Abstract
General Description
The Comprehensive Test of Phonological Processing–Second Edition (CTOPP-2; Wagner, Torgesen, Rashotte, & Pearson, 2013) is an individually administered, norm-referenced measure designed to assess phonological processing skills related to reading. This instrument evaluates three components: phonological awareness, phonological memory, and rapid naming. Components are assessed through subtests measuring elision, blending, phoneme isolation, short-term memory, rapid naming, and segmentation.
The CTOPP-2 is appropriate for individuals ages 4-0 through 24-11. The test consists of 10 core and two supplemental subtests. Nine core and one supplemental subtest are for ages 4 to 6; seven core and both supplemental subtests are for ages 7 to 24. Four core subtests are timed. Completion of the core subtests requires approximately 30 min.
The CTOPP-2 has undergone considerable modification since the first edition. New normative data were collected, and items were added to eliminate floor effects and extend the age range downward. A Phoneme Isolation subtest was added for ages 7 to 24. Furthermore, the Examiner’s Manual was modified to include justification for the inclusion of rapid naming subtests and to present additional confirmatory factor analyses.
Scoring
The CTOPP-2 has five composites (M = 100, SD = 15). Phonological Awareness, Phonological Memory, and Rapid Symbolic Naming composites can be calculated for all ages, whereas the Rapid Non-Symbolic Naming composite is for ages 4 to 6 and the Alternate Phonological Awareness composite is for ages 7 to 24. Additional derived scores include subtest scaled scores (M = 10, SD = 3), age and grade equivalents, percentile ranks, and developmental scores.
Test Materials and Stimuli
Materials include an Examiner’s Manual, a Picture Book, Examiner Record Booklets for ages 4 to 6 and 7 to 24, and two audio CDs used for administering three core subtests and both supplemental subtests.
Technical Adequacy
Standardization
The CTOPP-2 normative sample included 1,900 individuals. Age levels were in 2-year intervals for ages 4 to 17; ages 18 to 24 were included as one group. Most age groups included an adequate number of participants, with at least 200 participants in each. However, the 18- to 24-year age group included too few participants, with only 115 participants for this large age range. The sample was representative of the 2010 U.S. Census based on geographic region, gender, ethnicity, Hispanic status, exceptionality status, family income, and education level of parents. There were no data for urban/rural residence. Stratification of selected demographic variables is provided by age.
Reliability
Bracken (1987, 2000) recommended that coefficients of at least .80 serve as the minimum acceptable standard by which to judge subtest reliability, and .90 for composites. Coefficient alphas were provided for untimed subtests at 15 age intervals. At all ages, Elision, Sound Matching, and Segmenting Nonwords had correlations of .80 or higher. Subtests with coefficients below .80 included Blending Words (ages 8 and 10), Phoneme Isolation (age 9), Memory for Digits (ages 5 to 7, 12, and 14), Nonword Repetition (ages 6 to 10, 12, 13, 15, and >17), and Blending Nonwords (ages 7, 9, and 10); these coefficients ranged from .70 to .79. The Phonological Awareness and Alternate Phonological Awareness composites met or exceeded .90 for all ages, except Phonological Awareness at age 10, which had a coefficient of .88. Coefficients for the Phonological Memory composite were lower, ranging from .81 to .88.
Because the Rapid Naming subtests are timed, typical methods for examining internal consistency reliability are inappropriate. Thus, to obtain an estimate of error for internal consistency, an alternate form was created and results for both forms were compared. Correlations met or exceeded .80 for the Rapid Object Naming and Rapid Color Naming subtests at all ages. Correlations were below .80 for Rapid Letter Naming at ages 4 and 12, and Rapid Digit Naming at age 13; these three coefficients ranged from .73 to .78. Composite correlations were all above .90, except Rapid Symbolic Naming at ages 4, 12, and 13, and for Rapid Non-Symbolic Naming at age 4; where coefficients ranged from .88 to .89. Thus, internal consistency of the CTOPP-2 is lower than desired for some subtests and composites at certain ages, but does not fall below .70 for subtests or .80 for composites.
For content bias, the authors evaluated the internal consistency for selected subgroups including gender, race, ethnicity, and learning disability status. Only 2 of 81 correlations were below .80 for subtests, and only 4 of 34 composites had correlations below .90. Thus, scores do not appear to be biased for these subgroups.
Stability of scores over time was assessed using a 1- to 2-week retest interval with 144 participants divided into three age groups. Many subtests had adequate reliability, but reliability for composites was problematic. Each of the age groups had three or more subtests with coefficients ranging from .70 to .79. Furthermore, Phonological Memory was the only composite with a correlation of at least .90 for ages 4 to 6. Composites at all other ages ranged from .72 to .89.
Validity
The format and items of each subtest were chosen based on expert feedback and a decade of research. Items that did not discriminate well between individuals with and without phonological deficits, or that were too easy or difficult were eliminated based on item discrimination and item difficulty analyses. Further results from logistic regression suggested no item bias in terms of gender, race, or ethnicity.
To demonstrate construct validity, five studies compared the CTOPP-2 with other measures of phonological processing. These included the Children’s Test of Nonword Repetition (Gathercole & Baddeley, 1996), the Early Reading Assessment (Hammill, Pearson, Hresko, & Hoover, 2012), the Rapid Automatized Naming and Rapid Alternating Stimulus Tests (Wolf & Denckla, 2005), the Test of Memory and Learning–Second Edition (Reynolds & Voress, 2007), and the Test of Word Reading Efficiency–Second Edition (Torgesen, Wagner, & Rashotte, 2012). Corrected correlations with other measures of phonological awareness ranged from .45 to .85 for subtests, and .72 to .82 for the composite. Corrected correlations with other measures of phonological memory ranged from .41 to .74 for subtests, and .71 to .76 for the composite. Finally, corrected correlations with other measures of rapid naming ranged from .25 to .95 for subtests, and .27 to .91 for the composite. Thus, in most cases, the CTOPP-2 results correlate well with other measures of phonological processing.
The correlations of CTOPP-2 results with the two measures of reading (the Test of Word Reading Efficiency–Second Edition (Torgesen et al., 2012) and the Early Reading Assessment (Hammill et al., 2012) also provided evidence of how well CTOPP-2 results predict reading performance. The corrected coefficients with these two tests ranged from .25 to .70 for the Phonological Awareness composite, .20 to .64 for the Phonological Memory composite, and .56 to .80 for the Rapid Symbolic Naming composite. Thus, the CTOPP-2 has small to very large correlations with measures of early reading.
Results of confirmatory factor analyses support the authors’ claim that the test measures three related but separate constructs of phonological processing. For ages 4 to 6, correlations between the composites range from .34 (Phonological Memory and Rapid Non-Symbolic Naming) to .73 (Phonological Awareness and Phonological Memory). For ages 7 to 24, correlations between the composites range from .26 (Phonological Memory and Rapid Symbolic Naming, and Phonological Awareness and Rapid Symbolic Naming) to .67 (Phonological Awareness and Phonological Memory).
CTOPP-2 performance was found to be related to age, with average raw scores increasing with age. Furthermore, stronger correlations were evident for the younger participants, which the authors attributed to the link between phonological processing and beginning reading skills. Furthermore, group differentiation was addressed by comparing mean standard scores for the normative sample to those of eight subgroups. As expected, participants with attention deficit/hyperactivity disorder (ADHD) or a learning disability displayed the most significant deviations from the normative sample mean scores, whereas other subgroup mean deviations were all within the average range (i.e., gender and ethnicity).
CTOPP-2 performance was found to improve following instruction in phonological processing and reading skills, with average effect sizes ranging from 0.6 to 0.7. This supports the instrument’s usefulness in documenting response to intervention. Finally, an examination of item validity suggested that items on each subtest are highly correlated with each other.
Floor effects and item gradients
Tests with inadequate floors do not have enough items for low-achieving children and may inflate test results (Bracken, 1987, 2000). To have adequate floors, a raw score of 1 should equate to a standard score at least two standard deviations below the mean. The CTOPP-2 has adequate floors for all subtests except the following: Elision, which does not have an adequate floor until age 6, and Blending Words, Sound Matching, and Blending Nonwords, which do not have adequate floors until age 5 to 6. The Phonological Awareness Composite does not have an adequate floor until age 5 to 6, but other composites have adequate floors. Thus, although the authors attempted to eliminate floor problems with this edition, examiners should consider that some composites and subtests do not have adequate floors until ages 5-6 or 6-0.
Finally, tests with steep item gradients do not discriminate well between children with and without deficits (Bracken, 1987, 2000). There are no item gradient concerns with the CTOPP-2.
Commentary and Recommendations
The CTOPP-2 can be a valuable tool in diagnosing phonological processing difficulties as well as for research purposes. A paucity of norm-referenced tests of phonological processing exists in general, but especially for 4-year-olds. Other instruments measuring this skill are not as extensive as the CTOPP-2, and typically only contain one or two subtests dedicated to phonological processing. In contrast, at all ages, the CTOPP-2 offers 9 to 10 subtests of sufficient length (30 or more items) measuring multiple aspects of phonological processing. This is important because this skill comprises several distinct abilities that cannot be measured by only one or two tasks.
The CTOPP-2’s suitability for use with preschoolers is its main advantage. Research has suggested that poor phonological processing is implicated in later reading difficulties (Bradley & Bryant, 1983). Therefore, early detection and intervention is important in preventing reading failure, and this instrument provides a means for early identification. However, examiners should be aware that due to floor problems, some scores may be inflated for 4- and 5-year-olds with deficits. There are other available instruments that do not have floor problems for 4-year-olds; however, these instruments do not measure phonological skills as extensively as the CTOPP-2.
Overall, the CTOPP-2 is an exciting addition to the range of norm-referenced assessments of phonological processing due to its depth and breadth as well as the extended age range. The research behind phonological processing, and the rationale for inclusion of each subtest and its format are also thoroughly explained in the manual. However, examiners should be cautious due to the CTOPP-2’s reliability. Internal consistency was lower than desired for some ages and subtests, as was test-retest reliability for several subtests and all the composites except for Phonological Memory at ages 4 to 6. Substantial evidence was presented in support of validity; however, because reliability is a prerequisite to validity, and because some of the indicators of reliability were lower than desired for the CTOPP-2, examiners should question whether results from insufficiently reliable subtests are valid. The CTOPP-2 correlates well with other measures of phonological processing and beginning reading measures, and items were selected based on research and careful analyses. Despite some limitations, the extensiveness of the skills measured by the CTOPP-2 contributes to its usefulness in identifying phonological processing deficits.
Footnotes
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.