The Practice of Social Research

Chapter Four.  Research Design

THE TIME DIMENSION
    Cross-Sectional Studies
    Longitudinal Studies
    Approximating Longitudinal Studies
    Examples of Research Strategies

    Sometimes, data are collected at one point in time, in what are called cross-sectional studies.  These are often the simplest and easiest but make it difficult to understand processes occuring over time.

    Longitudinal studies collect data at more than one point in time, making it easier to study change.  Trend studies are a series of cross-sectional studies over time, asking the same question of different samples drawn from the same population.  A string of political polls, showing a shift from Smith to Jones would be an example.  This would not let us see that there were also people shifting from Jones to Smith (though not as many) or what kinds of people were shifting.

    Panel studies follow a specific set of subjects over time.  Staying with the above example, we would know who shifted from Smith to Jones and who shifted from Jones to Smith as well as those who stayed pat.  We would be able to discover the kinds of people in each category helping us understand the changes taking place.

    A cohort study follows a subset of the population over time, using different samples each time rather than a fixed panel.  Studies of Vietnam veterans  or the Baby Boom generation would be examples of cohort studies.

    Often, cross-sectional data are used to make inferences about longitudinal processes.  At one point in time, we can learn that older people are more religious than younger people--and we may infer that people get more religious as they grow older.  Or we may compare college Freshmen and Seniors at one point in time and make inferences about the effects of the college experience.