Equally important to the model and design of sampling, is the type of study
we intend to perform.  A panel study would be most appropriate for this study
because it would allow us to examine the same set of people each time we
visited the schools.  From our literature review, we learned that such a study
would allow us to analyze overall trends in a student's academic performance,
using different means of measurement, and it would show the precise patterns
of persistence and change in the student's home.  By studying the same students over time, we will be able to follow the specific problems that arise both at
home and in school, and how that combination affects the students' academic
performance and overall excellence.

    However, as Marsh recorded, with a panel study there is a potential danger, known as panel attrition (1990).  He explained how throughout the course of
a year, several students may relocate or transfer to other schools.  Some students who participate in the first study may not be able to participate in the study later on in the year.  Similarly, we are concerned that students may get bored withparticipating in the study and choose to exit the study.  The danger is that those who offer a motivational incentive throughout the study in order to keep students interested and willing to participate.Thus far in the plan, we do not believe that there is any possibility that our research will harm those students who will participate in the study.  Although, we can guarantee that no physical harm will be caused, some students may get emotional when asked questions about their family, the structure of the family, and what goes on at home.

    However, "[a]ny study attempting to imply causation on the basis of
correlation must be interpreted carefully.  The results of the present
investigation is based on a single cohort of students of similar age.
Researchers often find that children from different family configurations differ on a wide variety of academic, attitudinal, and behavioral outcomes.  There is
seldom any adequate basis for determining whether these differences are
preexisting or are caused by family configuration changes," ---unknown.