We use bar charts on this site to visually represent responses to a survey or polling question where there has only been one period of fieldwork. Each bar shows the proportion of people who gave a particular response to the question; hover over the bar to see the percentage.
We use line graphs on this site to visually represent responses to a survey or polling question that has been asked over two or more periods of fieldwork. You will see a vertical row of dots for each period of fieldwork and can hover over these to see the percentages. The coloured lines join up the same answers (e.g. the percentages of people who responded ‘yes’ to a question in each fieldwork period), highlighting any changes over time in response to the question.
Sometimes when researchers are presenting polling or survey data, they remove all responses where people did not answer using one of the set response options – they may have said ‘Don’t know’, ‘Can’t choose’ or given an alternative answer. Removing these types of answers is especially useful for looking at voting intention as it presents the figures focussing only on those people who have decided how they will vote. On this site, we normally include ‘Don’t know’ figures where available but give users the option to remove them; selecting this will result in the rest of the figures automatically updating so that the total is 100%.
Why are there different versions of the question wording for one graph?
Sometimes the questions asked in surveys and polls are very long and other times they only make sense in the context of the questionnaire (for example, a question might be a following up on a topic introduced in the previous survey question). To make questions easier to search for, we always have a ‘web-friendly’ version, which is what you see in the question listing and title. The full version of the question in a bar chart is at the bottom of the page.
In line graphs, data from different sources are brought together so there may be slight variations between them in terms of question wording. For these, click on ‘About these data‘ under the graph to see the full questions, which may all be the same or there may be slight variations.
Why would data from different polls or surveys be combined on the same line graph?
Most line graphs on this site show repeat questions from the same company – eg a regularly asked by a pollster like Ipsos MORI or a question that is part of a large-scale. However, when different companies ask the same question – or questions that are almost identical to one another – we think it is helpful to users to see these combined in the same graph. You can find out more about the sources of the data in any graph by hovering over the dots, looking at it as a table or reading the full details in ‘About these data’.
Why does the total of responses sometimes appear to add up to 101% or 99%?
This is due to the ’rounding error’. All percentages on this site are rounded to the nearest whole number. If a set of three figures that add up to 100% are all rounded up this sometimes results in rounded figures that total 101%. For example, when shown to one decimal place, the figures 50.6%, 20.7% and 28.7% add up to 100% but when rounded to the nearest whole numbers (51%, 21% and 29%) add up to 101%
Occasionally researchers may wish to combine two or more response categories in order to reduce the complexity of the data and produce a broad overview of what the results are telling us. A common example of this is where a Likert scale is used to assess the extent to which respondents agree or disagree with a given statement.In this scenario, response options include ‘Strongly agree’, ‘Agree’, ‘Neither agree nor disagree’, ‘Disagree’, and ‘Strongly disagree’. In order to examine the number of people who broadly agree or disagree with this stance, it’s useful for researchers to be able to combine the figure for those who answered ‘Strongly agree’ with the figure for those who answered ‘Agree’ (and, equally, the figure for those who answered ‘Disagree’ with the figure for those who answered ‘Strongly disagree’).