Note: This post is part of the “classic” RIP series, which is described in this initial post for context. (Originally posted on May 28, 2024)
One concept from the broader universe of psychological (and psychometric) theory that I’ve wanted to touch upon for some time now is what is known as “the correspondence principle.” This is a concept that I’ve brought up more than a few times in chatting with various members of the lab, and I figure that it is important enough to warrant its own RIP.
The correspondence principle is a foundational concept in psychology that has its origins in attitudes research and cognitive consistency theories. Put simply, this is the idea that I often describe as “when you ask people what they think or feel, you need to put it to them at the same level of specificity as the behaviors/outcomes you actually care about.”
A bit of history/backstory.
In the late 1960s, Allan Wicker published an extensive review highlighting the inconsistent and weak links between attitudes and behaviors. This created quite a stir among researchers, as attitudes have long been considered a central concept in psychology. It was widely believed that a person’s thoughts and feelings should strongly predict their actions. For example, if someone values health and fitness, you’d expect them to avoid fast food and smoking. However, research repeatedly found weak or no correspondence between what people said they felt and what they actually did. This led some prominent scholars to question the utility of “attitudes” as a construct and the value of measuring thoughts, feelings, values, beliefs, etc.
However, this skepticism spurred a wave of research aimed at uncovering the reasons behind these inconsistencies. Over the following decades, scholars identified several key methodological and conceptual factors influencing the attitude-behavior relationship. One of the most crucial theoretical developments from this period was the formalization of “the correspondence principle” by Ajzen and Fishbein, published a little less than a decade after Wicker’s review.
Ajzen and Fishbein observed that many studies reporting weak correlations between attitudes and behaviors often mismatched the specificity of the attitudes and behaviors being measured. For instance, predicting support for a plastic bottle tax based on a general pro-environmentalist attitude was flawed. Broad attitudes are not well-suited to predict narrowly defined behaviors. Instead, for accurate predictions, the specificity of the attitude measure must match the specificity of the behavior. If you want to predict support for a plastic bottle tax, you need to ask people how they feel about a plastic bottle tax, not just their general attitude toward environmentalism. Consequently, while general attitudes correlate poorly with specific behaviors, specific attitudes show much stronger correlations with specific behaviors.
Why is this important?
The correspondence principle is crucial in psychology for several practical reasons, particularly in the context of measurement and assessment across various domains. Its importance is most straightforward in the use of self-report measures, but it extends to other forms of psychological assessment as well. By emphasizing the alignment between what people think and how they act, the correspondence principle helps us understand, measure, predict, and influence human behavior across various contexts.
One of the primary reasons the correspondence principle is important is its role in enhancing the accuracy of psychological measurements. If you want to find correspondence between two measures, you must ensure that the things they measure actually correspond to one another. When developing and using self-report measures, it is essential to ensure that the questions accurately capture the attitudes and beliefs they are intended to measure and that these align conceptually with the outcomes we want them to predict. This basic idea helps us choose the right “level” of things to measure when studying beliefs, attitudes, etc., leading to more reliable and valid data for predictive tasks.
Critically, while the correspondence principle is vital for self-report measures, the general idea extends to other types of assessment. For example, observational methods and physiological measures can benefit from understanding the alignment between reported attitudes and observed behaviors. For instance, while someone might report a positive attitude towards healthy eating on surveys, observational studies might find that they frequently eat fast food, and their physiological measures, such as high cholesterol levels, reflect poor dietary habits. It’s a simple example, for sure, but thinking through these links can help inform study design where you cannot get observational data of behavior — if you want to do research on the links between attitudes and physical health, simply asking someone about their beliefs on “healthy eating” might not be sufficient. This comprehensive perspective helps us design assessment tools that provide a coherent picture of psychological phenomena.
You might have to trust me on this one, but believe me when I say that the correspondence principle—and the broader idea of measuring two things at an appropriate level of abstraction/specificity if you think they should correspond to each other—is truly a foundational idea in psychological research. Yet, it’s woefully underappreciated and consistently overlooked, even by experienced researchers. The principle is “rediscovered” regularly, and I often find that a subset of researchers is continually reminding everyone that this principle matters and that many ongoing debates in the field can be (at least partially) resolved by returning to this idea. So, let’s all discover this idea together—it’s something that will hopefully stick with you all for the rest of your research careers and beyond. And who knows, maybe you’ll be able to pass it on to others so that, eventually, everyone appreciates the full gravity of the correspondence principle and designs far fewer studies that end up being a complete “swing and miss” simply because they unknowingly measured things at the wrong, non-corresponding levels.
With that, then, please read and enjoy:
Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84(5), 888–918. https://doi.org/10.1037/0033-2909.84.5.888
P.S. — Note that there is now a rich literature that leans heavily on this principle, and I can’t possibly give it a fair treatment in a single Slack post. But, if you want to pursue it further, starting with Ajzen & Fishbein is definitely the way to go.