Challenges in Measuring Longitudinally
Posted on December 10, 2007
John E. Bates (bio) outlines some measurement challenges related to children's development.
Measurement of the same construct at different ages is one of our big challenges, of course. And the dilemma is always, do you change the measure to be more appropriate for a different developmental stage? And I don’t have a solid answer on this. We’ve done it both ways. We’ve found some real advantages to keeping the measure the same as much as possible across development, but there are also some ways you can adjust in your analysis for slight changes in your measurement approach. So we don’t do it casually.
It really depends on the individual situation or the particular concept we are trying to measure. The work we are just getting started now, we're studying self regulation in young children, and we want to be able to bridge the span of 30 months to 42 months, and a lot of development and self regulation happen during that phase of development. And we want to come up with tasks that are going to be valid for that whole one-year period. So we don’t want to have something that has a floor effect at the outset and a ceiling effect at the end.
We want to have something that has meaningful variation possible through out. And so that is a challenge finding all those tasks that will work in that area. So that is one particular place in which we have to think about measurement issues. Another kind of measurement issue has to do with when, in questionnaire measures, measures of adjustment or competencies in social behavior, you have really different repertoires at different ages, so stuff you are asking eight year olds about may not be very relevant to the stuff that you are asking thirteen year olds about, and vice versa.
So you may have the similar concepts, like externalizing behavioral problems, but you might want to be asking different questions at different stages.
Measurement intervals is another topic in this kind of research. You have to think what is a meaningful amount of time for a change to have occurred, and it depends on what your goal is. If you are just trying to, in a general sense, track growth, it is not critical that you see every single change in the kids, so then you pick some fairly convenient intervals that allow you to see some growth.
There's not a lot of people doing this kind of research, but if you had research where you really wanted to see the coming online of a particular skill, then you’d really have to be very thoughtful about your longitudinal intervals and you might have much shorter intervals and fewer participants, and these are the kinds of things that we haven’t fully resolved, but we are learning.
Excerpted from interview with researcher at the 2007 SRCD Biennial Meeting in Boston, MA.
Please note that the feedback is viewed only by 4researchers staff and is not intended for communication with individual contributors.
Use the form below to submit feedback about this article. If you would like a response, please be sure to include your e-mail address.