Researcher Elizabeth Goldfarb's research on the cognitive styles of risky drinkers was the topic of a recent episode of the podcasts, Under the Cortext, produced by APS. You can listen to the 15- minute discussion between Dr. Goldfarb and Dr. Özge G. Fischer Baum on the podcast, or you can read the transcript of the discussion here. The podcast does a great job of describing risky drinking and explaining some of the downsides of risky drinking.
What is risky drinking? Dr. Goldfarb describes it as follows:
…risky drinking is basically engaging in patterns of alcohol intake that are associated with higher risk for negative outcomes, both for physical well-being and for mental health. [For example] regularly engaging in binges where you consume a large amount of alcohol in a short amount of time, or just persistently having overall high levels of intake throughout your week.
The cognitive processes that Dr. Goldfarb studies are generalization and overgeneralization. Generalization is a type of learning:
…say a dog bites you. You want to use that to inform how you interact with dogs in the future, right? You don’t want to just walk up to another dog, stick your hand in front of its nose, and I wonder what’s going to happen now. You want to be able to use your past knowledge, right? But you don’t want to take that to an extreme and start being really afraid of even seeing any dog far away from you across the street. So we’d think about the first one of just taking your knowledge and using it adaptively as generalization. But taking it too far would be a form of overgeneralization.
Goldfarb's team used a card game to study overgeneralization in both risky and lighter drinkers. They found that, compared to lighter drinkers, risky drinkers tend to overgeneralize associations with alcohol. To find out how they operationalized "overgeneralization," see if you can follow their study paradigm based on her description:
… we told people that they were going to play a card game to earn tokens that would add to a financial bonus they’d get at the end of the study. So first they learned which cards would earn them the tokens, and then they got to choose in the second phase, which cards did they want to play with? What was happening under the hood in this game was that one card was paired with tokens that were actually pictures of alcohol, and the other card was paired with tokens that were just pictures of common objects. And the reason we did that was to see, are they really focused on alcohol related reward, or is it just any kind of reward that they’re going to generalize more? And then the third card was very rarely paired with any kind of token.
[we found that]…people who engaged in riskier patterns of drinking behavior stretched these alcohol associations specifically out farther. So [later in the game] they showed this tendency [to overgeneralize–they wanted to use] cards even if they didn’t look very much like the one that was paired with alcohol, they still want to approach it, they still want to play with it.
Next, the research team dug deeper; they wanted to test if the overgeneralization effect might also occur for punishment learning:
…we did the study again, but now they were penalty tokens. They were going to take away money from our participants. And [risky drinkers] also over generalized alcohol punishment. They showed a tendency here to avoid cards that looked anything like the one that was initially paired with an alcohol punishment. [This pair of studies is] showing that there’s just this overall tendency to take what you learn about alcohol and spread it out to overgeneralize alcohol associations in a way that’s really specific to alcohol. But valence general, it doesn’t matter if it’s gain or loss.
You might be asking, why do we care how a risky drinker plays a card game? The podcast host asked that question, too, and Goldfarb replied:
…we know that memory plays a huge role in substance use. You can’t relapse in a place where you used to use if you don’t remember that you ever used there before, but we don’t know what it is that they’re remembering. So many of the projects we’re doing in the lab are trying to isolate. We have all these different memory systems, all these different types of representations, which are the ones that are really motivating this problematic behavior. And once we understand which memories they are, we can start to think about how can we target them, how can we help people form different, perhaps more protective associations
This is a challenging study to work with, but keep in mind that at the heart of every study are its variables.
a) Classify the variables in the first study (reinforcement learning).
| Variable name | How was this variable operationalized? | What are the variable's possible levels? | Is this manipulated or measured? |
b) Which kind of study is this? Correlational or experimental?
c) See if you can sketch a scatterplot, bar graph, or line graph of the results they described. Later you can compare your sketch to the results in the empirical article (Figure 2, especially panel b).
The empirical article reporting this study appears in Psychological Science; it contains images of the cards used and full descriptions of the results.