In recent years the use of traditional statistical methods in educational research has increasingly come under attack. In this article, Ronald P. Carver exposes the fantasies often entertained by researchers about the meaning of statistical significance. The author recommends abandoning all statistical significance testing and suggests other ways of evaluating research results. Carver concludes that we should return to the scientific method of examining data and replicating results rather than relying on statistical significance testing to provide equivalent information.

No abstract

The relationship between the relative difficulty of passages and the number of unknown words in passages was investigated. In Study I, 219 students in Grades 3, 4, 5, and 6 were given 100-word passages sampled from textbooks and library books (trade books) and then asked to underline each word that they did not know. In Study II, 60 graduate students were given a similar task involving 120 passages at the following levels of difficulty: junior high, senior high, college, and graduate school. Relative difficulty (or relative easiness) was determined from the difference between a measure of the reading ability of the student in grade equivalent (GE) units and a measure of the difficulty level of the material in GE units. The results from both Study I and Study II can be summarized as follows: (a) when the material is relatively easy then close to 0% will be unknown basic words, (b) when the material is relatively hard, then 2% or more will be unknown basic words, and (c) when the difficulty of the material is matched closely to the ability of the reader, then around 1% will be unknown basic words. These findings do not support the theory that free reading results in large vocabulary growth because free reading is likely to involve relatively easy material that contains few, if any, unknown words. Indeed, these findings call into question the practice of devoting large amounts of classroom time to free reading if the purpose is to increase vocabulary or reading level.The purpose of this research was to determine the relationship between the number of unknown words in a passage and the relative difficulty of the passage. The functional relationship between the relative difficulty of a passage and the degree to which a passage can be comprehended was studied by Carver (1990a). He found a linear relationship that was readily summarized by an equation for a straight line; the accuracy of comprehension while engaged in normal reading, called rauding, was predictable from the relative difficulty of the material.It seemed important to determine the relationship between the number of unknown words in a passage and the relative difficulty of the passage because it has been theorized that any kind of reading (relatively easy, matched, or hard) improves vocabulary. For example, Nagy and Anderson (1984) have contended that "beginning in about third grade, the major determinant of vocabulary growth is the amount of free reading" (p. 327), and Krashen (1989) concluded that "free reading is the best way to build vocabulary" (p. 455). Yet, if individuals read relatively easy material and this material contains few if any unknown words, then it is difficult to understand how reading a large amount of this material could increase the breath of one's vocabulary.None of the advocates of free reading to achieve vocabulary growth have qualified their generalizations by placing restrictions on the relative difficulty of the material. For example, Nagy, Herman, and Anderson (1985) concluded from their research that "our results st...

Notes that because of design and development techniques, most tests focus upon either between-individual differences or within-individual gain. These 2 primary dimensions of tests have been called psychometric and edumetric. The psychometric dimension has been focused upon traditionally, so that most of the standardized tests being used to measure gain or growth have not been developed or evaluated from an edumetric standpoint. The danger inherent in this practice is that a treatment effect may appear to be small or insignificant simply because the test was designed to be sensitive to individual differences rather than gain. The differences between psychometric and edumetric dimensions on the Sequential Tests of Educational Progress (Reading) are compared with respect to the purpose of the test, item selection, reliability, validity, and interpretation of scores. (15 ref)

No abstract