Correlational research investigates relationships between two variables without manipulating either variable. It can be used for explanatory purposes to help explain behaviors, or for prediction purposes to predict outcomes. Common correlational techniques include scatter plots, regression analysis, multiple regression, and factor analysis. Threats to internal validity like subject characteristics, location, instrumentation, and mortality must be evaluated. The basic steps in correlational research are problem selection, sampling, instrumentation, design, data collection, and evaluating threats to validity.