Last registered on November 17, 2018


Trial Information
General Information
Initial registration date
November 15, 2018
Last updated
November 17, 2018 9:55 PM EST

This section is unavailable to the public. Use the button below to request access to this information.

Request Information
Primary Investigator
Aarhus University
Other Primary Investigator(s)
Additional Trial Information
On going
Start date
End date
Secondary IDs
Social emotional learning programs have been found to lead to immediate improvements in cognitive, social and emotional competences. Meanwhile, most evidence to date refers to the United States, and most other countries lack locally tailored teaching materials for socio-emotional learning. Further, there is a lack of knowledge about which subgroups benefit more. Such knowledge is important, because it could provide evidence relevant for both explaining and addressing inequality in educational achieving across subgroups of pupils. Knowledge about longer-term impacts on academic achievement is also called for. This protocol describes an experimental evaluation of a recently developed social emotional learning program implemented in Denmark. The evaluation combines survey data with register-based data, where the latter source allows for tracking of participant outcomes with minimal risk of attrition.
External Link(s)
Registration Citation
Simonsen, Marianne. 2018. "Perspekt." AEA RCT Registry. November 17.
Experimental Details
Treatment classes receive instruction using PERSPEKT 2.0 - a set of teaching materials aimed at training pupils’ emotional, personal and social skills to improve individual well-being as well as the social and learning environment in the classroom. It fulfills the four criteria for best implementation practice, SAFE; it is Sequenced in that there is coordinated progression of activities and practices to build competencies of the pupils; it is Active as it includes a number of participatory elements, such as role plays; it is Focused in terms of having allocated specific time and program elements to build specific SEL competencies; and it is Explicit in terms of having identified specific SEL competencies, that it aims to strengthen [12]. The material bears resemblance to PATHS (Promoting Alternative Thinking Strategies) and Second Step, both widely used social-emotional learning programs developed in the US, which have been subjected to several RCT based evaluations.
PERSPEKT 2.0 exists in three age-appropriate modules (Module I, II and III), targeting respectively grades 0-3, 4-6 and 7-9. Treatment classes are instructed using module II, which targets grades 4-6 and consists of 15 chapters, each of which is designed to take 45-60 minutes to complete. Exercises in the material are a variation over conversations, classroom exercises and small group activities. Some chapters offer specific tools, such as key phrases or steps, for children to use in different situations. Roleplaying and games are included as a means of drawing attention to and practicing different skills.
In treatment classes, instruction in PERSPEKT 2.0 is initiated in August 2018, at the beginning of the school year. To the extent possible, instruction in successive chapters will be spaced by one week, however schools are allowed some flexibility in timing, in order to accommodate other planned activities (e.g. thematic weeks or class trips) and teacher absences. The entire course must be completed by the end of February 2019. Instruction can be provided by either teachers or pedagogues associated with the class. While it is recommended that the same instructor, typically the class teacher, teaches the entire course, up to two teachers may, under special circumstances, be involved.
PERSPEKT 2.0 is designed to require no special training of instructors. A pilot study conducted in three Danish primary schools during the school year 2017/18 has confirmed that the teaching materials are self-contained, and that teachers typically spend less than half an hour to prepare for each chapter. Instructors in treatment classes are introduced to the materials exclusively through a video that demonstrates classroom practice. The teaching material itself is available through a custom-built web application, though a printed version of the material is also available upon request. Teachers are equipped with personal usernames and passwords and once logged in, instructors can read the chapter and exercise instructions and display project exercise materials on a smartboard in the classroom. In addition, instructors can easily keep track of the progression of their class(es) through the material at the level of individual exercises. This delivery of the program through a user-friendly online platform is intended to make it easy to implement and scale the program.
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
Our primary outcome is a measure of attitudes towards school and emotional well-being in the classroom.
Primary Outcomes (explanation)
We base the measure on the recently implemented national well-being indicators that are collected during the first quarter of each calendar year. For our primary outcome we use responses to the survey from the calendar year following the provision of treatment (i.e. within the same school year as treatment is administered).
Secondary Outcomes
Secondary Outcomes (end points)
Pupil academic performance; pupil problem behavior; pupil social emotional learning skills; pupil emotional distress; teacher absenteeism
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
We employ a two-level cluster randomized trial for children in two adjacent school cohorts (Grades 4 and 5) within the same school. There is otherwise no blocking..
Experimental Design Details
Not available
Randomization Method
We computerized the randomization via an unpredictable random sequence using STATA. In practice, we generated a sequence of 77 uniformly distributed random numbers on the (0,1) interval corresponding to the number of schools at the time of randomization. This sequence was applied to an ordered list of the numerical school identifiers. Values larger than 0.5 indicate that 5th grade pupils are offered PERSPEKT 2.0, while 4th grade pupils are offered TAU, and vice versa for values lower than 0.5.
Randomization Unit
School cohort. We have randomly allocated schools into teaching PERSPEKT 2.0 in either 4th grade or 5th grade such that all schools implement PERSPEKT 2.0 in only one of the two grade levels
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
77 schools
Sample size: planned number of observations
7200 pupils
Sample size (or number of clusters) by treatment arms
77 school cohorts in treatment; 77 school cohorts in control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Our power analysis uses data for 4th and 5th graders in the 2016/2017 academic year at the 77 schools enrolled in the study with consent to share administrative data. We measure the intraclass correlation coefficient to .056 and the coefficient of variation to .51. With a power of 0.80 and a significance level of 0.05, using a two-sample means t-test without conditioning on covariates results in a Minimum Detectable Effect (MDE) of 0.080 on our main outcome variable, the social well-being indicator. Conditioning on grade level (4th or 5th grade), the baseline outcome and the interaction between the two yields an MDE of 0.070. In both calculations we cluster standard errors at the school-grade-level. An MDE of 0.07 corresponds to 0.110 standard deviations on the social well-being indicator.
IRB Name
IRB Approval Date
IRB Approval Number