The components of SRL measured in recent studies vary significantly. In their investigations of SRL, the studies reviewed here included measures of motivation, learning strategies, self-concept, metacognition, learning behaviors, learning-related skills, and false belief. This variety of constructs suggests that there exists some debate as to the components of SRL. This is consistent with findings reported by Dinsmore and colleagues (2008), who examined theoretical definitions and measurement strategies in 255 studies on metacognition, self-regulation, and/or SRL. They found variability
across studies in both the definition and measurement of SRL, although the congruence between the definition and the measure within a study was greatest for SRL compared to metacognition and self-regulation.
As with many of the skills discussed in this report, self-regulation is difficult to observe. Many measurement approaches rely on students to report whether and how they are engaging in self-regulation, so the very act of measuring self- regulation intervenes in the student’s learning environment and may affect the skill being investigated. Reliance on self-report also limits what can be learned about self-regulation in younger children, who are not as able to articulate their mental processes. At those ages, SRL research depends more heavily on parent and teacher ratings of those observable behaviors that are assumed to be indicative of psychological events relevant to self-regulation. This methodological limitation may hamper or alter researchers’ understandings of the early components and processes that are part of SRL. This section gives an overview of the number and types of measures currently used in the field. Table 4-1 lists SRL measures and some of their key features, including psychometric information when available in the articles reviewed.
Measuring Self-Regulated Learning as an Aptitude
Winne and Perry (2000) described SRL as either an aptitude or an event.
When SRL is seen as an aptitude, it is abstracted over multiple self-regulation events and measurement formats. SRL may be measured using questionnaires, structured interviews, and parent or teacher ratings. SRL as an aptitude is found to vary within individuals over time, across tasks and settings, and across individuals, which raises the question of whether it is stable enough to be called an aptitude, or a trait.
Self-Report Questionnaires. Self-report questionnaires are prevalent because of their convenience, low cost, and simplicity. SRL self-report questionnaires usually ask students to generalize across learning experiences and may be administered in concert with or separate from SRL tasks. In this review of the literature, no single self-report questionnaire was found to be used with much greater frequency than any other (see Table 4-1). Eleven self-report measures were identified, and only one of these was used in more than one study. This variety of measures suggests that researchers are still struggling to define SRL sufficiently or that multiple types of SRL exist, each of which requires a different set of questionnaire items.
As noted above, younger students will not be able to respond to self- report questionnaires because of literacy requirements and the metacognitive demands inherent in such instruments. Accordingly, studies that used self- report measures started around the beginning of middle school. Studies of children younger than this relied on parent or teacher ratings or observational measures.
Structured Interviews. This type of SRL measure consists of a highly structured set of specific items, often with skip patterns determined by students’
responses. Structured interviews are different from think-aloud procedures (described below) because they do not take place during a specific learning task. An example of a structured interview for SRL is the Self-Regulated Learning Interview Schedule (SRLIS) (Zimmerman & Martinez-Pons, 1986).
One challenge to working with structured interviews is the need to train coders to score the content.
Measuring Self-Regulated Learning as an Event
When SRL is seen as an event, it is a more localized phenomenon that is defined with a beginning and end point in time. Measures of SRL as an event include think-aloud measures, error detection tasks, trace methodologies, parent and teacher ratings, and observations.
Think-Aloud Measures. Think-aloud measures vary in how structured they are, but they all ask students to report verbally on their cognitive processes while they are engaged in a specific learning task. Researchers rely on think- aloud protocols to help them map out models of SRL. As with self-report questionnaires, these measures are not likely to be as well suited for use with younger populations who may have insufficient vocabulary to communicate their mental processes.
Error Detection Tasks. This type of SRL measure is designed to introduce errors into task materials and then observe whether students detect the errors and, if so, how students proceed. Students may or may not be told beforehand that errors are present, and their detection of the errors may be measured by asking them to mark the errors found or by eye fixations, which assume students will attend longer to errors than other task features.
Table 4-1. Measures of motivation: Key features
Measure Name Data Source Subscales or Components
No. of Studies Using this
Measure
Intended
population psychometric properties Peg-tapping measure
of inhibitory control, item selection measure of attention shifting, unexpected contents and changed locations tasks
Observation 1 3- to 5-year-olds Not reported.
Children’s Behavior Questionnaire
Teacher/parent report
Anger, approach, attention, and inhibitory control
1 3- to 5-year-olds 16 items
Internal consistency reliability (α): .74 to .91 (teacher ratings); .51 to .71 (parent ratings)
(Source: Blair & Razza, 2007) Strategic Flexibility
Questionnaire (SFQ)
Self-report Adaptiveness, inflexibility, and irresoluteness in self- regulatory control
1 10th graders 21 items
Internal consistency reliability (α): .80 (Source: Cantwell, 1998)
Survey of Learning Behaviors (SLB)
Self-report Self-regulation: self-monitoring subscale; Self-regulation:
knowledge acquisition subscale, self-efficacy scale
1 15- to 22-year olds 29 items (13 ,9, 7 per scale)
Internal consistency (α): .61 to .89 depending on subscale (Source: Chularut & DeBacker, 2004)
Self-Regulation Strategy Inventory—Self-Report (SRSI-SR)
Self-report 1 9th and 10th
graders
45 items
Internal consistency reliability (α): .92 (composite); .72 to .88 (subscales) Construct validity for composite score: All three scales from the SRSI-SR loaded onto one higher-order factor, with factor loadings ranging from .83 to .71.
Discriminant validity: The SRSI-SR scales loaded on a separate high-order factor than other measures (i.e., the Task Interest Inventory and Perceived Instrumentality Inventory).
(Source: Cleary, 2006) Self-Regulatory Skills
Measurement Questionnaire (SRSMQ)
Self-report 1 6th and 7th graders 33 items
Internal consistency reliability (α): .86 (Source: Eom & Reiser, 2000) Self-Regulation Test for
Children (SRTC)
Self-report NA 1 5- to 8-year-olds 1 item
Test-retest reliability: r = .92 Convergent validity: r = .35 to .40 (Source: Howse, Lange, et al., 2003) Instrumental Competence
Scale for Children (COMPSCALE)
Teacher report Motivation, behavior 2 5- to 8-year-olds 18 items
Test-retest reliability: r =.86 Predictive validity: r = .61
(Source: Howse, Calkins, et al., 2003) The Emotion Regulation
Checklist
Parent report Negativity/liability scale, emotion regulation scale
1 Preschool and
kindergarten students
Internal consistency reliability (α): .77 (negativity/liability scale), .68 (emotion regulation scale)
Convergent validity: r = .50 (Source: Howse, Calkins, et al., 2003)
Table 4-1. Measures of motivation: Key features
Measure Name Data Source Subscales or Components
No. of Studies Using this
Measure
Intended
population psychometric properties Peg-tapping measure
of inhibitory control, item selection measure of attention shifting, unexpected contents and changed locations tasks
Observation 1 3- to 5-year-olds Not reported.
Children’s Behavior Questionnaire
Teacher/parent report
Anger, approach, attention, and inhibitory control
1 3- to 5-year-olds 16 items
Internal consistency reliability (α): .74 to .91 (teacher ratings); .51 to .71 (parent ratings)
(Source: Blair & Razza, 2007) Strategic Flexibility
Questionnaire (SFQ)
Self-report Adaptiveness, inflexibility, and irresoluteness in self- regulatory control
1 10th graders 21 items
Internal consistency reliability (α): .80 (Source: Cantwell, 1998)
Survey of Learning Behaviors (SLB)
Self-report Self-regulation: self-monitoring subscale; Self-regulation:
knowledge acquisition subscale, self-efficacy scale
1 15- to 22-year olds 29 items (13 ,9, 7 per scale)
Internal consistency (α): .61 to .89 depending on subscale (Source: Chularut & DeBacker, 2004)
Self-Regulation Strategy Inventory—Self-Report (SRSI-SR)
Self-report 1 9th and 10th
graders
45 items
Internal consistency reliability (α): .92 (composite); .72 to .88 (subscales) Construct validity for composite score: All three scales from the SRSI-SR loaded onto one higher-order factor, with factor loadings ranging from .83 to .71.
Discriminant validity: The SRSI-SR scales loaded on a separate high-order factor than other measures (i.e., the Task Interest Inventory and Perceived Instrumentality Inventory).
(Source: Cleary, 2006) Self-Regulatory Skills
Measurement Questionnaire (SRSMQ)
Self-report 1 6th and 7th graders 33 items
Internal consistency reliability (α): .86 (Source: Eom & Reiser, 2000) Self-Regulation Test for
Children (SRTC)
Self-report NA 1 5- to 8-year-olds 1 item
Test-retest reliability: r = .92 Convergent validity: r = .35 to .40 (Source: Howse, Lange, et al., 2003) Instrumental Competence
Scale for Children (COMPSCALE)
Teacher report Motivation, behavior 2 5- to 8-year-olds 18 items
Test-retest reliability: r =.86 Predictive validity: r = .61
(Source: Howse, Calkins, et al., 2003) The Emotion Regulation
Checklist
Parent report Negativity/liability scale, emotion regulation scale
1 Preschool and
kindergarten students
Internal consistency reliability (α): .77 (negativity/liability scale), .68 (emotion regulation scale)
Convergent validity: r = .50 (Source: Howse, Calkins, et al., 2003)
Table 4-1. Measures of motivation: Key features
Measure Name Data Source Subscales or Components
No. of Studies Using this
Measure
Intended
population psychometric properties Laboratory Assessment of
Temperament–Preschool Edition
Observation Frustration 1 Preschool students Inter-rater reliability of coding: Three variables were coded for two tasks:
latency to frustration, duration of frustration, and intensity of frustration.
Reliabilities for the coded latency to frustration scores were within 2 seconds difference in 86% of the judgments.
Duration of frustration codes were reliable within 2 seconds difference for 86% of the judgments.
Intensity of frustration: Cohen’s kappa = .74 (Toy task), .88 (Circle task);
percent agreement = .82; (Toy task), .93 (Circle task) (Source: Howse, Calkins, et al., 2003)
Control, Agency, and Means–
Ends Interview (CAMI)
Self-report Agency subscales for effort and ability
1 8- to 11-year-olds Not reported.
State Measurement Scale Self-report Awareness, self-checking, planning, cognitive strategy use, and effort
1 10th through 12th
graders
Unknown number of items
Construct validity for composite scale: Confirmatory factor analysis determined that the four subdomains had large intercorrelations (.85 to .97).
(Source: Malpass et al., 1999) Cooper-Farran Behavioral
Rating Scales (CFBRS) (Cooper & Farran, 1991)
Teacher report Work-related subscale 1 Elementary
students
16 items
Internal consistency reliability (α): .94 to .99 depending on item (Source: McClelland et al., 2006)
School Attitude Assessment Survey (SAAS)
Self-report Academic self-perceptions, attitude toward school, motivation/self-regulation, and peer attitudes
1 Middle and high
school students
51 items
Internal consistency (α): .85 to .89 (subscales)
Construct validity: Confirmatory factor analysis of the four-factor model yielded comparative fit index = .95.
(Source: McCoach, 2000) Self-Regulated Learning
Interview Schedule (SRLIS)
Self-report Motivation, metacognitive, behavioral
2 High school
students
Not reported.
Motivated Strategies for Learning Questionnaire (MSLQ)
Self-report Self-regulated learning (self-regulation and cognitive strategies subscales), motivation (intrinsic motivation and self-efficacy subscales)
1 5th- and 6th-grade
students
Internal consistency reliability (α): .67 to .83 (subscales) (Source: Shores & Shannon, 2007)
Cognitive Assessment System (CAS)
(Das & Naglieri, 1985)
Observation Visual search, crack-the-code from the planning scale
1 6th graders Not reported.
Self-Regulated Learning Questionnaire (SRLQ)
Self-report Self-concept, motivation, learning strategies
1 8th graders Not reported.
(continued)
Table 4-1. Measures of motivation: Key features
Measure Name Data Source Subscales or Components
No. of Studies Using this
Measure
Intended
population psychometric properties Laboratory Assessment of
Temperament–Preschool Edition
Observation Frustration 1 Preschool students Inter-rater reliability of coding: Three variables were coded for two tasks:
latency to frustration, duration of frustration, and intensity of frustration.
Reliabilities for the coded latency to frustration scores were within 2 seconds difference in 86% of the judgments.
Duration of frustration codes were reliable within 2 seconds difference for 86% of the judgments.
Intensity of frustration: Cohen’s kappa = .74 (Toy task), .88 (Circle task);
percent agreement = .82; (Toy task), .93 (Circle task) (Source: Howse, Calkins, et al., 2003)
Control, Agency, and Means–
Ends Interview (CAMI)
Self-report Agency subscales for effort and ability
1 8- to 11-year-olds Not reported.
State Measurement Scale Self-report Awareness, self-checking, planning, cognitive strategy use, and effort
1 10th through 12th
graders
Unknown number of items
Construct validity for composite scale: Confirmatory factor analysis determined that the four subdomains had large intercorrelations (.85 to .97).
(Source: Malpass et al., 1999) Cooper-Farran Behavioral
Rating Scales (CFBRS) (Cooper & Farran, 1991)
Teacher report Work-related subscale 1 Elementary
students
16 items
Internal consistency reliability (α): .94 to .99 depending on item (Source: McClelland et al., 2006)
School Attitude Assessment Survey (SAAS)
Self-report Academic self-perceptions, attitude toward school, motivation/self-regulation, and peer attitudes
1 Middle and high
school students
51 items
Internal consistency (α): .85 to .89 (subscales)
Construct validity: Confirmatory factor analysis of the four-factor model yielded comparative fit index = .95.
(Source: McCoach, 2000) Self-Regulated Learning
Interview Schedule (SRLIS)
Self-report Motivation, metacognitive, behavioral
2 High school
students
Not reported.
Motivated Strategies for Learning Questionnaire (MSLQ)
Self-report Self-regulated learning (self-regulation and cognitive strategies subscales), motivation (intrinsic motivation and self-efficacy subscales)
1 5th- and 6th-grade
students
Internal consistency reliability (α): .67 to .83 (subscales) (Source: Shores & Shannon, 2007)
Cognitive Assessment System (CAS)
(Das & Naglieri, 1985)
Observation Visual search, crack-the-code from the planning scale
1 6th graders Not reported.
Self-Regulated Learning Questionnaire (SRLQ)
Self-report Self-concept, motivation, learning strategies
1 8th graders Not reported.
Trace Methodologies. Traces are “observable indicators about cognition that students create as they engage with a task” (Winne & Perry, 2000, p. 551), for example, “the frequency and pattern of highlighting text, accessing various supports for learning, and obtaining feedback from efforts to learn”
(Zimmerman, 2008, p. 170). Other types of trace methods include think-aloud protocols and study diaries. These traces record students’ methods of learning and when linked with immediate academic outcomes, students can see which strategies work best for them. Trace methods are more easily used in older students who have the metacognitive abilities necessary to report on their SRL processes.
Parent and Teacher Ratings. Ratings of observable SRL behaviors can be provided by parents, teachers, or other educators. Parent and teacher ratings may not be preferable to self-report due to the internal nature of much of SRL, but the advantage of such ratings is that they can provide information for young children who cannot report these phenomena easily for themselves.
Observations. Observational measures of SRL behaviors provide advantages over some other methods because they collect information on the context of the student’s behavior, and they can be used with even the youngest students.
Observational data are sometimes supplemented with student interviews or quantitative data collected through self-report measures or student records.