A procedure for evaluating intervention effects on ability development is discussed. The method does not postulate the traditional and routinely utilized version of full measurement invariance over time for all measures, while accounting for possible latent structure changes. The approach permits examining growth or decline free from practice effects and natural development in previously and newly emerging subabilities pertaining to a studied general ability construct. The procedure is based on an application of latent variable modeling and can be straightforwardly employed with popular software. The outlined method is illustrated using numerical data in a two-group intervention setting.