Three Essays on the Measurement of Political Ideology

In this thesis, I present three essays on the problem of measuring political ideology. Having initially set out to understand how globalisation, demographic changes, and new ideological divides contributed to the ongoing political era we find ourselves in, I quickly encountered measurement problems which could not be avoided. I therefore address various measurement issues, with a particular emphasis on survey measurement, while attempting to address this wider backdrop. I begin first by broadly defining ideology, discussing measurement theory, then discussing how this theory applies to ideology. I outline past approaches, then summarise the three essays that make up this thesis. In the first essay, I directly address the problem of measuring the ideology of voters through survey research. This task can be difficult, and researchers often rely on ‘off the shelf’ datasets. Many of these datasets contain unbalanced Likert scales, which risk acquiescence bias. This paper proposes a strategy for dealing with this issue. I first demonstrate using two comparable datasets from the UK how unbalanced scales produce distorted distributions and can affect regression results. Then, building on past research that utilises factor analysis to eliminate the influence of acquiescence bias, I demonstrate how researchers can utilise a person intercept confirmatory factor analysis model to obtain factor scores corrected for acquiescence in the case of fully unbalanced scales. I conclude with practical recommendations for researchers and survey designers moving forward. In the second essay, I address age, period, and cohort effects as possible long-term drivers of both change and stability in political ideology in the electorate. However, the question as to the extent that the explanations these effects offer is consistent across countries has not been addressed. I therefore perform a comparative APC analysis of left-right political ideology. I run two side-by-side APC analyses of left-right positions, constraining effects to be common across countries in one and nesting them within country in the other. I pay special attention to the issue of how measures of ideology can be meaningfully compared, and develop a measure of relative ideology. I find evidence for ageing effects and life-cycle effects. Moreover, I find that while the constraint of common cohort effects is not a strong one, the constrain of common period effects is over-strong. Future research should focus first on better understanding this contrasting result, and second on developing absolute measures to better understand patterns of change and continuity in left-right ideological positions in the public. In the final essay, I address the problem that social democratic parties have been confronted with vis a vis the rise of second dimension issues. These issues often see social democratic parties facing a choice between competing portions of their own electorate. A particularly prominent second dimension issue is that of the EU: should social democratic parties take pro or anti-EU positions? I look at the case of the UK as an instructive example of this debate. In the build-up and aftermath of the 2019 UK General Election, a debate emerged regarding the optimal Brexit position for the Labour Party. This debate was ultimately without satisfactory conclusion as we do not observe counterfactual versions of reality - we witness only one version of events. I therefore estimate a narrow counterfactual, simulating how the Labour Party's vote share and seat count would have changed as its position on Brexit changes. I call this counterfactual narrow because I only consider the effect of these position changes on vote choice and turnout; and not any broader consequences. I run two simulations to compare the implications of pure proximity and proximity-categorisation models of vote choice. I generate seat predictions from the simulation results by using Uniform National Swing and Uniform Regional Swing. This allows me to assess the specific distributional claims made by those advocating for a more pro-Leave position for the Labour Party. I conclude this thesis by highlighting the contributions of these three essays not only as standalone papers, but as a cohesive whole. I take a recent example of research in the Financial Times (FT) that attempts the kind of measurement I perform across the three papers, but which falls short in terms of measurement inference. I show how each paper speaks to a different aspect of the FT article, and could have influenced it in a better direction. I conclude by arguing that measurement inference is ultimately a good thing for political science, as it will lead to more secure results and richer substantive interpretation.