Recall the definition of causality. To establish that one thing causes another, you need to demonstrate that those two things have a relationship with each other, that your proposed cause comes before the effect, and that you have eliminated other potential explanations for the effect. Those are the requirements, whether you are using quantitative or qualitative data in your analysis. But there are also two things that are nice to have when you are trying to demonstrate causality – an experimental design, which we’ll cover in a few weeks, or the causal mechanism. In this video, I discuss causal mechanisms in more detail and how we use them to demonstrate cause and effect through qualitative research.
There are many ways to establish causality through qualitative analysis. This video discusses mechanisms and process tracing, and the other two short videos assigned for today cover counterfactual analysis and necessary conditions. Of the three, process tracing – or a detailed description of how a cause leads to an effect – is the most common.
Process tracing sounds simple – just describe how things happened – but is difficult to do well. The key to effective descriptions of causal mechanisms is a clear structure. You need to cover each step of the process in a similar way, with a similar level of detail. A causal diagram can help you do this. A causal diagram is just a picture of how your variables relate to each other – drawing one forces you to think through exactly what steps you need to describe on the path from cause to effect. For example, John Owen argues that there are two pathways through which liberal democracy generates the democratic peace. Liberalism incorporates both certain ideas and certain institutions, both of which generate constraints on a government’s ability to declare war on another democracy. So in his four historical examples – cases where democracies might have gone to war with each other – he needs to describe each of these steps in the process – what the debate around war between democracies consisted of, and what institutions prevented leaders interested in declaring war from doing so.
Causal diagrams can also help you identify potential problems in an argument. For example, this is my causal diagram for my dissertation research. The link between the strength of a state’s institutions and its party system is described clearly, but it is complicated. There are a lot of steps and potential outcomes. Stronger research would focus in more narrowly on one outcome or a cause closer to the effect I’m examining (divisions in a party system) rather than an underlying cause like state capacity.
What is essential, however, is that no matter what your diagram, you systematically describe the relationship between each set of variables in your narrative about the case you are studying.
The last two points I want to make on qualitative causal inference are about one type of causal argument. Often times in historical analysis, we make what is called a path dependent argument.
Path dependency relies on the idea of critical junctures. A critical juncture is a key point in history when change is possible. Events such as the fall of communism or 9/11 – maybe Covid – open up opportunities to change institutions or policies that are otherwise stable. The institutions and practices that you set up during that time period, then, become self-reinforcing – they will be hard to change again.
For example, the book Why Nations Fail argues that a country’s initial political institutions have a big impact on its later economic development. Institutions that establish rule of law allow for effective regulation of markets and protection of property rights, promoting innovation and growth. Institutions that allow corrupt practices facilitate the development of monopolies and discourage investment, leading to economic stagnation.
The essence of path dependent arguments is that it is hard to switch from one path to another – fair legal institutions allow for self-correction, while corrupt institutions encourage further corruption. It takes a dramatic upheaval – a critical juncture – to create an opportunity for change.
So if you want to make an argument based on path dependency, you need to describe the initial critical juncture, then what happened afterwards. You will typically describe how the institutions and practices established in the critical juncture became self-reinforcing – how they keep themselves in place. And you will definitely need to describe these processes similarly – using the same variables -- in every case you study.
To summarize, if you want to use historical analysis to describe a causal mechanism, you need to be organized about it. You can use a causal diagram to describe the variables you think are relevant to your argument, then use process tracing to describe how each of those variables played out in each of your cases. When appropriate, you can also describe a critical juncture when institutions changed, then how that established path-dependent development moving forward.