Educational Validity Revisited

1987 ◽  
Vol 12 (2) ◽  
pp. 96-102 ◽  
Author(s):  
David W. Test ◽  
Fred Spooner ◽  
Nancy L. Cooke

In 1983, Voeltz and Evans introduced a set of criteria for establishing educational validity. Their intent was to improve the documentation of quality educational programs for learners with severe disabilities. Although the concept of educational validity is sound, we feel that Voeltz and Evans were not justified in rejecting single-subject research methodology as a vehicle for assessing educational validity. The present paper (a) provides a summary of the arguments of Voeltz and Evans against the use of single-subject research designs in establishing educational validity, (b) addresses each of the major concerns of Voeltz and Evans with single-subject research methodology, and (c) demonstrates how single-subject research methodology can be used to demonstrate educational validity.

1987 ◽  
Vol 12 (2) ◽  
pp. 103-106 ◽  
Author(s):  
Ian M. Evans ◽  
Luanna H. Meyer

Our article on educational validity summarized the major questions to be addressed for the evaluation of educational outcomes in programs for students with severe disabilities (Voeltz & Evans, 1983). In particular, we argued that the predominant emphasis upon single-subject research designs and the demonstration of the internal validity of intervention experiments were not sufficient for educational validity—a concept that requires systematic attention to larger issues of meaning-fulness in relationship to criterion environments. In this paper we respond to the arguments of Test, Spooner, and Cooke (1987) that single-subject design methodologies are capable of expansion to address educational validity. Based upon both theory and empirical data, we maintain that the serious limitations of the existing traditional methodologies continue to be problematic, so that we encourage movement toward a more comprehensive evaluative framework. Such a framework is critical to ensure that services and practices for persons with severe disabilities will be guided by research findings as well as social values.


1985 ◽  
Vol 62 (8) ◽  
pp. 516-522 ◽  
Author(s):  
FRANK L. COLLINS ◽  
RUTH A. BAER ◽  
RONALD L. BLOUNT

2003 ◽  
Vol 13 (2) ◽  
pp. 133-144 ◽  
Author(s):  
Christopher F. Sharpley

Although the last 20years have seen a focus upon evidence-based therapies, there are arguments that much of the so-called “evidence” produced is, in fact, irrelevant to the mental health practitioner in the field, principally because of the use of large-scale group designs in clinical controlled studies of the effectiveness of one therapy over another. By contrast, and with particular relevance to the practitioner who is both scientist and therapist, single subject research designs and methodologies for data analysis can be applied in ways that allow for generalisation to everyday practice. To inform the readership, the rationale underlying n = 1 studies is described, with some explanation of the major designs and their application to typical cases in guidance and counselling. Issues of inferential deductions from data, variations of design, data analysis via visual and statistical procedures, and replication are discussed. Finally, a case is argued for the introduction of n = 1 reports within the Australian Journal of Guidance and Counselling to better inform the readership about clinical research findings relevant to their practices.


1989 ◽  
Vol 14 (2) ◽  
pp. 93-97 ◽  
Author(s):  
Floyd F. Robison ◽  
D. Keith Morran ◽  
Diana Hulse-killacky

2021 ◽  
pp. 143-154
Author(s):  
Charles Auerbach

Meta-analytic techniques can be used to aggregate evaluation results across studies. In the case of single-subject research designs, we could combine findings from evaluations with 5, 10 or 20 clients to determine, on average, how effective an intervention is. This is a more complex and sophisticated way of understanding differences across studies than reporting those changes qualitatively or simply reporting the individual effect sizes for each study. In this chapter, the authors discuss why meta-analysis is important to consider in single-subject research, particularly in the context of building research evidence. They then demonstrate how to do this using SSD for R functions. Building upon effect sizes, introduced in Chapter 4, the authors illustrate the conditions under which it is appropriate to use traditional effect sizes to conduct meta-analyses, how to introduce intervening variables, and how to evaluate statistical output. Additionally, the authors discuss and illustrate the computation and interpretation of a mean Non-Overlap of All Pairs in situations which traditional effect sizes cannot be used.


Sign in / Sign up

Export Citation Format

Share Document