Implementation fidelity for math intervention: Basic quality ratings to supplement adherence.

Although it is common for researchers to assess implementation fidelity (IF) within the context of math intervention, IF assessments are often restricted to intervention adherence. Further, the degree to which IF influences observed outcomes is commonly ignored. The current study examined the relationship between three aspects of IF and the math performance of 1,340 grade 4 through 8 students who participated in an evidence-based math intervention. A series of multilevel regression models were fit to the data, including a final model with an indicator of intervention adherence, intervention delivery quality, and intervention engagement. A significant and positive association was observed between students' math performance and intervention engagement; however, a similar relationship was not observed when examining the impact of adherence and the quality of intervention delivery on students' math performance. Results are discussed in the context of implementation research for educational interventions. (PsycINFO Database Record (c) 2020 APA, all rights reserved).

Accuracy of student performance while reading leveled books rated at their instructional level by a reading inventory.

Identifying a student's instructional level is necessary to ensure that students are appropriately challenged in reading. Informal reading inventories (IRIs) purport to assess the highest reading level at which a student can accurately decode and comprehend text. However, the use of IRIs in determining a student's instructional level has been questioned because of a lack of research. The current study examined the percentage of words read correctly with 64 second- and third-grade students while reading from texts at their instructional level as determined by an IRI. Students read for 1 min from three leveled texts that corresponded to their instructional level as measured by an IRI, and the percentage of words read correctly was recorded. The percentage read correctly correlated across the three books from r=.47 to r=.68 and instructional level categories correlated from tau=.59 to tau=.65. Percent agreement calculations showed that the categorical scores (frustration, instructional, and independent) for the three readings agreed approximately 67% to 70% of the time, which resulted in a kappa estimate of less than .50. Kappa coefficients of .70 are considered strong indicators of agreement. Moreover, more than half of the students with the lowest reading skills read at a frustration level when attempting to read books rated at their instructional level by an IRI. The current study questions how reliably and accurately IRIs identify students' instructional level for reading.