Survey Design

Population

The population group of the KEEP consists of senior students (third grade students) among middle schools, high schools, and vocational and technical schools.

Population for the Panel Survey

The survey population of the KEEP nationwide is 4,175 middle schools and high schools that had more than 31 second grade students in 2003, and that were located in cities and towns across the country. (Refer to < Figure 1 > and < Figure 2 > )

  • Second grade students of 2003 graduated to third grade in 2004. Considering the level of student movements within and between schools, data was only collected from schools with over 31 students at second grade level.
<Figure 1> Selection of the Population for the Panel Survey
Classification Total number of schools Jeju Island Islands, remote areas No. of students No. of Population for the Panel Survey
0 1~30
Middle school 2,903 42 166 44 402 2,249
Vocational high school 748 12 31 16 58 631
General high school courses 1,497 17 62 47 76 1,295
General high schools Tech and Vocational high schools running general high school courses (1,297)
(200)
(17)
(0)
(38)
(24)
(39)
(8)
(36)
(40)
(1,167)
(128)
Total 5,148 71 259 107 536 4,175

1) 53 of branch schools are included.

2) Two out of 14 schools that are changed from technical vocational high school category into the general one, are included.

3) Among 199 comprehensive high schools and 11 ones that adopts the comprehensive high school system, those who run general high school courses are included.

<Figure 2> Number of Schools and Students by Region
Region No. of schools No. of students
Middle schools Vocational high schools General high school courses Middle schools Vocational high schools General high school courses
Total General high schools General high school courses in vocational high schools Total General high schools General high school courses in vocational high schools
Seoul 357 78 204 204 0 120,826 25,456 94,686 94,686 0
Busan 159 46 82 82 0 47,991 16,731 29,755 29,755 0
Daegu 107 21 56 56 0 36,474 8,699 24,416 24,416 0
Incheon 87 28 49 49 0 35,793 10,133 21,709 21,709 0
Gwangju 69 13 45 45 0 20,506 5,056 13,601 13,601 0
Daejeon 73 13 39 38 1 20,227 4,527 14,601 14,386 215
Ulsan 42 11 25 25 0 16,452 3,828 10,287 10,287 0
Gyeongki 392 110 229 182 47 132,889 28,181 80,688 75,351 5,337
Gangwon 83 34 63 44 19 16,537 5,386 11,380 8,554 2,826
Chungbuk 91 30 43 43 0 18,409 7,056 11,638 11,638 0
Chungnam 155 36 70 69 1 22,791 7,120 16,279 16,175 104
Jeonbuk 123 52 83 59 24 23,446 7,587 16,096 13,545 2,551
Jeonnam 156 51 78 73 5 21,583 8,337 14,660 14,350 310
Gyeongbuk 169 59 112 102 10 30,529 8,458 22,958 21,632 1,326
Gyeongnam 186 49 117 96 21 39,964 8,645 28,677 26,710 1,967
Total 2,249 631 1,295 1,167 128 604,417 155,200 411,431 396,795 14,636

1) Students of general high school courses at vocational schools are excluded.

Sampling Methods

The first step in sample selection of the KEEP is the stratification of all schools in the country, (vocational schools are stratified by school type). The second step is stratified cluster sampling to select the schools for each region by the ratio of the number of the students, (ratio of the students by school type in case of vocational high schools), and to select classes and students from the selected schools.

School Sampling Process

  • Middle schools and general high schools (Refer to<Figure 3> and <Figure 4> )
    • Stratification: The country is divided into 15 regions (Seoul, six metropolitan cities and eight provinces.(Jeju island was excluded)
    • School Sampling: 100 schools were initially selected as samples for each level.
  • Vocational and technical high schools (Refer to <Figure 5> )
    • Stratification: Schools are stratified into three groups by school type, (technical high schools, commercial high schools, and others). The ratio of students of the six types of vocational high schools, including household high schools, fisheries high schools, business high schools, comprehensive high schools, and marine high schools, is low compared to the total number of students. Therefore, they are grouped as “others”.
    • School sampling: 100 schools were selected..
<Figure 3> Number of Sample Schools by Region (Middle Schools)
Region Enumeration district by region Ratio of the number of students by region (%) No. of sample schools by region
No. of schools No. of students
Seoul 357 120,826 19.99 20
Busan 159 47,991 7.94 8
Daegu 107 36,474 6.03 6
Incheon 87 35,793 5.92 6
Gwangju 69 20,506 3.39 3
Daejeon 73 20,227 3.35 3
Ulsan 42 16,452 2.72 3
Gyeongki 392 132,889 21.99 22
Gangwon 83 16,537 2.74 3
Chungbuk 91 18,409 3.05 3
Chungnam 155 22,791 3.77 4
Jeonbuk 123 23,446 3.88 4
Jeonnam 156 21,583 3.57 3
Gyeongbuk 169 30,529 5.05 5
Gyeongnam 186 39,964 6.61 7
Total 2,249 604,417 100 100
<Figure 4> Number of Sample Schools by Region (General High Schools)
Region Enumeration district by region Ratio of the number of students by region (%) No. of sample schools by region
No. of schools No. of students
Seoul 204 94,686 23.01 23
Busan 82 29,755 7.23 7
Daegu 56 24,416 5.93 6
Incheon 49 21,709 5.28 5
Gwangju 45 13,601 3.31 3
Daejeon 33 14,601 3.55 3
Ulsan 25 10,287 2.50 2
Gyeongki 229 80,688 19.61 20
Gangwon 63 11,380 2.77 3
Chungbuk 43 11,638 2.83 3
Chungnam 70 16,279 3.96 4
Jeonbuk 83 16,096 3.91 4
Jeonnam 78 14,660 3.56 4
Gyeongbuk 112 22,958 5.58 6
Gyeongnam 117 28,677 6.97 7
Total 1,295 411,431 100 100
<Figure 5> Number of Sample Schools by School Type (Technical and Vocational High Schools)
Classification Number of survey population by school type Ratio of the number of students by region (%) No. of sample schools by region
No. of schools No. of students
Technical 202 63,059 40.63 41
Commercial 194 59,279 38.20 38
Others 235 32,862 21.17 21
Total 631 155,200 100 100

Class sampling process

  • Four classes were selected from each selected school.

Student sampling process

  • Five students were selected from each selected class using a systematic sampling method.

New Sample Design(Additional Samples for the Fourth Year)

The needs of the new sample design

  • The existing samples of the KEEP for the first year is 2,000 third grade students at middle schools, 2,000 third grade students at general high schools, and 2,000 third grade students at vocational high schools. (Refer to <Figure 6> ).
<Figure 6> Existing Sample Design
School type Existing sample numbers
2004 third grade middle schools 2,000
2004 third grade, high schools General high schools, third grade 2,000
Vocational high schools, third 2,000
Total 6,000
  • The sample number of third grade students at middle schools in the first year was smaller than the sample number of third grade students at high schools of 4,000 students. The third grade middle school students entered general high schools and vocational high schools in the second year.
  • The addition of the samples of third grade students at middle schools is required given the changes of the characteristics of the samples (entering general high schools and vocational high schools).
    • The additional samples of third grade students at middle schools are essential in order to analyze the future course of the cohort by vocational school type (technical, commercial, and others), as 30% of the students entered vocational high schools.
    • There is also not enough information for special purpose high schools, (scientific and foreign language schools), in consideration of the influence of those schools.

Number of new samples

  • The number of samples by school type was determined by considering the changes of the characteristics of the samples, (entering general and vocational high schools), and the necessity for more information about special purpose high schools. (Refer to <Figure 7> ).
<Figure 7> Number of New Samples (Additional Samples for the Fourth Year)
School type New sample number
Third grade at vocational high schools 1,000
Third grade at special purpose high schools Foreign language high schools, third grade students 300
Science high schools, third grade students 300
Total 1,600

1) 50 schools x 20 students = 1,000 students

2) 10 schools x 30 students = 300 students

Population of the new samples

  • The population of the new survey is second grade students at vocational high schools, foreign language high schools, and science high schools as of 2006. The same samples will be used in 2007.
    • The reason why second grade students were used for the sampling is that 60% of all students enter college after graduating from second grade in science high schools, and second grade students are used for the samples to include those students accordingly.
    • The new samples apply to second grade students as of 2006 and the actual survey was conducted among third grade students in 2007.
  • The new population for the panel survey is 641 from vocational and foreign language/science high schools with more than 31 second grade students as of 2006, located in cities and town across the country. (Refer to <Figure 8> ).
    • Second grade students of 2006 graduated to third grade in 2007. Considering the level of student movements within and between schools, data was only collected from schools with over 31 students at second grade level.
<Figure 8> The Decision Process of the Survey Population (New Samples)
School type Number of schools Jeju Island Special regions No. of students Survey population
0 1~30
Vocational high schools 723 12 32 17 61 601
Foreign language high schools 29 1 0 4 0 24
Science high schools 19 1 1 1 0 16
Total 771 14 33 22 61 641

The number of vocational high school students does not include the number of students in general high school courses. (For example, comprehensive high schools with less than 30 students, without counting students taking general high school courses, were excluded.)

Sampling process for new samples

  • Vocational high schools (Refer to <Figure 9> )
    • Stratification: Schools are stratified into three groups by school type, (technical high schools, commercial high schools, and others). The ratio of students of the six types of vocational high schools, including household high schools, fisheries high schools, business high schools, comprehensive high schools, and marine high schools, is low compared to the total number of students. They are grouped as “others”.
    • School selection: 50 schools were selected.
  • Foreign language high schools (Refer to <Figure 10> ).
    • Stratification: Stratified into two types by school establishment type (public or private)
    • School selection: 10 schools were selected.
  • Science high schools
    • School selection: 10 schools were selected randomly.
<Figure 9> Sample schools by school type (vocational high schools)
School type Number of survey population by school type The ratio of student numbers by school type (%) Number of allocated sample schools by school type
Number of schools Number of students
Technical high schools 206 61,855 43.0 21
Commercial high schools 188 54,002 37.5 19
Others 207 28,150 19.5 10
Total 601 143,962 100 50
<Figure 10> Sample Schools by School Establishment type (Foreign Language High Schools)
School establishment type Number of survey population by school type The ratio of student numbers by school type (%) Number of allocated sample schools by school type
Number of schools Number of students
Public 7 1,272 18.37 2
Private 17 5,652 81.63 8
Total 24 6,924 100 10

Class sampling process (new samples)

  • Vocational high schools
    • Four classes were selected from each selected school
  • Foreign high schools and science high schools
    • Five classes were selected from each selected school

Student sampling process (new samples)

  • Vocational high schools
    • Five students were selected from each selected class using a systematic sampling method.
  • Foreign high schools and science high schools
    • Six students were selected from each selected class using a systematic sampling method.

Weights

  • Cross-sectional weights as well as longitudinal weights are required in the longitudinal surveys.
    • The general weight application method for cross-sectional surveys can be used for the weights of the first year.
    • The longitudinal weights must be considered in addition to the cross-sectional weights for the weights of the second year and later.
  • The weights are generally calculated over three steps: unequal selection probability compensation, non-response compensation, and post-stratification compensation.

Unequal selection probability compensation

  • Selection probability is calculated during the sample selection process. The selection probability compensation weight is calculated as the reciprocal of the selection probability.

Non-response compensation

  • The difference between the survey target sample number and the actual survey sample number is due to the presence of non-response samples.
  • The response ratio is calculated as the actual sample number to the survey target sample number. The non-response compensation weight is calculated as the reciprocal of the response ratio.

Post-stratification compensation

  • The total sum of the weight calculated by applying the selection probability, the response ratio, and the survey population number will be the same using this compensation method.
  • The post-stratification compensation weight is calculated as the ratio of the actual survey population number to the total sum of the weights.