The methods of sociological research are basically those outlined in the preceding chapter and used by all scientists. As Karl Pearson has remarked, "The unity of all science consists alone in its method, not in its material. The man who classifies facts of any kind whatever, who sees their mutual relation and describes their sequences, is applying the scientific method and is a man of science" (1900, p. 12].
While scientific methods are basically alike Anticlerical Life (1976], reports interviews with II national sample of 2,700 households, inquiring as to their satisfactions and- dissatisfaction. They found that married people are happier than single people, that prosperous people are happier than' poor people and made many other interesting observations. If the study extends over time, describing a trend or making a series of before-and-after observations, it is called a longitudinal, duty. Thus, Levine and Meyer [1 m]studied changes in black and white enrollment in Kansas-City public schools between 1960 and 1974. They found that schools with a relatively small black enrollment (under 29 percent) were likely to "remain desegregated, while schools with a higher percentage of black students became aghast totally desegregated as a result of what has come to be known as "white flight.

The national public opinion polls (Gallup, Ha ris, Fahd Others) are cross-sectional studies, but if the same set of .questions is repeated at intervals over a period of years, longitudinal comparisons can he drawn.Longitudinal studies may be either prospective or retrospective. A retrospective study (often called an ex post fact study) works backward in time, using data that are already recorded. For example, Wonder and Averts (1950] used hospital records of 605 lung cancer victims and found that all but eight were cigarette smokers. When a retrospective study shows strong evidence of a relationship between two facts, the next step often is to see whether a prospective study will confirm the relationship. A prospective study begins with the present and carries observations forward over a period of time. Thus Dorn (1959) and Kahn (1966) followed the health history of 200,000 veterans for eleven years, finding that the pack-a-day-or-more smokers were sixteen times as likely to die from lung cancer as were nonsmokers, Prospective studies take a long time to complete and are often very costly, milking them one of the least common types  of research study.

Sometimes longitudinal conclusions are drawn from cross-sectional studies. A cross sectional study may show differences between age groups, and this is often interpreted as evidence of changing attitudes or behavior. For example, numerous studies have shown that young people are more permissive than older people about sex behavior and drug use. Does this mean that values are changing and that the values of today's youth will be everybody's values tomorrow? Or is this a life-cycle change, With the young growing more conservative as they grow older? A cross-sectional study will not tell this. Longitudinal conclusions from cross-sectional studies "are often. dead wrong. For . example; ever since "intelligence" testing began, cross-sectional comparisons have consistently shown that average IQ -seems to . peak in early adulthood and declines steadily thereafter. But these surveys were conducted during a period of steadily rising levels of .public education. Each thus compared better eat ed young people with less well educated older people. More recent longitudinal studies reassuring IQ s of the same persons over a period of years report no consistent decline , in IQ until old age, with some aspects of "intelligence" improving and others declining with advancing years ( Baltes, 1968; Baltes and Schaie, 1974]. Longitudinal conclusions can only be established by longitudinal studies, even though cross-sectional studies may suggest promising hypotheses.