Research design involves a set of decisions regarding what topic is to be studied, among what population, with what research method, and for what purpose. Chapter 4 of Babbie provides a general introduction to research design. It examines the following issues:

• Purposes of research inquiry
• Units of analysis
• Points of focus in social scientific research
• The time dimension in research
• The design of a research project
• The research proposal

1. Exploration
Much of social research is conducted to explore a topic or to provide a beginning familiarity with that topic. This type of research is typical when a researcher is examining a new interest or when the subject of study is relatively new and unstudied. Exploratory studies are more typically done to:

• To satisfy the researcher's curiosity and desire for better understanding about a particular topic.
• To determine the feasibility of undertaking a more careful study
• To develop research techniques and a sense of direction for future research.

2. Description
A major purpose of many social scientific studies is to describe situations and events. The researcher observes and then describes what was observed. The chief goals of descriptive research are to provide an accurate profile of a group and to describe a process, mechanism, or relationship. Because of its methodology, scientific descriptions are more accurate and precise than casual descriptions. Examples of descriptive studies are:

• U.S. Census - describes the characteristics of the population of the United States.
• A market researcher undertaking to describe people who use a certain product.

• Why did Americans vote for Clinton in the last presidential elections?
• Why do some cities have higher crime rates than others?
• Why do some children become delinquent?

Learning Check!
Discuss the three purposes for doing research. Describe a study that illustrates each purpose.

UNIT OF ANALYSIS
Units of analysis are those things we examine in order to create summary descriptions of all such units and to explain differences among them. Units of analysis deals with "what" or "whom" the social researcher aims to observe, describe and explain.

• Individuals: Individual people are the most typical unit of analysis in social scientific research. The individual is the unit of analysis in a survey in which 100 students are asked to rate their favorite courses at Northeastern. The individual is the unit of analysis because each individual student's response is recorded. Note that even though such data is gathered at the individual level, it is not reported at that level - we combine descriptions of individuals to provide a composite picture of the group the individuals represent. Statements that we might make from such observations are:

• Groups: Social groups might also be the unit of analysis for social scientific research. When social groups are the units of analysis, their characteristics may be derived from characteristics of the individuals in the group. A study of all gangs in a city would provide data on the racial and ethnic composition of the gangs studied. Other units of analysis at the group level could be married couples, families and friendship groups.

• Organizations: Examples would be corporations, churches, colleges, army divisions, church congregations and academic departments.

• Social Artifacts Another unit of analysis is the social artifact, or any product of social beings or of their behavior. One class of social artifacts includes concrete objects such as books, poems, paintings, songs, poetry, and jokes. Social interactions form another class of social artifacts suitable for scientific research. Examples are weddings, court cases, traffic accidents and race riots.

POTENTIAL ERRORS INVOLVING MISUSE OF UNIT OF ANALYSIS
A. Ecological Fallacy: Ecological units refer to groups or sets or systems: something larger than individuals. Ecological fallacy occurs when we assume that something learned about an ecological unit says something about the individuals making up that unit. This fallacy occurs when the observations are gathered at a higher level or aggregated unit of analysis, but the causal explanations being made is about a lower or disaggregated unit. It is always possible that patterns observed at the aggregate level do not reflect behavior at the individual level.

Example: A researcher finds that crime rates are higher in cities having large African American populations than in those with few African Americans. She then concludes that African Americans are more likely to commit crimes than other Americans. Note that we do not know who commits crimes in these cities.

B. Reductionism: Refers to an overly strict limitation on the kinds of concepts and variables to be considered as causes in explaining a broad range of human behavior. Reductionism of any type suggests that particular units of analysis or variables are more relevant than others.

• sociological reductionism - sociological variables (such as norms, values, roles) are considered
• economic reductionism - economic variables (such as supply and demand, marginal value) are considered.
• psychological reductionism - psychological variables (such as personality traits, childhood traumas).

• Characteristics: refer to states of being. For example, we might characterize individuals by gender, religious affiliation, marital status, country of origin, etc.
• Orientations: these reflect attitudes and beliefs.
• Actions: these are actual behaviors. Examples would include voting, buying, investing, and studying for an exam.

TIME DIMENSION IN RESEARCH

• cross-sectional Studies: Cross-sectional research is the simplest and least costly type of research design because observations are made at one point in time. Cross-sectional studies can be exploratory, descriptive, or explanatory. The major limitation of cross-sectional research studies is that their design makes it difficult to explain phenomena that occur over time. For example, do the descriptions of a study conducted today tell us anything about what happened ten years ago or what will happen a couple of years from now?

• Longitudinal Studies permit observations over an extended period of time. There are several types of longitudinal research designs but all involve observing something at more than one time. Longitudinal studies are more costly and more complex than cross-sectional studies, but they are also more powerful, especially when a researcher wants to answer questions about social processes or change. Three special types of longitudinal studies are discussed in Babbie: trend, cohort, and panel studies.

• Trend Studies: Trend studies are those that examine changes within some general population over time. They can be thought of as repeated cross-sectional designs. A good example would be a comparison of U.S. Censuses over time.

• Cohort Studies: A cohort is a specific group of people who go through a demographic or social event at the same time: marriage, birth, divorce, students who took Soc 337 in fall of 1998. A cohort study examines more specific subgroups (cohorts) as they change over time.

• Panel studies: A panel study is a longitudinal study in which exactly the same individuals (or sample) are observed at different time periods. Panel research is formidable to conduct and very costly. Panel studies face a special problem of panel attrition or loss of respondents over the course of the study. Tracking people over time is difficult; some respondents are lost through death, lack of interest or change of residence.

• Making logical inferences (see the University of Hawaii drug use study discussed in Babbie, page 104).
• Examining the time order of variables.
• Asking retrospective questions (Asking people to recall their pasts).
• Examining age differences

The table below contains census data collected in 1970, 1980, and 1990. Column variables represent the year the data was collected. The rows represent the year that the respondents were born. There are three cohorts in this table: people born in 1930, 1940, and 1950. The cells are labeled A through I. The number within each cell represents the respondents' age at the time of measurement.

Use the table below to identify which cells represent cross-sectional, panel, cohort and trend studies.

HOW TO DESIGN A RESEARCH PROJECT & THE RESEARCH PROPOSAL: Read these sections of the chapter 4 very carefully. They will help you design your own research project and provide an overview of the topics to be covered for the rest of the semester. Try to understand the various components of a research project.

• Research interest, idea or theory
• Literature review
• Conceptualization of concepts
• Choice of a data collection method
• Operationalization
• Population and sampling
• Observations
• Data processing
• Analysis
• Application (Reporting results and assessing their implications)