Progress in International Reading Literacy Study (pirls)

Teacher population. 
The target teacher population consists 
of all teachers linked to the selected students. Note that these 
teachers are therefore not a representative sample of 
teachers within an education system. Rather, they are the 
teachers who teach a representative sample of students in 
grade 4 within the education system. 
School population. 
The target school population consists of 
all eligible schools containing one or more fourth-grade 
classrooms. 
Sample Design 
PIRLS uses a two-stage stratified cluster sample design. The 
first stage consists of a sample of schools, which may be 
stratified; the second stage consists of a sample of one or 
more classrooms from the target grade in sampled schools. 
First-stage sampling selects individual schools with a 
probability proportionate to size (PPS) approach, which 
means that the probability is proportional to the estimated 
number of students enrolled in the target grade. Substitution 
schools are also selected to replace any schools that are 
originally sampled but refuse to participate. The original 
and substitution schools are selected simultaneously. In the 
second stage of sampling, one or two fourth-grade classes 
are randomly sampled in each school. 
PIRLS guidelines call for a minimum of 150 schools to be 
sampled in each education system, with a minimum of 4,000 
students assessed. A sample of 150 schools yields 95 
percent confidence limits for school-level and classroom-
level mean estimates that are precise to within 16 percent of 
their standard deviations. Countries with small class sizes or 
less than 30 students per school are directed to consider 
sampling more schools, more classrooms per school, or 
both, to meet the minimum target of 4,000 tested students. 
For countries choosing to participate in both PIRLS and 
PIRLS Literacy, the required student sample size is 
doubled—i.e., around 8,000 sampled students. Countries 
could choose to select more schools or more classes within 
sampled schools to achieve the required sample size. 
Because ePIRLS is designed to be administered to students 
also taking PIRLS, the PIRLS sample size requirement 
remains the same for countries choosing also to participate 
in ePIRLS. 
In the United States, the PIRLS 2001 sample consisted of 
3,763 fourth-grade students from 174 schools (after 
substitution). In 2006, the U.S. sample consisted of 5,190 
fourth-grade students from 183 schools (after substitution). 
For the 2011 data collection, there were 370 U.S. schools, 
after substitution, consisting of 12,726 fourth-grade 
students. The reason for a larger sample in 2011 was due to 
the coinciding administration of the Trends in International 
Mathematics and Science Study (TIMSS). To accommodate 
this concurrent administration, schools with at least two 
grade 4 classrooms were asked to participate in both studies, 
with one classroom being randomly assigned to TIMSS and 
the other to PIRLS. 
In the United States, one sample was drawn to represent the 
nation at grade 4 for PIRLS 2011. In addition to this national 
sample, a state public school sample was also drawn at 
grade 4 for Florida, which chose to participate in PIRLS 
separately from the nation in order to benchmark their 
student performance internationally. The sample frame for 
public schools in the United States was based on the 2011 
National Assessment of Educational Progress (NAEP) 
sampling frame. The 2011 NAEP sampling frame was based 
on the 2007–08 Common Core of Data (CCD). The PIRLS 
2011 data for private schools were from the 2007–08 Private 
School Universe Survey (PSS). Any school containing at 
least one grade 4 class was included in the school sampling 
frame. 
The U.S. PIRLS 2016 national school sample consisted of 
176 schools, which was higher than the international 
sampling minimum of 150 to offset anticipated school 
nonresponse and ineligibility. A total of 158 U.S. schools 
agreed to participate in PIRLS 2016, including 131 from the 
original sample and 27 sampled as replacements for 
nonparticipating schools from the original sample. Of the 
158 U.S. schools that participated in PIRLS, 153 also