Presentation on theme: "Aimee R. Holt, PhD Middle Tennessee State University"— Presentation transcript:
1Aimee R. Holt, PhD Middle Tennessee State University Data Analysis within an RtI2 Framework: Linking Assessment to InterventionAimee R. Holt, PhDMiddle Tennessee State University
2A systematic and data-based method for addressing academic concerns: What is RTI2?A systematic and data-based method for addressing academic concerns:identifyingdefining &resolvingBrown-Chidsey & Steege (2010)RtI2 is a way of thinking about and doing instruction. It is a mindset focused on finding the instruction that works best for each child.
3RTI2 is a general education initiative…. Components of RTI2High-quality instructionFrequent assessment of academic skillsData-based decision makingBrown-Chidsey & Steege (2010)Components of RTI2High-quality instruction (evidenced based)Frequent assessment of academic skillsData-based decision makingThe purpose of assessment throughout the RtI2 process is to facilitate instruction.Within an RtI2 framework assessment FOR learning replaces assessment OF learning.
4Problem Identification Problem SolvingAt each tier within RTI2, a problem solving model is employed to make decisionsProblem IdentificationAnalyze the Results of ImplementationDetermine Next StepsAnalyze the Assessment Plan ResultsDevelop an Intervention PlanDefine the ProblemDevelop an Assessment PlanImplement PlanProgress MonitorPlan EvaluationProblem Analysis
6Universal Screeners LEAs are required to: Administer a nationally normed,skills-based universal screenerto students at their grade levelLEAs are required to:Administer a nationally normed, skills-based universal screener to students at their grade levelThe screener selected must be Skill-based – because it will be used to determine whether students demonstrate the skills necessary to achieve grade level standards.Universal screening data can indicate which students need additional instruction. They are NOT diagnostic. They are a type of formative assessment rather than summative.While the primary goal of summative assessment is to determine how much has been learned, formative assessment data can be used to identify specific skills that need to be taught and how best to teach them.Tier 1 instruction should be effective for about 80% of the students. When fewer than 80% of the students are successful with Tier 1, core instruction needs to be IMPROVED
7For K-8, Universal Screeners should be administered 3X per year In grades 9-12, there are multiple sources of data that can be reviewed, such as:EXPLORE, PLAN and ACT; Tennessee Comprehensive Assessment Program (TCAP) which includes Writing (TCAP-WA), End of Course (EOC), 3-8 Achievement and in , Partnership for Assessment of Readiness for College and Careers (PARCC); TVAAS
8Characteristics of Appropriate Universal Screening Tools Helps answer questions about efficiency of core programAligns with curriculum for each grade levelSkills mastery aligns with state mandated year-end assessmentIkeda, Neessen, & Witt (2008).Timing the measures increases their standardization and sensitivity to change (Riley-Tillman et al., 2013)
9General Outcome Measures (GOM’s) Skill Based Measures 3 Types of CBM’sGeneral Outcome Measures (GOM’s)Skill Based MeasuresSub-skill Mastery Measures
10General Outcome Measures GOMssample performanceacross several goals at the same timecapstone tasksEx. Oral reading fluencyCan be used forscreening (benchmarking),survey & specific level assessmentprogress monitoringGOMs are used to sample performance across several goals at the same time by using capstone tasks that are complex in the sense that they can only be accomplished successfully by applying a number of contributing skillsEx. Oral reading fluencyCan be used for screening (benchmarking), survey & specific level assessment as well as progress monitoring
11Skills-Based Measures SBM are similar to GOM’s but can be used when capstone tasks are not availableEx. Math computationCan be used forscreening (benchmarking),survey & specific level assessmentprogress monitoringSkills- based measures are designed to accomplish many of the functions of GOM’s but can be used when capstone tasks are not availableEx. Math computationCan be used forscreening (benchmarking),survey & specific level assessmentprogress monitoring
12Subskill Mastery Measures SMMs are very narrow in focusEx. Names of lettersShould not be used for benchmarking(exception… early skills such as Letter Naming Fluency, Letter Sound Fluency, Number Naming Fluency)
13Example Reading Skills Typically Assessed by Universal Screeners GradeAreas Typically Assessed by Universal Screeners6thOral Reading Fluency ; Reading for understanding5thOral Reading Fluency; Reading for understanding4thOral Reading Fluency; Reading for understanding3rd2nd1stLetter Naming Fluency (beginning); Phonemic Awareness; Phonics; Word Identification Fluency;Oral Reading Fluency (end)KLetter Naming Fluency; Phonemic Awareness:Early Phonics Skills including Letter Sound Fluency
15Making Decisions about Group Data Review universal screening data to answer the following questions:Is there a class wide problem?Who needs a Tier II intervention?Be sure to examine students at the marginDoes anyone need Tier III now?
16As a guideline, students below the 25th percentile would be considered “at-risk.” Students who exceed grade level expectations may be considered advanced.
17Who needs a Tier II or Enrichment? Winter Benchmark for ORF:90th %- 153;25th % - 72;Winter Benchmark for Maze:90th % - 25;25th % - 9;Instructional level criteriaFor contextual reading – 93-97% correctFor most other academic skills – 85-90% correctORFMaze26 /98%154/100%154/85%26 /79%68/ 95%09 /94%Example of 3rd Grade Winter Benchmark DataRow 1 needs enrichmentRow 2 is sacrificing speed for accuracy – focus should e on increasing accuracy – can be done with instruction in Tier 1Row 3 is accurate but slow; probably needs Tier 2 intervention focused on fluencyRow 4 is slow and inaccurate; possibly needs Tier 2 intervention focused on phonics skills development68/88%08 /80%
18Examining students at the Margins Winter Benchmark for ORF:90th %- 153;25th % - 72;Winter Benchmark for Maze:90th % - 25;25th % - 9;Instructional level criteriaFor contextual reading – 93-97% correctORFMaze75/96%11 /100%80/100%10 /97%73/82%11/75%Example of 3rd Grade Winter Benchmark DataRow 1: Fluency is just above 25th % but accuracy is good; Probably does not need Tier 2 - work on fluency in Tier 1Row 2: Although mazes is slow the student is accurate and ORF is well above the 25th % and highly accurate; might not need Tier 2Row 3: Fluency is just above 25% but accuracy is poor; Probably does need Tier 2 instruction
19Identifying who needs Tier III Winter Benchmark for ORF:25th % - 72;10th % -44Winter Benchmark for Maze:25th % - 9;10th % - 6Instructional level criteriaFor contextual reading – 93-97% correctORFMaze46 / 76%6 / 80%42 / 83%5 / 75%
20Referral to Tier II Decision Tree Core literacy instruction has been implemented with fidelity≥80% of student needs are met by core instructionDifferentiated instruction has been provided in a small group within core literacy instructionStudent has been present for ≥75% of instructional daysStudent has passed vision and hearing screeningData indicates performance below the 25th% on universal screening of student achievement compared to national normsAdditional Assessment data supports universal screening data
21What do we mean by linking assessment to intervention?
22Linking Assessment to Interventions…. Research has shown that effective interventions have certain features in common:Correctly targeted to the student’s deficitAppropriate level of challenge (instructional range)Explicit instruction in the skillFrequent opportunities to practice (respond)Provide immediate corrective feedback(e.g., Brown-Chidsey & Steege, 2010; Burns, Riley-Tillman, & VanDerHeyden, 2013; Burns, VanDerHeyden, & Boice, 2008;)An intervention can not be effective if it is not targeting the student’s skill deficit within a skills based hierarchy.Once the target and the level of challenge have been appropriately identified, then interventions should be linked to that data. The intervention identified that include explicit instruction, with frequent opportunities for the student to practice those skills. Immediate corrective feedback will be essential.
23Academic Instruction in Reading Both NCLB and IDEA require that instruction in the general education setting cover all 5 areas of reading identified by the National Reading PanelPhonemic AwarenessPhonicsFluencyVocabularyText Comprehension Strategies
24Linking the 5 skill areas to 3 SLD areas Basic Word ReadingPhonemic AwarenessPhonicsReading FluencyFluencyReading ComprehensionVocabularyText Comprehension Strategies
25Phonological Awareness A metacognitive understanding that words we hear have internal structures based on soundResearch on PA has shown that it exerts an independent causal influence on word-level reading. (Berninger & Wagner, 2008)Phoneme – smallest unit of speechThe English language has phonemes
26PhonicsAlphabetic principle - Linking phonological (sound) and orthographic (symbol) features of language (Joseph, 2006)Important for learning how to read and spellNational Reading Panel –students with explicit AP instruction showed benefits through the 6th gradePhonological awareness is a prerequisite skill
27Word Reading Skills - (McCormick, 2003) Word identification: the instance when a reader accesses one or more strategies to aid in reading words (e.g., applying phonic rules or using analogies)Decoding – blending sounds in words or using letters in words to cue the sounds of others in a word (Joseph, 2006)Word recognition: the instant recall of words or reading words by sight; automaticity
28Fluency“ The ability to read a text quickly, accurately, and with proper expression” (NRP, p.3-5)Most definitions of fluency include an emphasis on prosody – the ability to read with correct expression, intonation and phrasing (Fletcher et al., 2007)National Reading Panel -Good reading fluency skills improved recognition of novel words, expression during reading, accuracy and comprehensionWhen decoding is an automatic process, reading a connected text becomes effortless and therefore requires little conscious attention; thus more cognitive resources are available for higher-order processing of the meaning of the text (Fletcher et al, 2007)Fluent readers have learned to automatically recognize words and to increase reading rate while maintaining accuracy.Fluency provides a bridge between word recognition and comprehension.Fluent readers recognize words and comprehend at the same time.
29Vocabulary & Text Comprehension Skills Vocabulary knowledge – including understanding multiple meanings of words; figurative language etc..Identifying stated detailsSequencing eventsRecognizing cause and effect relationshipsDifferentiating facts from opinionsRecognizing main ideas – getting the gist of the passageMaking inferencesDrawing conclusionsAlthough they can often recall details, many students with RC difficulties struggle with getting the gist (i.e., formulating main ideas)This leads to problems summarizing and/or drawing inferencesThey are less accurate answering explicit (stated directly in the reading) questions based on their readings than those without RC difficulties and they can have significant difficulties answering implicit (must make inferences to answer) questionsThese problems may also occur when listening
31So you have identified your “at risk students”- now what? You will need to conduct Survey Level Assessment (SLA) for these studentsSurvey Level Assessment (SLA)Can be used to: (a) provide information on the difference between prior knowledge and skills deficits to be used to plan instructional interventions & (b) serve as baseline for progress monitoringBecause of the potential number of students involved in Tier 2 activities the focus should be on efficiency.Note: Baseline data is a a minimum 3 stable data points. If the baseline data are variable then more data points will need to be collected.Tier 2 should not require in-depth functional analysis, however, some additional data might be necessary.Students needing Tier 3 services will require more in depth functional assessments before they begin their interventions.
32“What is the CATEGORY of the problem” Why is it important to conduct Survey Level Assessments before beginning Tier II interventions?The primary question being addressed by the survey level assessment at Tier II is“What is the CATEGORY of the problem”(What is the specific area of academic deficit?)(e.g., Riley-Tillman, Burns, Gibbons, 2013)Evidenced based interventions are only valid for specific purposes with certain groups of studentThe universal screening data may only tell you that a problem exists, NOT what the category of the intervention needs to beIf you mismatch an evidenced based intervention with a problem it was not designed to address, there is NO reason to think it will work.
33An Example of Survey Level Assessment Using DIBELS 1) Start at student’s grade level2)Test backwards by grade until the student has reached the “low risk” benchmark for a given skill•Low risk/ established indicates the student has “mastered” that skillGradeCBM AssessedBenchmarked6thOral Reading FluencyFall, Winter, Spring5th4th3rd2nd1stWinter, SpringNonsense Word FluencyPhoneme Segmentation FluencyLetter Naming FluencyFallKWinter, SpringLetter Naming FluencyInitial Sound FluencyFall, Winter
34For example….. In reading comprehension & fluency =comprehension interventioncomprehension low fluency, but decoding =fluency interventioncomprehension fluency decoding, but phonemic awareness skillsdecoding interventionRiley-Tillman et al., (2013)It might be inappropriate to place a 3th grade student who does poorly on MAZE fluency in a comprehension group especially if they are still having decoding problems.In reading:A student with low comprehension but sufficient fluency would likely benefit from a comprehension interventionA student with low comprehension and low fluency, but acceptable decoding, would likely need a fluency interventionA student with low comprehension, low fluency, and low decoding, but sufficient phonemic awareness skills would likely require a decoding intervention
35Let’s look at Michael a 2nd grade student At the fall benchmark, he was identified on ORF as being in the some risk range.His score was 30 wcpmSurvey level assessment were conducted using:DORF 1st grade – (fluency)DNWF 1st grade – (decoding)DPSF 1st grade – (phonemic awareness)ProblemIdentificationProblemAnalysis
37What next….You link your assessment data to an intervention that targets the category of skill deficit that was identifiedYou select progress monitoring probe(s) that assess that skillYou set the student’s goal for improvementYou can use ROI & Gap Analysis Worksheets to help with this
38What progress monitoring is not… It is NOT an instructional method or interventionThink of progress monitoring as a template that can be laid over goals and objectives from an assortment of content areas
41Referral to Tier III Decision Tree Tier II intervention(s) have occurred daily for 30 minutes in addition to core instructionIntervention logs attached(3) Fidelity checks completed and attachedImplementation integrity has occurred with at least 80% fidelityStudent has been present for ≥75% of intervention sessions Tier II intervention(s) adequately addressed the student’s area of need
42 Tier II intervention was appropriate and research-based Research based interventions are:□ Explicit□ Systematic□ Standardized□ Peer reviewed□ Reliable/valid□ Able to be replicatedProgress monitoring has occurred with at least weekly data points –OR bi-monthly data points Gap analysis indicates that student’s progress is not sufficient for making adequate growth with current interventions
43Does a student require Tier III intervention? Step 1: Need to check to see if the data can be interpretedA minimum of 8-10 data points, if progress monitoring every other week, OR data points, if progress monitoring weekly to make a data-based decision to change to Tier III.Christ (2006) found that formative assessment data can fluctuate especially the greater the age of the student being monitored, therefore a minimum of approximately 8 to 10 data points are necessary to make reliable decisions.
44Step 2: Examine Rate of Improvement You can compare the student’s actual ROI to the goal that was establishedYou can use the ROI worksheetsLet’s complete one for Michael
45Completing the ROI Worksheet for Michael Assessment Used: DIBELS NWFStudent’s score on first probe administered: 28Student’s score on last probe administered: 37Fall benchmark expectation: 24Spring benchmark expectation: 50 Step 1____________-_____________/_________=___________Spring benchmark expectationFall benchmark expectationNumber of weeksTypical ROI (slope)First we calculate the typical ROI for this skill5024360.72
463728130.69Next we will determine the student’s current ROI
471.440.720.721.08Now take the Typical ROI we calculated in Step 1 which was 0.72 and multiple it X 2 for an aggressive ROI goal or 1.5 for a reasonable ROI goal.Next determine if the student’s ROI which was 0.69 < the Reasonable or Aggressive ROI goal – in this case it is
48You also can visually analyze the graphed progress monitoring data Calculate the trend line of the intervention data points and compare it to the aim (goal) line.If the slope of the trend line is less than the slope of the aim line, the student may need to be moved to Tier III.Especially if it appears that given the student’s current ROI that they will not meet year end grade level standards
49Dual Discrepancy-A student should be deficient in level and have a poor response to evidenced-based interventions (slope) to the degree that he/she is unlikely to meet benchmarks in a reasonable amount of time without intensive instruction to move:between Tier II to Tier III as well as between Tier III and referral for a comprehensive special education evaluation.(e.g., Brown-Chidsey & Steege, 2008; Lichenstien, 2008)
51Specific Level Assessment Functional analysis of skillsAre used to:(a) identify specific skills deficits;(b) students prior knowledge; &(c) serve as baseline for progress monitoringspecific level assessments rely primarily on subskill mastery measures.“drill down” to specific deficitsFocus on a functional analysis of the learner’s skillsAre used to:identify specific skills deficits;(b) provide SP with additional information about a students prior knowledge; &(c) serve as baseline for progress monitoringWhile survey level assessments rely primarily on GOMs and some skill based measure, specific level assessments rely primarily on subskill mastery measures.Specific level assessments “drill down” to specific deficits (i.e. vowel patterns)
52Functional Analysis RIOT/ICEL Matrix R- review I – interview O – observeT - testI – instructionC – curriculumE – environmentL- learnerProblems in Tier III should be explored through multiple sources of data and possible hypotheses should be generated.
53Linking Assessment Data to Intervention at Tier III The learnerfocus on alterable learner variablesidentify academic entry level skillsThe tasklevel of the material the student is expected to masterThe instructionresearch-based methods and management strategies used to deliver curriculumMatch = SuccessInstructionStudentTaskSAY 1st:We conduct a functional analysis of academic skills within an instructional hierarchy to determine what factors may be contributing to the mismatch between actual and desired levels of performance.Prerequisite skills are analyzed to determine if any weak or missing.
54Targets for Academic Instructional Materials Instructional levelcontextual reading – 93-97% correctother academic skills – 85-90% correctProduce larger gains more quicklyGravois, T.A., & Gickling, E.E. (2008). Best practices in instructional assessment. In A. Thomas & J.Grimes (Eds.), Best practices in school psychology (5th ed., pp ). Bethesda, MD: National Association of School Psychologists.Instructional level criteriaFor contextual reading – 93-97% correctFor most other academic skills – 85-90% correctResearch has shown that interventions with materials that stay within this range of difficulty produce larger gains more quickly than those that do not.
55Phonemic Awareness Hierarchy identifying initial, final & medial sounds in wordsAlliterationblending individual sounds to make a whole wordBlendingbreaking a whole word into it’s individual partsSegmentingDeleting: saying the new word created by omitting a syllable or individual sound in a wordSubstituting: changing the initial, final, or medial sound in a word to create a new wordReversing: saying the sounds of a word in reverse order to create a new wordManipulatingIn order to recommend an appropriately targeted intervention, it is important to identify the level within the hierarchy where the student is in the instructional range. Examples of intervention activities for each of these levels can be found at the Florida Center for Reading Research website.Daly, Chafouleas, & Skinner (2005)
56Let’s look at Michael again….. Specific Level Assessment –Phonics:Decoding Skills testDevelopmental Spelling AnalysisSight words:Graded word listPhonemic Awareness:LAC 3ProblemAnalysisContinuing the case we began in the module addressing Tier II, as you recall, Michael was a 2nd grade student referred because he fell below benchmark in the fall on ORF. Survey level assessments determined that he was still having difficulty with phonics/decoding skills but had established phoneme segmenting skills for 3 & 4 phoneme words as measured by DIBELS. He received 13 week of a standard protocol intervention focused on phonics. Although he made gains his ROI was not acceptable and the team decided to refer him to Tier III.Following the problem solving model, his problem has been identified in general terms – his ORF is below grade level and lack of phonics skills is considered to be contributing to his low reading fluency skills, however, now in Tier III problem analysis is going to much deeper to try to establish the functional relation between prior knowledge/skills and his limited ROI
57Linking specific level assessment data to interventions…. Basing interventions on direct samples of student’s academic skills has been shown to result in larger effect sizes than interventions derived from other dataThis is also known as a skill by treatment interactionBurns, Codding, Boice & Lukito, (2010)While normative cognitive and neuropsychological assessment measures can be used as part of specific level assessment data collections, how one uses these measures to guide intervention design should be done with the focus on academic skills.For example, lets consider two middle school students with a student difficulties related to math problem solving due. Sam is having difficulties because of limited math vocabulary knowledge related to deficits in comprehension knowledge. Tami’s difficulties with math problem solving are due to her not being able to engage in multiple step processing to reach a solution as well as difficulties with determining essential from nonessential information which are skills related to fluid reasoning or even executive functioning. These two students would certainly needs a different focus during a Tier 3 intervention. The focus of the intervention in both cases should revolve around acquiring specific math skills not on addressing the underlying processing deficits.
60Level Central location of data within a phase often compared to benchmark (goal/aim line)can also look at mean or median for each phase(e.g., Daly III et all., 2010; Hixson et al., 2008; Riley-Tillman & Burns, 2009)Can conduct a Gap Analysis using the worksheet
61How the central location changes over time Slope/TrendHow the central location changes over timeWith academic data we are usually looking for an increase in skillsTarget students ROI can be compared with peer groups ROI or benchmark(e.g., Daly III et all., 2010; Hixson et al., 2008; Riley-Tillman & Burns, 2009)
622 approaches for analyzing slope Calculate ROI and compare to an identified peer group using the ROI worksheetPlot the trend line and compare the aim (goal) line to the slope (trend) line
63Variability Should be examined both within and between phases General rule- most of the variability in the data should explained by the trend line80% of the data points should fall with in 15% of the trend line
65Referral for SLD Evaluation Decision Tree Tier III Intervention(s) have occurred daily for 60 minutes in addition to core instructionIntervention logs attached(5) Fidelity checks completed and attachedImplementation integrity has occurred with at least 80% fidelityStudent has been present for ≥75% of intervention sessionsTier III intervention(s) adequately addressed the student’s area of need
66Referral for SLD Evaluation Decision Tree Tier III intervention was appropriate and research-basedResearch based interventions are:□ Explicit□ Systematic□ Standardized□ Peer reviewed□ Reliable/valid□ Able to be replicatedProgress monitoring has occurred with at least weekly data points –OR bi-monthly data points at Tier IIIGap analysis indicates that student’s progress is not sufficient for making adequate growth with current interventions
67Referral for SLD Evaluation Decision Tree The following have preliminarily been ruled out as the primary cause of the student’s lack of response to intervention□ Visual, motor, or hearing disability□ Emotional disturbance□ Cultural factors□ Environmental or economic factors□ Limited English proficiency□ Excessive absenteeism
68Deciding to refer for SLD evaluation As part of the teams decision to refer for an SLD evaluation, a Gap Analysis should be conductedLet’s look at how to complete the Gap Analysis worksheet with Michael
69Current benchmark expectation Gap AnalysisAssessment Used: 2nd ORFStudent’s current benchmark performance: 66Student’s current rate of improvement (ROI): 1.3Current benchmark expectation:90End of year benchmark expectation:Number of weeks left in the school year:5Is Gap Significant?________/ ________= _________□ Yes □ NoCurrent benchmark expectationCurrent performanceCurrent gapFallWinterSpring90661.4
70Conducting a Gap Analysis Step 29066242454.8241.318
72SEMAdditionally, we cannot ignore issues such as interpreting CBM scores in light of SEM or CI when those scores are used for such as diagnoses and eligibility determinationsFor more detailed discussion including suggested SEM guidelines for oral reading fluency scores in grades 1-5 see:Christ, T. J. Silberglitt, B. (2007). Estimates of the standard error of measurement for curriculum-based measures of oral reading fluency. School Psychology Review, 36, ppJust as SEM and CI are psychometric issues important for the interpretation of normative achievement and IQ measures, in the future we can expect that we will need to consider these issues with CBM data. Researchers are beginning to explore this issue in more detail. Here is a source for those of you who might be interested in reading more about these issues.
73Use of Progress Monitoring in Special Education Because CBM datacan be directly tied to skill development necessary to be successful in the curriculum,they possess a higher level of sensitivity, andallows for graphic representation;they allows for development of a higher quality IEPProgress monitoring should continue after the IEP is initiatedExit criteria can be set to determine if early reevaluation can be completed due to student success.