Self-regulated learning
Freiburg, May 25, 2021
Many people are familiar with thinking they have understood a text, but when they go to explain it to someone else, they suddenly don’t know what to say. That is the focus of Dr. Anja Prinz’s work. Self-assessment of reading comprehension is one area of research for the research associate in the Department of Empirical Teaching and Learning Research. But she also knows one thing for certain: There are simple means to control and improve one’s own text comprehension. Annette Hoffmann spoke with her about the meta-analyses she has recently published together with Dr. Stefanie Golke and Prof. Dr. Jörg Wittwer from the Department of Educational Science.
Self-assessment in reading texts is important when learning is less instructor-driven. New meta-analyses show that people who had read texts on a computer were better at assessing their understanding of the content. Photo: Sandra Meyndt
What do we actually accomplish when we read a text?
Anja Prinz: A lot. One of the basic requirements is that we understand individual words and sentences. However, it is particularly challenging to understand the deeper meaning of texts, for example of described processes and contexts. Often, we have to draw our own conclusions from the information. In our meta-analyses, however, the primary focus was on self-assessment of one’s own comprehension. This assessment is important so that one can effectively regulate one’s own learning in the next step. For example, by selecting appropriate learning strategies to improve comprehension. During the pandemic, self-assessment is particularly important because learning at home can be less externally controlled, for example by teachers. So here it is particularly important that students themselves can monitor and accurately assess their understanding.
What kinds of texts do the study subjects receive?
For the meta-analysis, we pooled over 90 studies that used a variety of texts. Most have used expository texts, i.e., explanatory texts. However, some studies also used narrative texts. We combined these studies statistically and thus had a large database. This allowed us to obtain more reliable results than by using individual studies. It turned out that for both types of texts, self-assessment is generally very inaccurate.
What groups are your results based on?
The studies I summarized in the meta-analysis had no age limit. They included children from elementary school age all the way up to 70-year-olds.
We gain reading experience as we age and most likely, our own assessment regarding text comprehension increases, right?
Exactly, that’s what the meta-analyses have shown us. In general, learners are not good at realistically assessing their own understanding. But university students are better at it than school-age children. However, researchers believe that students can already monitor their own understanding by the end of elementary school. However, they need more support in doing so.
"Learners can realize in advance that they don't just want to remember facts and details, but to understand the deeper messages of a text," says Anja Prinz, explaining one of the methods for controlling and improving text comprehension. Photo: Patrick Seeger
Did the study also provide inferences about differences in female and male readers?
Unfortunately, this could not be investigated using the data in the meta-analyses. However, individual studies show that it depends somewhat on the domain. Boys tend to overestimate themselves more in science, whereas girls underestimate themselves more often.
Does it make a difference whether the texts were printed out in front of the study participants or whether they read them on the screen?
Yes, it made a difference. Interestingly, the assessment was more accurate when the texts were read on the computer. We explained this to ourselves in such a way that the learners are more and more used to reading texts on the screen. And that they can therefore also monitor themselves better, or more routinely, when they learn through this medium.
How can people reliably increase their self-assessment?
We looked at funding measures in a second step. What stood out was the test expectation. If learners are informed in advance that they are expected to have a deeper understanding of the text and that it is not a matter of remembering individual facts, they can also better assess their deeper understanding. Typically, learners assume that they are expected to remember facts such as years and are not focused on a deeper understanding of the text. However, it is also very beneficial to create or complete diagrams related to the texts you read. This helps to check if you can visually represent the connections or processes in the text and thus see if you have understood a text well. Other effective measures include creating summaries, writing out key words, and generating so-called concept maps. Explaining content to yourself can also be effective. However, it is important here to generate high-quality self-explanations, i.e., to explain the underlying processes and relationships to oneself in relative detail. Rereading texts can also have an added value, but this is rather low compared to the other methods.
Which method is most effective?
That is hard to say. However, we found that the test expectation was the most effective. This pleased me because learners as well as teachers can easily implement this method. Learners can realize in advance that they don't just want to remember facts and details, but to understand the deeper messages of a text. Then they usually monitor themselves better to see if they are successful in doing so. Teachers can also inform students of this learning goal and, if appropriate, present sample questions aimed at capturing deeper understanding.
How should students go about it?
Students can also set their learning goal in advance, consider what they want to learn, and monitor and assess themselves accordingly. They should also guard against false estimation, such as that shorter texts are easier to understand. Also, people in positive moods tend to overestimate more. So in the sense, I'm in a good mood, so I also understood the text well. Instead, however, you should rather check how well you can actually explain the content.