The self-reference effect

In lab experiments, adults and teenagers are more likely to remember trait words (like ‘beautiful’) if they are related to the self (option 1), than if they are related to someone else (option 2) or processed non-socially (option 3). By adulthood, our self-concept (containing what we know and feel about ourselves) is highly detailed and elaborate. By associating the item with our self-concept we achieve a depth of processing that is perhaps unparalleled by any other concept we hold.  For example, we are likely to be able to draw on episodic memories of past events where our beauty was judged, and feel the emotional consequences of this judgment. We are not so informed or invested about the ‘history’ of George Clooney’s beauty. Nor are we likely to pay as close attention to a task that involves categorising someone else, or judging words by their physical properties, as we would if the task was made ‘self-referent’.

In ‘real’ life, the self-reference effect is expressed by a tendency to pay attention to self-referent information (think of your reaction when you overhear your own name mentioned in conversation) and, when self-focused, to evaluate ourselves relative to salient standards (are we proud about what we are overhearing?). This cognitive activity is likely to create a strong memory trace, allowing us to store useful information about ourselves and further elaborate our self-concept. From an evolutionary perspective, the ability to prioritise our attention and memory to the events that are most relevant for us is very profitable. We are primed to react to events that are likely to concern us, and store information to inform our decisions in the future.

Although quite a lot is known about mature self-reference effects, relatively little is known about how these effects develop. We know that children have developed a basic self-concept by about 2 years of age. To the extent that the self-reference effect is driven by attention, it should be possible to witness advantage for self-referent stimuli as soon as the self-concept comes ‘online’ in toddlerhood. However, the strength of self-reference effects may increase with age, since growth in the self-concept may be required for in depth processing of self-referent material. Our project is designed to test these ideas about the development of the self-reference effect. We aim to find out if simple ‘attention grabbing’ self-reference effects develop before more complex ‘self-evaluation’ effects, and if the strength of these effects are related to growth in other areas of the self-system (the self-concept and autobiographical memory). Very little is known about how the self-system develops, and self-reference effects may provide a valuable ‘tool’ to explore development in this area.

The traditional lab based self-reference paradigm is not appropriate for young children since they don’t have the vocabulary to process trait adjectives, so it is necessary to design new tests for use with children. A simple way to measure the development of self-reference effects in early childhood is to pair to-be-remembered material with a concrete representation of the self such as a self-photograph (see Cunningham et al, 2014; Ross et al, 2011). This supports children in seeing the to-be-remembered material as self-referent.

One can then ask the children to process the material in simple or more complex ways. For example, children can be asked whether the to-be-remembered object appears to the left or right of their self-image. Children can do this simply by attending to the stimuli. Alternatively, a child might be asked if they ‘like’ the to-be-remembered object, this requires more complex self-evaluative processing. For both tasks, we can assess the strength of the self-reference effect by comparing the children’s memory for the objects shown with their own photograph, with their memory for objects shown with a peer’s photograph. Importantly, these tasks are appropriate from early to late childhood and allowed us to directly compare the memory performance of different age groups in our study (our project tests 3- to 10-year-olds), producing the first comprehensive picture of the early development of the self-reference effect.