PROJECT HEAD START WAS CONCEIVED IN PART, TO INCREASE THE ORDINARILY REDUCED NUMBER OF CONNECTIONS BETWEEN THE CHILD AND OTHER PEOPLE. FOUR QUESTIONS WERE ASKED. (1) WOULD THE DEVELOPMENT OF SELF-SOCIAL CONSTRUCTS OF HEAD START CHILDREN DIFFER FROM THE DEVELOPMENT OF CHILDREN IN A CONTROL GROUP. (2) WOULD HEAD START CHILDREN DEVELOP APPROPRIATE SOCIAL TRUST. (3) WOULD THE TEACHERS' COGNITIVE STYLES AFFECT THE DEVELOPMENT OF SELF-SOCIAL CONSTRUCTS. (4) WOULD THE TEACHERS' PERCEPTIONS OF HEAD START CHILDREN AFFECT THE DEVELOPMENT OF SELF-SOCIAL CONSTRUCTS. TWO-THIRDS OF THE STUDENTS, OR 978 CHILDREN IN THE DELAWARE SUMMER 1965 HEAD START PROGRAMS WERE GIVEN A PRE-TEST ON ELEVEN TASKS FROM THE SELF-SOCIAL SYMBOLS TASKS, AND 945 OF THE CHILDREN WERE POST-TESTED, 100 NON-HEAD START CHILDREN OF A COMPARABLE BACKGROUND WERE USED AS A CONTROL GROUP AND TESTED IN THEIR HOMES AT THE SAME TIME AS THE HEAD START STUDENTS. EIGHTY PAIRS OF CHILDREN WERE PRE-TESTED ON A SHARING TASK MEASURING SOCIAL TRUST, AND 20 PAIRS WERE POST-TESTED. RESULTS OF THE TESTS SHOWED THAT THE HEAD START EXPERIENCE PRODUCED POSITIVE CHANGES IN SELF AND SELF-OTHER RELATIONSHIPS. HEAD START CHILDREN GAINED A PERCEPTION OF SELF AS BEING SIMILAR TO OTHERS AND TENDED TO MAINTAIN SELF AS CENTRAL. CONTROLS SHOWED A SHIFT TOWARD LOWER SELF-ESTEEM AND A LACK OF CHANGE FROM A SELF-DIFFERENT TO A SELF-SAME RESPONSE. (CO'D)