A comparison of two self-instructional methods for improving spelling in high school and college : a twenty-six classroom experiment

Graduation date: 1965This study compared two self-instructional methods for\ud improving spelling in high school and college: a traditional-deductive presentation of spelling rules and lists of words to exemplify\ud each rule versus a programed-inductive presentation leading\ud the learner to observe the spelling behavior of words and to make\ud generalizations.\ud Both the traditional-deductive and the programed-inductive\ud materials had been initially prepared for use in non-credit, self-help, corrective spelling classes at Oregon State University. Both\ud emphasized the predictable behavior of common word roots adding\ud suffixes to form common derivatives. Each presented the regularities\ud of language. Neither attempted instruction in the irregularities\ud called "spelling demons."\ud In booklet form, the programed materials comprised eight\ud sequential units in vertical, linear format whose 130 frames required\ud written responses and provided immediate reinforcement and continuous\ud knowledge of results because each frame carried its own answer.\ud They made no provision for error because steps were small and the\ud purpose of instruction was to make spelling so clear that success\ud was assured.\ud The hypothesis of the study was that if during experimentation\ud and comparison a difference in performance in spelling appeared\ud between the two modes of self instruction, the difference would favor\ud the programed-inductive method.\ud The hypothesis was tested by 606 high school and college students\ud in 26 paired experimental classrooms, one class of each pair\ud using the traditional-deductive mode of instruction, the other the\ud programed-inductive. An additional nine classrooms without match-mates raised the aggregate number to 842 students in 35 classrooms\ud and provided additional, useful data. Seven western Oregon high\ud schools and Oregon State University participated in the study.\ud Findings and Conclusions:\ud 1. An analysis of covariance compared the student's raw score\ud on the Traxler High School Spelling Test, Form 1, before instruction\ud with his mean score made on 13 tests during instruction but found no significant difference attributable to method alone. Apparently,\ud method as exemplified in this study was not a decisive variable.\ud 2. An analysis of variance showed that on the terminal Traxler,\ud Form 2, after instruction the boys using the programed-inductive\ud method made higher scores than did the boys using the tradition-deductive, a difference significant at the .05 level. Method appeared\ud to make a significant difference among boys. No such difference\ud appeared among girls.\ud 3. An analysis of variance showed girls to be better spellers\ud than boys, not only before instruction but throughout and after instruction.\ud In the light of similar findings appearing repeatedly in\ud other studies, a difference favoring girls in spelling would seem to\ud be a characteristic difference between the sexes.\ud 4. A comparison of group means on the percentile norms of\ud the Traxler Tests showed 34 of the participating classes falling\ud short of average performance regardless of scholastic ability. The\ud only group to exceed the norms was a class of university seniors\ud ready to teach English. Students in this study did not spell as well\ud as did their predecessors, the standardization group ten years ago.\ud 5. General improvement resulted from students' self-instructional efforts irrespective of method used. In this study,\ud effort appeared a more decisive factor than method