Executive Summary
Chapter 3 Sampling and Response Rates
3.2 Sampling Procedures
from that of primary and secondary schools. The steps involved are listed as follows:
1. A total of 17 special schools belonging to the above 4 categories were included in the sampling frame.
2. For each stratum (i.e. each category of schools), schools were sampled purposefully by seeking advice from experts in the special school sector on the general level of ICT use for learning and teaching in the schools included in the sampling frame.
3. Two additional schools were also selected from each category as the replacement schools, with the exception of the category Schools for Children with Visual Impairment for which no replacement school was available.
3.2.1.2 Sampling of primary and secondary schools
The sampling process for primary and secondary schools was conducted as follows:
1. The sampling frames contained school identity number, school size and overall student ability level with reference to P4 and S1 enrolments in the 2005/2006 school year (The target grades for administration of the PAs in the 2006/2007 academic year were P5 and S2 for the primary and secondary levels respectively. However, the sampling had to be completed before summer 2006. Therefore, the sampling frame was based on P4 and S1 enrolment information.). The schools were grouped into sampling strata based on the overall student ability level.
2. The number of schools to be sampled in each stratum was determined using the following formula:
60
grade target entire in the students of
number total
stratum specific
in the students of
number total
3. Schools within a stratum were listed in descending order of school size (the number of students is known as the measure of size (mos)) in the target grade. The cumulative measure of size (cmos) is then calculated from the first to the last schools for all schools.
4. A sampling interval for primary and secondary schools was defined by dividing the total number of students in the entire population of the target grade in Hong Kong by the desired number of schools to be sampled. For example, the total number of students in P4 in January 2006 was 67493; therefore the sampling interval for primary schools was 1124.88.
5. A random number between 0 and 1 was then chosen from a random number table. For example, the random number selected for primary schools was 0.2975. This number was then multiplied by the sampling interval to give us the random number that would be used to start the selection procedure. In this case, the start number was 335. Given this starting random number, the 2nd, 3rd, 4th, …numbers were obtained by just adding the sampling interval to the initial random number, thus generating the numbers 335, 335+1124.88=1460; 1460+1124.88=2585, etc.
6. A school was selected into the sample if a number generated fell between the cmos of the preceding school and the cmos of that school. For example, in Table 3.1 below, school 7
was selected because the number 1460 is within the cmos of that school, namely, between 1409 and 1606. The 2 schools following a selected school on the list were then designated as replacement schools in case the selected school was unable to participate in the study.
For example, schools 3 and 4 are replacement schools for the sampled school 2.
Table 3.1 shows a partial listing of schools in the sampling frame to illustrate how the sample and replacement schools were selected.
Table 3.1 Sampling procedure of schools School
Identity Number
Implicit Stratum
Measure of Size (mos)
Cumulative Measure of Size (cmos)
Random Number Sample Status
1 H 303 303
2 H 243 546 335 Selected
3 H 234 780 Replacement 1
4 H 217 997 Replacement 2
5 H 212 1209
6 H 200 1409
7 H 197 1606 335+1124.88=1460 Selected
8 H 197 1803 Replacement 1
9 H 195 1998 Replacement 2
10 H 194 2192
11 H 191 2383
12 H 189 2572
3.2.2 Sampling of Classes
One intact class of the target grade level was sampled from each of the sampled schools to participate in the PAs. As it could generally be assumed that class sizes were very similar within the same school in Hong Kong, only random sampling was conducted at the class level for each sampled school to select one class out of all the classes at the target grade level in the school. The teachers teaching the sampled classes in the assessed KLAs (which were the Mathematics and Chinese Language teachers of the sampled class at the primary level, and the Science and Chinese Language teachers of the sampled class at the secondary level) formed the sample of teachers to complete the Teacher Questionnaire.
3.2.3 Sampling of Students
Three PAs (the technical proficiency tasks and the information literacy tasks for two KLAs) in addition to the Student Questionnaire were administered at primary and secondary levels to the students. PAs of the kind designed and administered in this study are actually cutting edge research
even at an international level and there are not many examples of such in the research literature.
Literature related to the design and administration of PAs in IL for special school students cannot be located. It is expected that special arrangements will need to be made for conducting PAs for students in special schools and the inclusion of special schools will allow us to explore the feasibility and necessary adaptations for using this type of PAs with special school students.
Therefore, only students in secondary 2 or equivalent were to take part in this study from the special school sector.
In order to reduce the assessment load on the sampled students, each student only had to complete two of the PAs. Hence, students in each of the sampled classes were randomly assigned into one of the 3 groups, each of which took a different combination of two out of the 3 PAs. An example of the detailed arrangement at the primary level is illustrated in Table 3.2.
Table 3.2 Performance assessments conducted at a sampled P5 class Sequence of Online
Tasks for Students
Group 1 Group 2 Group 3
1 Student Questionnaire Student Questionnaire Student Questionnaire
2 Technical
proficiency tasks
IL tasks for Chinese Language
IL tasks for Mathematics
3 IL tasks for
Chinese Language
IL tasks for Mathematics
Technical proficiency tasks
In order to reduce the scoring and coding load, it was planned that not all of the completed work from students would be marked. For primary and secondary schools, 4 students would be randomly sampled from each of the 3 groups of students so that there would be assessment results from 12 students in each sampled class resulting in a total student sample of 720 students at each level. For special schools, completed work of 6 students in each sampled class (i.e. 2 students for each PA) would be marked and resulted in a total of 30 students in the sample. Appendix 3.1 summarizes the sample sizes for the different groups of respondents in the three school sectors as proposed in the study.