Page 335 - Contributed Paper Session (CPS) - Volume 4
P. 335

CPS2253 Gabrielle Palermo
               1.  Introduction
               Geres was designed to follow the imitial sampled pupils along five waves, from
               2005 to 2008, and the new pupils that joined the classes of original pupils,
               since  they  continued  studying  at  a  sampled  school.  The  surveys  additions
               could be from the same initial population or a different population rather than
               the one at wave 1. Additional pupils from the same population can have their
               selection probability computed if the sampling frame is available. However, it
               would be necessary to add all students studying with them in 2005 (1° grade),
               and track them over the waves. As a consequence, the πhk would change for
                                                                       (1)
               all selected schools, and the sample representation for the study population
               could fail.
                   On the other hand, some of the new students could be studying in a grade
               above, and repeat the year at some point, joining a surveyed classes. Likewise,
               an additional student could be studying in another place that is not part of
               the  area  covered  by  the  survey.  Thus,  pupils  that  are  from  a  different
               population at wave 1 do not have an inclusion probability.
                   Indirect sampling is crucial when one wants to study rare populations that
               are difficult to identify or populations of interested that are not listed; since a
               probability sample cannot be selected directly. The idea of indirect sampling
               happens when a sample from a given target population B can be reached
               through  links  with  another  sample  of  a  population  A,  then,  these  two
               populations  are  somehow  connected.  Indirect  sampling  is  a  form  of
               probability sampling because it is based on a sampling frame (Lavallée, 2007),
               and it started to be used in cross-sectional surveys to make inference about
               the population of interest.
                   The figure 1 shows the different ways that pupils could be included or
               followed by Geres. The pupils that are outside school were not selected then.
               And pupils could repeat or skip a grade more than once, but the figure only
               shows pupils that were attending up to one year less or more than the main
               grade, in other to make it simpler.
                   The main grade is the one initial pupils are expected to be studying at a
               given wave. The first and the second waves happened at the same academic
               year, and the further ones after one year. Thus, the main grade for each wave
               is:
                      Wave 1: 1st grade;
                      Wave 2: 1st grade;
                      Wave 3: 2nd grade;
                      Wave 4: 3th grade;
                      Wave 5: 4th grade.





                                                                  324 | I S I   W S C   2 0 1 9
   330   331   332   333   334   335   336   337   338   339   340