The eagle eyed among you may have noticed a new feature appear recently on WhatScotlandThinks.org. If you go to any page with a graph or table and scroll down a little, there is now a button called ‘Edit Responses’. Click on this and you will see that it allows you to combine the response categories for any question.
Let us look at an example of how someone might want to use this feature. If you were interested in how likely people are to vote in the referendum, you might come across this graph showing the results of a question that Ipsos Mori have repeatedly asked on the topic:
We can see quite readily that the majority of people say they are ‘certain to vote’, with typically around three-quarters choosing the answer 10 on this 10 point scale. That means, however, that the proportions choosing any of the other nine possible responses are all small, and it is difficult to see what is going on in the rest of the graph.
To resolve this, you can click on ‘Edit Responses’ then tick, say, the boxes next to 1, 2, 3, 4 and 5 so that they comprise your first ‘group’. If you then click on ‘add another custom group’ you can tick 6, 7, 8 and 9 so that the form a second group. Then click on ‘view combined responses and you will see this graph:
It is now easy to see that the proportion of those answering between 6 to 9 on this scale is consistently higher than the proportion answering 1 (absolutely certain would not vote) to 5. If you change to table view, you can see that on most occasions when this question has been asked the number of individuals with a score of between 6 and 9 has been around double those with a score of 5 or less. This would seem to underline the message that if people do what they say they will do, we should expect a very high turnout in September.
One obvious use for this new facility is to summarise the results for questions that give respondents five possible answers ranging from ‘strongly agree’ to ‘strongly disagree’. You might simply want to be able to see how many agree, whether strongly or not, and how many disagree. One such example is the following question in which respondents in England were asked whether they thought England should become an independent country.
To see how many people agree and how many disagree with this proposition, we can use the new tool to merge those who ‘strongly agree’ with those who ‘tend to agree’ into one group, and those who ‘strongly disagree’ or ‘tend to disagree’ into another.
The graph now brings out how evenly divided the responses proved to be in response to this question. At 38%, those who disagree are almost equally matched by the 34% who agree.
One other point to note. If you scroll down the page of this second example, you can see that the middle response category ‘Neither agree nor disagree’ has been selected to form ‘Group 2’ even though it is not being combined with any other response option. Doing so ensures that that response is placed in the middle of the graph; otherwise anything that is not selected as a ‘group’ is placed at the right hand side. The graph would, of course, still make sense with the middle category placed on the far right, but it is easier to interpret if the response categories are placed in this more logical order.
Many of questions in the Scottish Social Attitudes survey consist of similar five point agree/disagree scales. Now you can combine the ‘strongly agrees’ and the ‘agrees’ (and the ‘strongly disagrees’ and ‘disagrees’) to see more clearly what Scotland thinks about a wide range of issues.