Search

A case is made for representing quantitative methods in use in the social sciences within a unified framework based on structural equation methodology (SEM). Most of the methods now in use are shown in their SEM representation. It is suggested that the visual and verbal representations of SEM are of most use, while specific estimation and hypothesis testing methods play a lesser role. The change to a common verbal and visual representation for quantitative methods allows better understanding of quantitative methods for students, easier communication among researchers, and new improvements in quantitative approaches to social research. (Contains 8 figures and 15 references.) (Author/SLD)

Students in grades 1-6 who were part of the norming sample for the Kaufman Test of Educational Achievement took both a word identification task, Reading and Decoding, and a spelling test. Each word in both tests was coded for linguistic components: number of phonemes, consonant blends, vowel digraphs, consonant digraphs, r-controlled vowels, silent markers, and regular or irregular pronunciation/spelling of the word. For each student a regression analysis was performed to predict whether the student could successfully pronounce (spell) the word using the linguistic components as predictors. Regression weights were then used in various multivariate analyses along with overall word identification and spelling performance to investigate relationships among the variables. Correlations among the two sets of variables (word identification and spelling linguistic components and achievement) indicated generally high correlations at all grades among linguistic components and achievement, and between word identification and spelling achievement. Structural equation models were developed at each grade. Specific patterns varied at each grade level that appeared to be consistent with instructional emphases at each grade level in both reading and spelling. For example, phonemic length was important at grades 1-2 but not beyond for both reading and spelling, while components such as vowel digraphs and silent markers varied with grade in their importance for reading and spelling. Findings suggest that word regularity is more important to spelling than reading beyond grade 2, as students encounter increasingly difficult spelling words. Regularity in word identification does not appear to play a role in upper grade activities. (Contains 10 references, 1 table and 8 figures of data.) (Author)

A study examined the developmental shifts in the importance of linguistic components of words to single word decoding. Subjects, 1200 children ages 6-7, 8-9, and 10-12.5 years who were the normative sample of the Kaufman Assessment Battery for Children (K-ABC), were tested using a word recognition test, the Reading Decoding Subtest of the K-ABC. Each word was decomposed into nine structural components important to linguistic information: number of phonemes, graphemes, syllables, morphemes, consonant digraphs, vowel digraphs, r-controlled vowels, consonant blends, and silent markers. Results indicated that: (1) for all three age levels, phonemes, graphemes, and syllables were so highly correlated and correlated so similarly to success in decoding that only phonemes were used; (2) for all three age levels, within subject regression coefficients of the seven variables on right-wrong score exhibited the same direction of regression but relative importance varied; and (3) phonemic complexity accounted for the highest amount of variance at each level, with other variables shifting across age in a pattern consistent with a stage development model for reading. Findings suggest questioning the concept of a true automaticity in reading. (One table of data is included; 18 references are attached.) (Author/RS)

Given the complexity of classroom reading observations, maintenance of measurement reliability is a concern to researchers. Sources of variation contributing to the unreliability of many measurements may be either (1) lasting and specific, (2) lasting and general, (3) temporary and specific, or (4) temporary and general. Lasting-specific sources of variations, such as observers' attitudes toward the coding system and differences between training conditions and actual classroom settings, can be controlled through appropriate criteria for observer selection and training procedures that are content specific to the actual observation tasks. Lasting-general sources of variation include observers' observation skills and the supervision condition, which can be controlled, in part, by researchers and observers for discussing recently completed observations. Although temporary-specific sources of variation, such as observer fatigue, attention span, and memory, are not always under experimenter control, they can be managed with some success by such techniques as regulation of observation periods. Temporary-general sources, including location, surroundings, and physical condition of the observer, are more amenable to researcher control. If an observer is ill, for example, the observation session can be cancelled. The systematic consideration of these sources of variation is critical to understanding classroom interaction and behavior. (MM)

A meta-analysis of 29 separate studies investigating pretest effects was conducted. Outcomes of the studies (achievement gains or attitude improvements) were computed as standardized differences between pretested and non-pretested groups. Eleven other variables were coded for each outcome. Initial descriptive statistics were indicative of differences between randomized and nonrandomized studies, so all further analyses were based on the 110 randomized group outcomes. For all outcomes, the average effect size was .34, indicating the general elevating effect of pretest on posttest. Cognitive outcomes were raised .51, attitude outcomes .33, and others about .27 standard deviations. Seventy percent of all effects were positive, and 93% of the cognitive effects were positive. Duration of time between pre- and post-testing was not related to effect size, although effect size is generally smaller for durations over one month. Year of publication, sample size, presence of experimental treatment, and sameness or difference of pretest and posttest were not significantly related to effect size. Researchers must continue to include pretest as a design variable when it is present, and to estimate its effect. (Author/CP)

Score on the Mathematics Test of the National Teacher Examinations (NTE) was the dependent variable in a stepwise multiple regression using personological variables of in-service teachers as regressors. Two variables--number of semester hours credit in mathematics, and grade level of teaching--accounted for 61 percent of the observed NTE score variance. Predicted and obtained NTE scores were compared for a trichotomous division of teachers into low, middle, and high achievement. The prediction seems adequate for use in classifying teachers by achievement in research studies without requiring costly and time-consuming testing. (Author/HM)

This study focused on the evaluation of the delivery system of five different projects supported by the National Science Foundation. Seven general factors proposed by Havelock were used to identify the difference among various delivery systems. A series of on-site visits, interviews, observations, and questionnaires were used to describe the major characteristics of the delivery system of each project. A panel of nine judges used the collected information to rate each delivery system against Havelock's factors. A two-way analysis of variance with repeated measures was used to analyze the data. Results showed that differences existed among projects in the extent to which they met the factors taken as a group. (HM)

A major deficiency in classical test theory is the reliance on Pearson product-moment (PPM) correlation concepts in the definition of reliability. PPM measures are totally insensitive to first moment differences in tests which leads to the dubious assumption of essential tan-equivalence. Robinson proposed a measure of agreement that is sensitive to different test difficulty and gives a practical statistic to estimate reliability in the presence of known form variation in difficulty. Robinson's measure of agreement appears to be a useful alternative to the generalizability coefficient, as it provides a more conservative estimate of reliability under conditions of parallel form differences in mean. This is likely to be especially useful when examining inter rater reliability when internal consistency of the raters is poor. Robinson's measure does not seem advantageous for highly reliable parallel tests such as are encountered in standardized testing programs. A simulation study is presented to illustrate the degree of the coefficient's sensitivity to form difficulty variance. Robinson's measure of agreement and the intraclass correlation are computed for each simulation and their values are compared. (author/RL)

This study is based on a questionnaire survey of teachers trained over a three-year period in a program centered at the University of Wyoming, the Portal School Project (PSP), and a similar survey of a random sample of teachers in Rocky Mountain States neighboring the region served by the PSP. Teachers were asked certain demographic questions and if they had used certain National Science Foundation (NSF) curricula for each semester between 1970 and 1974. They were also asked to rate their level of usage of the curricula. Possible urban-rural differences, cross-year differences, and junior-senior high differences for secondary curricula were investigated. Use of NSF curricula was about 40% for elementary and 68% for secondary curricula. The PSP group tended to use the NSF curricula longer than the control sample. (Author/BB)

The derivation and analysis presented in this report allow the experimenter to perform multifactor contrasts on correlation coefficients for several populations in an experimental design. Cells of the design constitute independent groups of subjects on which two measures have been taken and correlated. Measures need not be identical for each group. An example is given in which the dependent variable represents the correlation between attitude and achievement for six secondary school science classes stratified by ability (three levels) and grade level (two levels). An overall relationship was observed between science attitude and knowledge of science for junior and senior high school students of science. No difference due to achievement was found, but the grade difference was significant. (Author/BB)

Generalizability theory is based upon analysis of variance (ANOVA) and requires estimation of variance components for the ANOVA design under consideration in order to compute either G (Generalizability) or D (Decision) coefficients. Estimation of variance components has a number of alternative methods available using SAS, BMDP, and ad hoc procedures. This study examines the effect of these methods on coefficeints for a large sample of subjects under complex designs. The Wechsler Adult Intelligence Scale-Revised national standardization sample provided the data base. Subject (n=1880) variables (age, gender, etc.) provided variety for the construction of complex factoral designs. Three designs were used to investigate variance component estimation and resultant g-coefficeints. The designs included: (1) a two-factor subject x item design; (2) a three-factor subject x item x age design; and (3) a four-factor subject x item x age x gender design. These designs allowed estimation of both within and between subject effects. Suggestions are made concerning equal and unequal situations. (Author/PN)

The current state of usage of regression models in analysis of variance (ANOVA) designs is empirically examined, and examples of several statistical errors made in usage are presented. The assumptions of the general linear model are that all predictors are known without error of measurement and are fixed with no replication or sample variation; in the population, errors are normally distributed independently with variance, and errors are independent of all predictors. The rules for construction of the ANOVA allow the expected mean squares to hold just as if the levels of each factor had been randomly sampled. Analysis of Covariance (ANCOVA) combines the elements of regression analysis with design, albeit in a restricted manner. The homogeneity of regression coefficients is a parameter restriction from the design view point. The regression weight associated with a given covariate level is discussed. Most regression approaches to ANOVA and ANCOVA assume a fixed factor model under all design specifications. A major oversight of expected mean squares has contributed to the current lack of concern for the level of generalizability warranted from the design specification. Aptitude treatment interaction models are examined. (Author/CM)

Methodological deficiencies inherent in expert-novice reading research make it impossible to draw inferences about curriculum change. First, comparisons of intact groups are often used as a basis for making causal inferences about how observed characteristics affect behaviors. While comparing different groups is not by itself a useless activity, progressing directly to training is premature at best. Second, the think-aloud protocol technique is often used for inferring a subject's cognitive structure of subject matter. This method is inappropriate because it assumes that the organization of this structure resides consciously in a person's mind and can be verbally reproduced. Third, retrospective methods have been employed to infer causality by selecting groups currently differing and discovering differences in their past on putative causal variables, which are then inferred to have caused the present differences. While this technique must be used in historical analyses, it becomes suspect when the inferences are used to speculate on implications for current practice. Finally, techniques employed in naturalistic inquiry often confuse a change in methodology with a change in the discipline being studied, and rely heavily on impressionistic, one-shot observation for many facts. (JD)