This study evaluated the properties of a 37-item survey designed to measure a school's implementation of a multi-tiered support system (MTSS). Twenty-four elementary schools from one urban school district participated in the study and were grouped based on assigned level of implementation. Survey responses were collected from 148 different staff members. Internal consistency was evaluated using Cronbach's alpha which equaled 0.981, 95% CI [0.975, 0.986]. Interrater agreement, as measured by average measure intraclass correlation (ICC), was 0.979, 95% CI [0.973, 0.985] across all schools and raters. An ANOVA comparing the mean survey score total across groups did not reveal any significant differences [F (2, 96) = 0.41, p =0 .67], nor were there any significant correlations between survey score totals and student outcomes as measured by AIMSweb. Reliability of the survey was strong, but further refinement is needed to improve the survey's ability to discriminate between implementation levels.

This study examined the predictive validity and classification accuracy of four early literacy screeners with Spanish-speaking English language learners (Ss ELLs) at varying levels of English language proficiency. First grade Ss ELLs (N = 209) were screened in the fall and winter using Phoneme Segmentation Fluency (PSF), Developmental Spelling Test (DST), Nonsense Word Fluency (NWF), and Word Identification Fluency (WIF). The criterion measure administered in the spring was Oral Reading Fluency (ORF). Overall, the WIF screener was the strongest predictor of end-of-first grade ORF scores. At each assessment period, it was the most strongly correlated with ORF3 for the total sample and for students aggregated by language proficiency groups. Results of hierarchical regression models, with PSF, DST and DST-squared, NWF, and WIF entered in order, found that each addition resulted in a significant change in R-squared with respect to spring ORF scores. However, once all screeners were entered, WIF was the only significant predictor of ORF3 in the winter regression model, and one of two (with NWF) significant predictors in the fall. Language proficiency did not explain additional variance in ORF3 when added to the regression model after the literacy screeners. When it comes to classification accuracy, WIF had the largest area under the curve (AUC) of the fall and winter measures and the highest specificity (when sensitivity was selected as close to .90 as possible). Though it was not as strong as WIF’s, DST’s classification accuracy was much better than that of PSF. Findings indicate that the same cut scores can be used for ELLs at each level as for native English speakers (NESs). Results from this study contribute to the research suggesting that schools should use both WIF and a spelling measure, like DST, to assess first graders.