Statistics: Chapter 1
What are we doing, exactly? First, we’re doing science. The purpose of science is to describe, predict, and explain (and possibly control) natural phenomena. For Behavioral Science majors, this means people. (and sometimes rats). For Business majors, this means people, too. Science is about the interpretation of data based on a hypothesis. Hypothesis: Prediction of the outcome of a study.
Goals of science (increasing in complexity) Describe: most basic level of science. Observation of behavior and when it occurs. Predict: Identifying factors that indicate when events will occur. Requires that you’ve done some observation first. Explain: Identify the causes that determine when and why a behavior occurs.
How do we do these things? Descriptive Methods Observational: Nothing but watching and taking notes Naturalistic Laboratory Case Study: A small group of individuals, in detail Survey: Just asking questions about everyday behavior
Surveys: Tricky business Asking people about their habits and regular behaviors poses a few problems. When might people be less than honest about their responses? When might your sample not represent the population? Some folks might be more willing than others to respond.
How do we do these things? Predictive (Relational) methods Correlational: Assess the degree of relationship between two variables Variable: An event or behavior that has at least two values Quasi-experimental: Allows us to compare naturally-occurring groups of individuals. There are some characteristics that are not assigned by the experimenters.
More about correlations Correlations allow us to do some predicting. However, we cannot make the “jump” to causality. Correlations can be either positive or negative. http://www.emathzone.com/wp/wp-content/uploads/2014/10/positive-negative-corrrelation.jpg
More about quasi-experimental In quasi-experimental studies, we look at how a particular variable (behavior) changes across different existing or naturally-occurring groups. On-campus examples: sex/gender, athletic group, social club, state/city of origin, religious affiliation Some, but not all, of these examples are unethical for researchers to assign randomly or experimentally. Still not talking causality; alternative explanations may explain relationships between variables.
How do we do these things? Explanatory methods Experimental: Attempting to establish a cause-and-effect relationship through manipulation and control of variables.
More about experimental methods Certain requirements to be an “experiment” Researcher manipulates some aspect of the environment. (This is the Independent Variable) As many other factors as possible must be controlled. (The idea is that there is only one difference between two groups being measured). Researcher measures some behavior or outcome (Dependent Variable) Participants are randomly assigned to either “control” and “experimental” conditions
Some science big-picture issues One study/experiment isn’t enough to definitively explain relationships. Multiple studies should converge on a similar story. Getting evidence against a hypothesis isn’t a “fail” –could be explained by other factors. We never “prove” anything in science. The data we collect helps to support/ confirm/ give evidence for/substantiate a hypothesis or theory.