19.10.2014 Views

01 NRDC Dyslexia 1-88 update - Texthelp

01 NRDC Dyslexia 1-88 update - Texthelp

01 NRDC Dyslexia 1-88 update - Texthelp

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Developmental dyslexia in adults: a research review 79<br />

as a training not in first aid but in reconstructive surgery.<br />

Appendix 9 lists some useful books for teachers of reading.<br />

Teachers of both initial reading and adult literacy need to be well-informed about language<br />

and its acquisition.<br />

Adult literacy tutors need skills that are analogous with those of reconstructive surgeons,<br />

not first-aiders (see above).<br />

Evaluation studies<br />

Introduction<br />

An evaluation of a single intervention may ask only one question: ‘Is the intervention<br />

effective?’ A comparative evaluation will ask two questions: ‘Is either intervention effective?’<br />

and ‘Is one intervention more effective than the other one?’<br />

The appearance of effectiveness is no guarantee that the intervention is effective in reality. In<br />

naturalistic situations, where conditions likely to influence the outcome can not be<br />

standardised or even in some cases, identified, evaluation research is difficult. Often, there<br />

are a number of possible alternative explanations. The conscientious researcher will seek to<br />

rule them out before claiming effectiveness for an intervention. However, it is not uncommon<br />

for studies in this field to be undertaken by untrained researchers who have not been alerted<br />

to the pitfalls in their procedures and whose claims, in consequence, are flawed (although the<br />

flaws are seldom, if ever, a bar to media coverage).<br />

The validity of evaluation studies may be subject to a number of challenges (Gliner & Morgan,<br />

2000; Harrington et al., 2002; Pawson & Tilley, 1997; Torgerson et al., 2002; Troia, 1999). One<br />

such challenge comes from illusory change. Simply by statistical ‘regression’, or the testretest<br />

effect, an extreme score on initial testing may be followed by a more normal score on<br />

subsequent testing for reasons unrelated to anything that has happened in the interim. If the<br />

initial score is untypically low, possibly because of the unfamiliarity of the testing situation,<br />

the later score may be a more accurate reflection of the participant’s normal performance.<br />

A second group of challenges is associated with genuine changes which bear no specific<br />

relation to the intervention. Normal maturation will occur over time, irrespective of any<br />

intervention. A placebo (or expectation) effect may occur when the intervention gives<br />

participants the motivation to make their own adjustments (perhaps by reading more than<br />

they would otherwise have done) while a similar but temporary distortion may occur when<br />

unusual care and attention are lavished on the participants (so that they feel happier and<br />

become more receptive to ideas).<br />

The validity of evaluation studies can be challenged in a third way by shortcomings on the part<br />

of the experimenter. It may happen that conditions for inclusion in either the participant or<br />

control group are poorly specified, so that either or both groups are heterogeneous in ways<br />

that could affect the outcome of any comparison in their final test scores. There is a<br />

particular hazard from experimenter bias, no matter how innocent researchers’ intentions<br />

are, when they stand to make financial gains from an intervention programme of their own

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!