| Difference Between Humans and Molecules | How That Difference Makes Transferring Classical Scientific Methods to Psychology Difficult | Common Strategies for Dealing With that Problem |
| Individuals are influenced by many variables other than the treatment and can't be isolated from those variables by being put in a vacuum. | Internal validity (making solid cause-effect conclusions) is difficult because isolating humans from non-treatment variables to determine whether the treatment has an effect is difficult. | Use experiments: studies that use random assignment.
Do not make cause-effect conclusions based on studies that are not experiments. |
| Individuals have thoughts and feelings. | Construct validity (establishing that labels for variables are accurate) is difficult because thoughts, feelings, and other psychological constructs are invisible and subjective and thus cannot be objectively manipulated or seen. | Make inferences about constructs using objective recipes--operational definitions--and do research to test the construct validity of those definitions. |
| Individuals may behave differently if they know they are being studied. | Construct validity is difficult because what the researcher thinks is a reaction to a stimulus may be an act put on to impress or help the researcher. For example, participants may act in a way that supports what they think the researcher's hypothesis is or participants may act in a way that they think will impress the researcher. | Making responses anonymous may stop participants from acting in a way to impress the researcher, disguising the purpose of the study may stop participants from acting in a way that would support the hypothesis. |
| Individuals are all different and individuals behave differently at different times, places, and situations. | External validity is difficult because generalizing from the behavior of one set of individuals at one time, place, and situation is difficult. | Random sampling and replication |
| Individuals are extremely variable: They behave differently from each other and they behave inconsistently. | Knowing whether what you are seeing reflects a coincidence or reflect a consistent pattern is difficult. | Determine whether the finding is reliable using statistical significance tests--and designing the study so that its data meets the assumptions of those tests. |
| Individuals have rights and should be treated with respect. | Ethics: Many potentially interesting studies should not be done because they would be unethical. | Follow the APA code of ethics, especially in terms of getting approval from an ethical review board before doing the study, getting informed consent, allowing participants to withdraw from the study at any time, and debriefing participants. |