5.6: Getting Rid of Developmental Differences and Changes

At the same time, in developmental designs, there are two interesting kinds of study designs that may seem counter-intuitive, but can be useful in compiling a causal case: one in which the researcher tries to get rid of age effects and one in which the researcher tries to separate age and the potential causal variable.

Why would a researcher want of get rid of age effects?

We know it sounds paradoxical, but getting rid of age differences or changes in your data may be an effective strategy for identifying a sufficient causal bundle. Let’s think of an example with which we are already familiar—we see age differences in our cross-sectional study of cognitive performance and wonder whether these could be cohort effects. We know that there are cohort differences in education and health care, and so we can do use two strategies to see whether these cohort effects may underlie our age differences: (1) control for education and health status and see if our age differences disappear; or more interesting for relational developmentalists, (2) look for age differences within groups of our sample stratified by these conditions. We used the same strategy in our longitudinal study, remember? We looked within the conditions we thought might be contributing to our different pathways and found (to a contextualist, not surprisingly) that these groups showed different pathways.

This strategy can be used in any kind of study that includes a pattern of age differences or age changes, to look more closely at the factors that researchers think may contribute to these patterns. A simple example can be found in a study of cross-year changes in children’s use of help-seeking and concealment as ways of coping in the academic domain (Marchand & Skinner, 2007). Across the transition to middle school, students in sixth grade showed higher use of concealment and lower use of help-seeking compared to fifth graders. One possibility was that losses in motivational resources across this transition could be contributing to these age differences in coping, so two analyses were run: (1) one in which the same grade differences were examined, but with level of motivational resources held constant; and (2) groups who were high and low in motivational resources were created and grade differences within each group were examined. For both analyses, grade differences disappeared. The separate analysis of each of the motivational resources individually revealed that only a subset of them were effective in getting rid of the grade differences, suggesting that they might be more central in explaining the losses in coping capacity over the transition.

How else can a researcher get rid of age differences or changes?

We saw lifespan researchers trying to get rid of age differences that may have been based on performance factors that disadvantaged older individuals, through massive practice, untimed tests, and so on. Unfortunately, there is no list of strategies for trying to get rid of age differences or changes. Researchers must use their knowledge of the phenomenon and alternative possible explanations to systematically test which ones might be effective in wiping out age differences or changes. And, of course, only a subset of factors can be remediated in the present moment.

What are the strategies for separating age from causal processes?

Despite the fact that developmentalists are always looking for age-graded changes, none of us actually think that age per se causes anything. Age is an index for time at a particular biopsychosocicultural window, and so we think of it as a marker for some causal process that is taking place during that window. One of the most interesting causal designs is an attempt to separate age and the potential causal variable. A great example if provided by Baltes et al. (1977) in their introduction to lifespan research methods: As people age, their hearing normatively declines, but how much of this decline is inevitable and how much due to a history of exposure to noise? In typical samples, as participants get older, they have more of both: more age and more noise exposure. However, researchers can search for and select samples for whom those two factors are not confounded: young people who have very noisy jobs or who spend their time in noisy rock concerts, and older people who have very quiet jobs and live in very quiet areas. This allows researchers to “simulate” those age-graded factors that they think may be involved and see if they end up with no age differences-- or (more likely) different patterns of hearing loss for participants with different histories of noise exposure at different ages.

A second way of separating age and developmental factors is to zoom in on a particular narrow age range and to use a careful assessment of a specific developmental accomplishment that is scheduled for that age range. This assessment can then be used to distinguish people of the same age who have “crossed over” the developmental accomplishment from those who have not. Researchers can then see whether age or the developmental change is more important to the potential outcome of interest. One accomplishment that has been used this way is the differentiation between effort and ability, which typically takes place between 10 and 12 years of age (see box).

Separating Age and the Effort-ability differentiation. Children’s views of their ability make a big difference to their participation and success in school. An important developmental change underlying these perceptions is the cognitive shift in inferential reasoning that allows children to differentiate the cause of effort from that of ability. This development accompanies the shift to formal operational reasoning, which typically takes place between 10 and 12 years of age (Nicholls, 1984). Before this point, children see effort and ability as mutually diagnostic, in that greater effort implies more ability and vice versa; after they are differentiated, children come to believe that they have an inverse compensatory relationship in that more effort implies less ability, and greater ability entails less effort. So researchers could look at age-graded changes from ages 10 to 12, such as changes in ways of coping or the correlates of perceived control, and see whether these changes follow age or the developmental shift (Band & Weisz, 1990). Two such studies showed that age differences between these age groups disappeared after controlling for performance on tasks of effort-ability differentiation. More interesting, direct examination of differences according to levels of effort-ability differentiation revealed that differences in functioning walked right along those developmental steps. For example, before children differentiated effort from ability, their cognitive performances were correlated with their beliefs about their capacity to exert effort, but as effort was successively differentiated from ability, the primary correlate of performance shifted to beliefs about ability (Chapman & Skinner, 1989).