A CONCEPT ATTAINMENT TASK IN WHICH CHILDREN ARE ASKED TO IDENTIFY A CONCEPT THROUGH DETERMINING WHICH OF A SET OF OBJECTS ARE OR ARE NOT INSTANCES OF IT WAS USED TO INVESTIGATE THEIR PROBLEM-SOLVING BEHAVIOR. OF PARTICULAR INTEREST WAS THE QUESTION OF TO WHAT EXTENT THEY USED OR LEARNED THE STRATEGY OF VARYING ONE FACTOR WHILE HOLDING ALL OTHERS CONSTANT. IN THE FIRST OF THREE EXPERIMENTS A SAMPLE OF 144 FOURTH GRADERS WERE GIVEN 24 APTITUDE TESTS AND NINE CONCEPT ATTAINMENT TASKS. OF SEVEN APTITUDE FACTORS FOUND, ONLY TWO, CALLED SPATIAL ORIENTATION AND FIGURAL ADAPTIVE FLEXIBILITY, SHOWED SIGNIFICANT CORRELATIONS WITH PROBLEM-SOLVING PERFORMANCE. THE SECOND EXPERIMENT INVESTIGATED THE EFFECT OF THE PRESENCE OR ABSENCE OF ACTUAL OBJECTS ON CHILDREN'S PERFORMANCE. A SAMPLE OF 24 THIRD OR FOURTH GRADERS WERE EACH GIVEN THREE PROBLEMS TO SOLVE. OF FOUR GROUPS, ONLY THE ONE NEVER SHOWN ANY OBJECTS PERFORMED SIGNIFICANTLY MORE POORLY. THE THIRD EXPERIMENT WAS INTENDED TO COMPARE TWO METHODS OF TEACHING THE DESIRED PROBLEM-SOLVING STRATEGY, A PART-TASK METHOD WHICH ATTEMPTED TO TEACH COMPONENT SKILLS AND A WHOLE-TASK METHOD WHICH ATTEMPTED TO TEACH THE TOTAL SKILL. TWO GROUPS OF 18 FIRST GRADERS RECEIVED ONE OR THE OTHER FORM OF TRAINING, AFTER WHICH THEIR PERFORMANCE WAS COMPARED WITH A GROUP OF UNTRAINED FIRST GRADERS. ON RETENTION PROBLEMS THE PART-TASK GROUP WAS SIGNIFICANTLY BETTER THAN THE WHOLE-TASK GROUP, WHICH WAS IN TURN BETTER THAN THE UNTRAINED GROUP. (DR)