You’re most likely accustomed to surveys that ask you to fee your self on a scale of 1 to 10, whether or not they ask about your happiness, job satisfaction and even political leanings. Typically, these questions use a 0–10 scale as a substitute. However do these numbers actually matter, aside from their rank order? Is it attainable, as an example, that some survey respondents could be extra more likely to describe themselves with the #1 than the number 0 because 1 is a optimistic quantity and Zero will not be?

Understanding how the numeric labels on scales would possibly influence survey responses is an area of ongoing investigation for researchers. At Pew Analysis Middle, we’ve spent a while eager about this, too.

We lately ran an experiment testing whether or not the numeric labels utilized in one among our scales influenced how respondents reported their political ideology. We in contrast the outcomes from a survey query that used 0 because the minimal worth on a scale with one which used 1 as a substitute. (Different researchers have thought of how people respond to political ideology questions and the constraints of asking about political ideology extra typically.)

We performed this experiment as a part of a 2017 telephone survey about religion and pluralism in Western Europe.  To hold it out, we randomly assigned respondents in France, Germany and the UK our political ideology query with one among two 7-point scales: both 0–6 or 1–7. The complete query wording utilizing our conventional 0–6 scale was this: “Some folks discuss politics when it comes to left, heart and proper. On a left-right scale from 0 to 6, with 0 indicating excessive left and 6 indicating excessive proper, the place would you place your self?”

Our aim was to handle a number of questions:

1. Do totally different scale endpoints end in totally different frequency distributions across the ideological spectrum or in numerous charges of merchandise nonresponse?

2. Do in another way numbered scales have an effect on the broader left-right lean of the scale?

3. Do the two scales end in different demographic profiles for teams on the ideological spectrum (for instance, evaluating the demographics of the ideological heart on every scale)?

4. Is our understanding of how ideological teams really feel about major political parties and important societal issues influenced by which response scale we use to measure ideology?

Under, you’ll discover the outcomes of our experiment. For ease of dialogue, I’ve labeled the interior points of the scales regardless that the query itself solely assigned labels to the endpoints (that’s, Excessive Left and Excessive Proper). The ensuing scale (in ascending “numeric order”) is as follows: Excessive Left, Left, Middle-Left, Middle, Middle-Proper, Proper and Excessive Proper.

Distributions across the ideological spectrum

The priority that fewer people can be prepared to self-report as Excessive Left when that response is related to 0 rather than 1 is unfounded. In all three nations, respondents had been equally more likely to choose the Excessive Left reply class no matter which numeric scale was used. Equally, in every nation, item nonresponse charges (the charges at which individuals selected to not reply the ideology questions) had been constant between the two scales.

Moreover, whereas Left self-reports differed throughout the scales in all three nations, it was not within the route that will assist the speculation that folks can be unwilling to decide on the Excessive Left class when it’s assigned a price of 0. If this speculation had been true, we’d anticipate a spike of 0–6 scale respondents selecting Left (numeric worth 1) as in comparison with the 1–7 scale respondents selecting Left (numeric worth 2). In actual fact, extra respondents who had been assigned the 1–7 scale selected Left in every nation.

Throughout all six scale distributions we studied, there was a peak close to the center that tapered towards the dimensions’s ends. Within the UK, each scale variations peaked on the ideological Middle, whatever the numeric worth related to it. However in France and Germany, the modal class was related to the numeric worth 3, whatever the ideological label: The quantity Three represents Middle on the 0–6 scale, however Middle-Left on the 1–7 scale. There have been additionally vital variations between the scales on the Middle positions in France and Germany, with the 0–6 scales having extra Middle responses.

These outcomes counsel that when a scale is definitely divided in half — for instance, when the utmost worth is 6 reasonably than 7 — it’s extra doubtless for respondents to pick the midpoint. Previous research has discovered that respondents are more likely to assume that half the highest endpoint is a scale’s midpoint, so when half the highest endpoint will not be a solution possibility (e.g., 3.5 on the 1–7 scale), respondents searching for the central level on the dimensions could someday select 3 (not the dimensions midpoint) and typically select 4 (the precise scale midpoint). Those that acquired the 0–6 scale might extra simply discover the midpoint (3) by halving the highest endpoint.

Left-right lean of the scales

To find out if the in another way numbered scales affected the broader left-right lean of the scale, I condensed every seven-point scale into three common classes: Left-c, Middle and Proper-c. (The “-c” differentiates these classes from the equally named classes within the seven-point scales.) Left-c is a mixture of the Excessive Left, Left and Middle-Left classes, Proper-c is a mixture of the corresponding classes on the different finish of the scales, and Middle is just the Middle class in each scales.

In all three nations, utilizing the 1–7 scale elevated self-identification as Left-c compared to the 0–6 scale. Students have noted that the meaning of the center in political ideology scales can differ based mostly on context and respondent interpretation. Using Zero vs. 1 because the decrease endpoint might have altered the respondent notion of the query and its reply choices, resulting in totally different distributions.

Demographic differences

In all three nations, when utilizing the condensed distribution of Left-c, Middle and Proper-c, there aren’t any statistically vital variations between the 2 scales in terms of demographic structure (age, gender and schooling). For instance, inside every nation, the Left-c is equally divided between young and old, female and male, and extra and fewer educated — regardless of which scale is used.

When trying on the full 7-point ideology scales, solely 4 classes have ample pattern sizes (roughly 100 instances or extra) to make comparisons between the two scales. These are the Left, Center-Left, Center and Center-Proper categories. (Center is identical as within the condensed distribution so will not be thought of once more). Wanting on the demographic make-up of those classes, there is just one statistically vital distinction: Utilizing France’s 0–6 scale, a majority of the Left is male (56%), whereas on the 1–7 scale, a minority of the Left is male (41%).

Political party favorability and other societal points

We additionally requested respondents in every nation if that they had a really favorable, considerably favorable, considerably unfavorable or very unfavorable opinion of a number of nationwide political events. For this evaluation, I chosen two traditional parties from every nation (one celebration every from the left and the best): France’s Socialist Social gathering and the Republicans; Germany’s Social Democratic Social gathering (SPD) and Christian Democratic Union (CDU); and the UK’s Labour Social gathering and Conservative Social gathering. I in contrast “internet” favorability (combining very favorable and considerably favorable) towards the events for every ideological class between the scales.

Total, political ideology corresponds constantly to political celebration favorability, regardless of which scale is used. There are a couple of variations throughout the three nations, however nothing suggests a scientific distinction between the 2 ideology scales.

In the meantime, to get an concept of how the 2 scales work together with different matters, I in contrast the totally different measures of ideology throughout quite a lot of substantive questions. The objects chosen had been about faith (the share who say faith is essential of their life), nationalism (the share who fully/principally agree that “our individuals are not excellent, however our tradition is superior to others”), minority religions (the share who say Islam is basically incompatible with their nationwide tradition and values), immigration (the share who suppose the variety of immigrants to their nation needs to be diminished) and social points (the share who suppose abortion needs to be authorized in all or most instances). The survey didn’t ask questions on financial coverage, which might additionally relate to a person’s political ideology.

By and huge, respondents within the corresponding ideology classes from the totally different scales give the identical solutions to those questions. That is true for each the consolidated left-center-right categorization and the classes from the seven-point scales for which we’ve got ample pattern sizes to do an evaluation (roughly 100 instances or extra). Throughout all these factors of comparability, nothing suggests systematic variations between the ideology scales. And the person variations — even when pretty extensive — have a tendency to not change the overarching story of a given group’s stance on a specific subject. For instance, utilizing France’s 0–6 scale, 59% of the Left thinks abortion needs to be authorized, whereas on France’s 1–7 scale, 87% of the Left helps legalized abortion. This can be a statistically vital distinction, to make certain, however it nonetheless exhibits broad assist among the many Left.

Conclusions

If a researcher’s aim is to get a way of the relative dimension of the political left, heart or proper within a country, there might be significant variations relying on whether or not one uses the 0–6 scale or the 1–7 scale — particularly in terms of the top of the dimensions closest to 0. Further analysis is required to know which scale presents a extra legitimate measure of political ideology.

Nevertheless, an ideological group’s demographic profile, favorability towards particular political events and attitudes on different vital points don’t drastically change when totally different numeric values are utilized in a seven-point scale. The general impression of utilizing a 0–6 scale or a 1–7 scale as an impartial variable or management for social research appears minimal.

FAQ