katherine morgan graduate program: master of science in applied statistics

1
Katherine Morgan Graduate Program: Master of Science in Applied Statistics Case 1- Behavior Background : Functional Analysis (FA): used to identify variables that influence the occurrence of problem behaviors (i.e. noncompliance, disruptions, self-injurious behaviors…). Problematic behavior is identified and data collected to investigate under which situations the behavior occurs most often. Clinical specialist uses information to write specific behavior protocols to help reduce the behavior. Current Process : FA data collection includes: date of occurrence, antecedent precededing behavior(i.e. transition, demand, wants something and cannot have it), frequency of occurrence. Excel used to sum all frequencies per antecedent. Problem : Current method does not account for any environmental factors (such as time of day or therapist). Improvement : Utilized one case/student enrolled in an Early Intervention Program to demonstrate the effectiveness of SAS for investigating the function of behaviors. Frequency data was collected on noncompliance from Variables: time of day (AM or PM), therapist, antecedent, and task. Findings/Outcomes : Using Proc Freq procedure, identified that the student engages in noncompliant behaviors most often when the antecedent is “1” (demand at table) and when working with therapist “K” (table 1). SAS also useful to determine with which antecedent and task the behavior is most likely to co-occur (table 2). Case 1 SAS Code Case 2 - Academic Table 1 – Number of Behaviors by Therapist Using SAS Macros to Summarize Academic Achievement and Functions of Problematic Behaviors *Research Question 1: Which situations dose behavior occur?; %macro freq (data=, var=); proc sort data=&data; by &var; run; ods rtf; title "Number of Noncompliances"; title2 "By &Var for June 15 - June 19 2009"; proc freq data=&data; weight frequency; tables &var/chisq; run; ods rtf close; %mend freq; %freq (data=fa.data, var=antecedent) %freq (data=fa.data, var=Time_of_day) %freq (data=fa.data, var=task) *Research Question 2: Relationship between variables?; %macro freq2 (data=, var=, var2=); proc sort data=&data; by &var &var2; run; ods rtf; title "Number of Noncompliances"; title2 "By &Var and &var2 for June 15 - June 19 2009"; proc freq data=&data; weight frequency; tables &var*&var2/chisq cmh; run; ods rtf close; %mend freq2; %freq2 (data=data, var=therapist, var2=time_of_day) %freq2 (data=data, var=therapist, var2=task) %freq2 (data=data, var=therapist, var2=antecedent) %freq2 (data=data, var=task, var2=antecedent) Table 2: Partial Output of Frequency Table for Task by Antecedent Case 2 SAS Code *Research Question 1: Differnce between Enrollment Status; %Macro mastery (data=, var=, byvar=); proc sort data=&data; by &var; run; proc boxplot data=&data; Title "Boxplot of Number of Goals &Var by &byvar"; plot &var*&byvar/boxconnect=mean vaxis=axis2 haxis=axis1; inset mean stddev min max/ header = 'Overall Statistics' pos=tm; insetgroup mean min max stddev/header= 'Statistics by &byvar'; run; Proc anova data=&data; Title “One way ANOVA”; class &byvar; model = &var * &byvar/snk; Run; mend mastery; %mastery (data=student, var=mastery, byvar=FT) *Research Question 2: Difference between 2008 and 2009; proc sql; create table score_diff as select student08.id, mastery08, mastery09, mastered08, mastered09 from student08, student09 where student08.id=student09.id; quit; proc ttest data=score_diff; title 'Paried T-test of Percent Mastery by School Year'; paired mastery08*mastery09; run; Figure 1: Boxplot of Number of Goals Mastered By Classroom for School Year 2008 Background : Goal of the Early Intervention Program is to develop readiness skills to enable successful transition to community schools and kindergarten programs. The Early Intervention Program uses applied behavior analysis (ABA) to document and evaluate the child’s progress within their personalized intervention plans that include: academic, social, cognitive, and self-help skills. Improvement : SAS used to demonstrate how Early Intervention staff (Director, clinical specialists, and classroom teachers) may monitor progress by individual student of class aggregate. Proc Boxplot used to provide visual representation of the distribution of acquired academic goals between classrooms (Figure 1). Findings: Proc ANOVA was used to drill down differences in the average number of skills acquired by classroom; Results suggested that there is no difference between classrooms (F=1.08, p=0.3662). SAS used to create a table to be used to track student progress over time. Using Proc SQL , a table was created by combining 2 different data sets – each containing percent mastery of academic skills for a particular school year. This resulting table may be used to track student progress over time. Reference: Burlew, Michele. (2007). SAS Macro Programming Made Easy. SAS Institute Inc., Cary, NC. SAS Macros contain two main components: SAS macro variables and SAS macro programs. Macro variables can be used when you need to repeatedly insert a piece of text throughout the program. Macro programs use both Macro variables and programming statements in building SAS programs. Macro Programs are useful for doing repetitive tasks when analyzing data. The purpose of this project was to demonstrate how Macro programs can be used at an Early Intervention Program with children with Autism. Specifically how they can (1) help Clinical Specialist determine the cause (or “function”) of a probleamatic behavior and (2) aid in summarizing student academic information more efficiently. Both of which are important in writing effective academic and behavior interventions. Macro Benefits : The use of Macros has given me the flexibility to create a number of frequency tables using a variety of variable combinations. Staff in the Early Intervention Program are now able to investigate under which situation problematic behaviors occur. Previously, protocols were based on an analysis of one variable. Now, when writing behavior protocols, staff have access to more information. Another great benefit to the staff and the students is that protocols are more tailored to the individual student.

Upload: hisano

Post on 05-Jan-2016

24 views

Category:

Documents


0 download

DESCRIPTION

Using SAS Macros to Summarize Academic Achievement and Functions of Problematic Behaviors. Katherine Morgan Graduate Program: Master of Science in Applied Statistics. Case 1- Behavior. Case 2 - Academic. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Katherine Morgan Graduate Program: Master of Science in Applied Statistics

Katherine Morgan

Graduate Program: Master of Science in Applied Statistics

Case 1- Behavior

Background: Functional Analysis (FA): used to identify variables that influence the occurrence of problem behaviors (i.e. noncompliance, disruptions, self-injurious behaviors…). Problematic behavior is identified and data collected to investigate under which situations the behavior occurs most often. Clinical specialist uses information to write specific behavior protocols to help reduce the behavior.

Current Process: FA data collection includes: date of occurrence, antecedent precededing behavior(i.e. transition, demand, wants something and cannot have it), frequency of occurrence. Excel used to sum all frequencies per antecedent.

Problem: Current method does not account for any environmental factors (such as time of day or therapist).

Improvement: Utilized one case/student enrolled in an Early Intervention Program to demonstrate the effectiveness of SAS for investigating the function of behaviors. Frequency data was collected on noncompliance from Variables: time of day (AM or PM), therapist, antecedent, and task.

Findings/Outcomes: Using Proc Freq procedure, identified that the student engages in noncompliant behaviors most often when the antecedent is “1” (demand at table) and when working with therapist “K” (table 1). SAS also useful to determine with which antecedent and task the behavior is most likely to co-occur (table 2).

Case 1 SAS Code

Case 2 - Academic

Table 1 – Number of Behaviors

by Therapist

Using SAS Macros to Summarize Academic Achievement and Functions of Problematic Behaviors

*Research Question 1: Which situations dose behavior occur?; %macro freq (data=, var=);proc sort data=&data;by &var;run;ods rtf;title "Number of Noncompliances";title2 "By &Var for June 15 - June 19 2009";proc freq data=&data;weight frequency;tables &var/chisq;run;ods rtf close;%mend freq;%freq (data=fa.data, var=antecedent)%freq (data=fa.data, var=Time_of_day)%freq (data=fa.data, var=task)

*Research Question 2: Relationship between variables?;%macro freq2 (data=, var=, var2=);proc sort data=&data;by &var &var2;run;ods rtf;title "Number of Noncompliances";title2 "By &Var and &var2 for June 15 - June 19 2009";proc freq data=&data;weight frequency;tables &var*&var2/chisq cmh;run;ods rtf close;%mend freq2;%freq2 (data=data, var=therapist, var2=time_of_day)%freq2 (data=data, var=therapist, var2=task)%freq2 (data=data, var=therapist, var2=antecedent)%freq2 (data=data, var=task, var2=antecedent)

Table 2: Partial Output of Frequency Table for

Task by Antecedent

Case 2 SAS Code*Research Question 1: Differnce between Enrollment Status;%Macro mastery (data=, var=, byvar=);proc sort data=&data;

by &var; run;proc boxplot data=&data;

Title "Boxplot of Number of Goals &Var by &byvar";plot &var*&byvar/boxconnect=mean vaxis=axis2

haxis=axis1;inset mean stddev min max/ header = 'Overall

Statistics' pos=tm;insetgroup mean min max stddev/header= 'Statistics by

&byvar';run;Proc anova data=&data; Title “One way ANOVA”; class &byvar; model = &var * &byvar/snk;Run;mend mastery;%mastery (data=student, var=mastery, byvar=FT)

*Research Question 2: Difference between 2008 and 2009;proc sql;

create table score_diff asselect student08.id, mastery08, mastery09, mastered08,

mastered09from student08, student09where student08.id=student09.id;

quit;

proc ttest data=score_diff;title 'Paried T-test of Percent Mastery by School Year';paired mastery08*mastery09; run;

Figure 1: Boxplot of Number of Goals Mastered

By Classroom for School Year 2008

Background: Goal of the Early Intervention Program is to develop readiness skills to enable successful transition to community schools and kindergarten programs. The Early Intervention Program uses applied behavior analysis (ABA) to document and evaluate the child’s progress within their personalized intervention plans that include: academic, social, cognitive, and self-help skills.

Improvement:

SAS used to demonstrate how Early Intervention staff (Director, clinical specialists, and classroom teachers) may monitor progress by individual student of class aggregate. Proc Boxplot used to provide visual representation of the distribution of acquired academic goals between classrooms (Figure 1).

Findings: Proc ANOVA was used to drill down differences in the average number of skills acquired by classroom; Results suggested that there is no difference between classrooms (F=1.08, p=0.3662).

SAS used to create a table to be used to track student progress over time. Using Proc SQL , a table was created by combining 2 different data sets – each containing percent mastery of academic skills for a particular school year. This resulting table may be used to track student progress over time.

Reference:

Burlew, Michele. (2007). SAS Macro Programming Made Easy. SAS Institute Inc., Cary, NC.

SAS Macros contain two main components: SAS macro variables and SAS macro programs. Macro variables can be used when you need to repeatedly insert a piece of text throughout the program. Macro programs use both Macro variables and programming statements in building SAS programs. Macro Programs are useful for doing repetitive tasks when analyzing data. The purpose of this project was to demonstrate how Macro programs can be used at an Early Intervention Program with children with Autism. Specifically how they can (1) help Clinical Specialist determine the cause (or “function”) of a probleamatic behavior and (2) aid in summarizing student academic information more efficiently. Both of which are important in writing effective academic and behavior interventions.

Macro Benefits: The use of Macros has given me the flexibility to create a number of frequency tables using a variety of variable combinations. Staff in the Early Intervention Program are now able to investigate under which situation problematic behaviors occur. Previously, protocols were based on an analysis of one variable. Now, when writing behavior protocols, staff have access to more information. Another great benefit to the staff and the students is that protocols are more tailored to the individual student.