Chapter 7 Descriptive Methods
7.1 Distinguish between descriptive methods
Descriptive methods are non-experiments (as opposed to quasi-experiments or experiments). They have no manipulation and no random assignment.
| Descriptive Method | Definition | Example |
|---|---|---|
| Case study | In-depth analysis of experiences of a single individual or group | Psychologists at Children’s Hospital Los Angeles treated and studied Genie, a feral child to understand critical periods in human language development (1970s) |
| Single variable study | Use of a sample to measure a single variable; ideally, the sample is randomly selected from the population | Public opinion polls |
| Correlational methods | Any non-experimental study (no manipulation) analyzing the relationship between two variables | A survey asking participants how many servings of fruits and vegetables they eat daily along with the time it takes them to run one mile |
| Archival research | Data is gathered from existing records to test a theory or hypothesis | Paternoster (1983) analyzed court records and found that prosecutors were much more likely to request capital punishment for Black defendants than White ones. |
| Longitudinal research | Repeated measurements from a sample over a period of time | Since 2004, the Alzheimer’s Disease Neuroimaging Initiative has been tracking the development of Alzheimer’s Disease in adults over age 55. |
| Qualitative research | Researchers conduct open-ended data collection (for example, interviews) and generalize theory based on specific example (called inductive reasoning). Focus is on “why” and “how.” Usually no hypothesis testing and few quantitative variables. | Case studies are an example of qualitative research. |
7.2 Qualitative versus quantitative research
Descriptive methods fall into two broad categories. Quantitative research describes anything that is analyzed with statistics like t-tests or correlations. In quantitative research, researchers first form a hypothesis and then use the research study to provide evidence for it. In qualitative research, researchers start with a broad research problem and then use the research study to explore it. The data collected may be interview notes, books, descriptions, videos, or recordings. After qualitative data are collected, then researchers start making sense of it, often by summarizing it and describing its themes.
Quantitative research Researchers test a specific hypothesis to see if the study supports it (called deductive reasoning). Focus is on control, reliability, and validity. Usually involves quantitative data of some kind and hypothesis testing. | Experiments, quasi-experiments, single-variable studies, correlational methods |
Qualitative research | Researchers conduct open-ended data collection (for example, interviews) and generalize theory based on specific example (called inductive reasoning). Focus is on “why” and “how.” Usually no hypothesis testing and few quantitative variables. | Case studies are an example of qualitative research.
7.3 Selection threats can affect descriptive methods
Internal validity is the big concern with non-experiments. Recall that all threats to internal validity are confounding variables, either because of the third variable problem or the directionality problem. In the next unit, we will discover why experiments make it easier to establish internal validity.
The most common problematic third variables can be remembered as GAGES (Pelham & Blanton, 2019): Geography, age, gender, ethnicity, socioeconomic status.
GAGES are examples of selection threats. “Selection” means that something about the participants differs across the conditions at the start of the study. If you do a study to investigate political attitudes among people living in a big city on the west coast of the US versus people living in a small town in the southwest US, your measurement of attitudes might be affected not only by geographic location but also by socioeconomic status (or other demographic variables). We could say that selection is a threat to internal validity in this case.
Selection is not the only threat to internal validity, but it’s one of the hardest threats to address in non-experimental research designs.
7.4 Which statistics can be used for descriptive methods?
By now, it should start to be clear that any statistics can be used on any research design. The right statistic depends on the types of variables that are measured in the study. Does the design have continuous or discrete variables? Are any variables dichotomous? The nature of the variables affects which statistics can be used, not the research design.