Teaching Secondary Students to Write Effectively

(2015); Olson and Land (2008); Olson et 

al. (2016). 

75  Kim et al. (2011); Olson and Land (2008); 

Olson et al. (2016).

76  Gallagher, Woodworth, and Arshan 

(2015).

77  Ruth and Murphy (1988). 

78  Smith and Swain (2011). 

79  Mueller (2014).

80  Kentucky Department of Education (2012).

81 Ibid.

82  C.B. Olson. (2015).

83  National Writing Project (2016).

84  Adapted from “Thinking of Teaching” blog 

(http://thinkingofteaching.blogspot.ca/).

85  Lewin and Shoemaker (2011). 

86  Graham (in press).

87  Duke (2000).

88  The gradual release of responsibility model 

was coined by Pearson and Gallagher (1983).

89  Following WWC guidelines, improved out-

comes are indicated by either a positive 

statistically significant effect or a positive, 

substantively important effect size. The WWC 

defines substantively important, or large, 

effects on outcomes to be those with effect 

sizes greater than or equal to 0.25 standard 

deviations. See the WWC guidelines at http://

whatworks.ed.gov.

90  For more information, see the WWC Fre-

quently Asked Questions page for practice 

guides at http://whatworks.ed.gov.

91  This includes randomized control trials (RCTs) 

and quasi-experimental design studies (QEDs). 

Studies not contributing to levels of evidence 

include single-case designs (SCDs) evaluated 

with WWC pilot SCD standards and regression 

discontinuity designs (RDDs) evaluated with 

pilot RDD standards.

92  The research may include studies generally 

meeting WWC group design standards and 

supporting the effectiveness of a program, 

practice, or approach with small sample sizes 

and/or other conditions of implementation 

or analysis that limit generalizability. The 

research may include studies that support the 

generality of a relation but do not meet WWC 

group design standards; however, they have 

no major flaws related to internal validity 

other than lack of demonstrated equivalence 

at pretest for QEDs. QEDs without equivalence 

must include a pretest covariate as a statisti-

cal control for selection bias. These studies 

must be accompanied by at least one relevant 

study meeting WWC design standards. For 

this practice guide, the latter studies did not 

need to be considered because a sufficient 

number of studies meet WWC design stan-

dards for each recommendation.

93  American Educational Research Association, 

American Psychological Association, and 

National Council on Measurement in Educa-

tion (1999).

94  Fitzgerald and Markham (1987).

95  Murphy and Smith (2015).

96  Olson, Scarcella, and Matuchniak (2015).

97  Eligible studies that meet WWC group design 

standards with or without reservations are 

indicated by bold text in the endnotes and 

references pages.

98  A statistically significant finding is a finding 

that is unlikely to occur by chance. 

99  Substantively important findings are defined 

as those with an effect size greater than or 

equal to 0.25 or less than or equal to –0.25, 

as measured by Hedge’s g.

---

( 

87

 )