Description
Compares groups by fitting performance as a function of age, experience or ability in each group, instead of matching participants to controls at one point.
The traditional comparison of an atypical group matches it to controls of the same chronological age and, separately, of the same mental age: a deficit against both suggests deviance, a deficit only against age-matched controls suggests delay. Matching removes the developmental variable by design and so hides how performance changes with it. Trajectory analysis instead recruits each group across a wide range of the developmental variable and fits performance as a function of it. Groups are then compared on intercept (onset), slope (rate) and shape of their functions, and on how performance on one task predicts another within each group. The approach turns a single-point comparison into a comparison of functions and exposes groups that reach similar levels by different routes. Cross-sectional trajectories describe group-level change, not individual change, so longitudinal follow-up is needed for claims about individuals.
Attributes
- Discipline
- Statistics & Modelling
- Loss type
- Unmigrated
- Stage
- Discovery, Evaluation
- Data
- Quantitative
- Effort
- Weeks
Steps
- Choose the developmental or experience variable (age, months of product use, skill score) that the comparison should be expressed over.
- Recruit each group across a wide, evenly spread range of that variable rather than at a single matched level.
- Fit performance against the variable separately for each group, starting with a linear function and testing curvature where theory suggests it.
- Compare groups on intercepts, slopes and shapes using interaction terms in a single regression or mixed model.
- Relate performance on one task to another within each group to see whether skills are linked in the same way.
- Report where trajectories diverge and, if claims about individuals are needed, plan longitudinal follow-up.
Sources
- Wiley-Blackwell Handbook of Childhood Cognitive Development (Goswami, 2011)
Also known as
- Trajectory approach
- Cross-sectional developmental trajectories
- Developmental trajectory analysis versus chronological-age and mental-age matching
Categories
- Analytics & Behaviour
- Statistics & Modelling