Skip to main content
. 2012 Jan;40(1):49–59. doi: 10.1016/j.intell.2011.10.007

Fig. 1.

Fig. 1

Latent growth curve model of the level and change in general cognitive ability over 3 waves of the Lothian Birth Cohort 1921. The model shows a latent general cognitive ability factor (G) at each age comprising VF = Verbal Fluency, LM = Logical Memory, and Rav = Raven's Progressive Matrices. Items in rectangles are measured variables, those in ellipses are latent traits. The numbers adjacent to the arrows leading from intercept and slope to the latent cognitive factors are fixed by the investigator (the 4 refers to the 4-year period between age 79 and 83, similarly for the 8). The other numbers—beside those arrows going from measured variables to latent traits, and beside arrows between latent traits—are parameters estimated by the program. These can be treated like standardized partial beta weights, and when squared give the proportion of variance shared by adjacent variables. For the covariates, paths and parameter estimates are only given for those paths that were significant p < .05, except the path between intercept and slope which is included for reference. All parameter estimates are standardized and given to two decimal places, except the path between intercept and slope which is the unstandardized value. For sex, the reference category was male. Education is the number of years in full-time formal education/qualifications attained; smoking status is defined as never, ex or current.