoid starting a sentence with a number (unless it is a dot point list). Try to use an expression, such as “Half of respondents…” instead. If you simply must start the sentence with a number, write it as text, not numerals; so “Fifty per cent of respondents” not “50 per cent of respondents”.
  • With regards to using “%”, “percent” or “per cent” in the written analysis, this will differ across organisations. It is a good idea to get your hands on the style guide for the organisation for whom the report is written to make sure it conforms to their rules.

In addition to these common analysis mistakes, there are a couple of rules that I use when analysing data:

  • In a typical research analysis report (non-academic), after every sentence or paragraph ask “so what?”. That is, the analysis should provide meaningful information, not just put forth numbers (unless it is a pure technical report of course).
  • When reviewing a research report, do a search for the word ‘interesting’. If you have sentences such as “it is interesting to note…” or “Interestingly…” then delete them. Using ‘interesting’ to try and find meaning in data is a sure sign that it has no meaning.