Class 17: Research Enterprise in Psychology
Last updated: October 2, 2021
Consider these questions:
- #1 Do lie detectors work, that is, can we figure out who is lying and who isn't with a polygraph ("lie detector")?
- #2 Does the full moon cause people to do strange and abnormal things?
- #3 Does listening to classical music (e.g., by Mozart) make infants smarter?
- #4 Do people with opposite personalities attract each other?
The answer to all of these questions is FALSE
How would you research any of these questions?
What would you need to do to answer these questions?
Goals of science are "description, prediction, and control":
1 Measurement & Description (good but not yet enough)
- What the phenomenon does, looks like, etc.
- How much of the phenomenon exists (how heavy, tall, fast, hard, etc.)
2. Understanding & Prediction = Lawfulness (what general predictive statement can we make on the basis of understanding a phenomenon)
- Hypotheses: A tentative statement about the relationship between two variables
- Note that the concept of "hypothesis" in science is not the same as the concept of "theory" (see below)
- In a hypothesis, we predict that there IS a relationship
- Variables: any measurable conditions, events, characteristics, or behaviors that are examined or controlled in a research study
- Some "predictions" and "laws" in psychology
- The personality trait of Conscientiousness is robustly associated with better academic performance (Poropat, 2009), that is, students who are highly organized and efficient and avoid putting off required tasks receive higher grades in school.
- Serial Position Effect: When exposed to a long series of stimuli, people tend to remember those stimuli which came either at the beginning or the end of the series.
- Thorndike's Law of Effect: Behaviors that result in favorable effects are more likely to be repeated while behaviors that result in unfavorable/painful/unsatisfactory effects are less likely to be repeated
- Keynes’ fundamental law of economic behavior: “humans are disposed, as a rule and on the average, to increase their consumption as their income increases but not by as much as the increase in the income” (we spend more as we earn more, but we don't spend all of the increase in our earnings, we save some of the increase.)
3. Application of Laws & Control (Doing something with the laws)
- Scientists hope ultimately to use the laws they discover to help human beings lead better lives
- Thus, there are always moral issues involved in science
Theory in science ≠ (is not the same as) "theory" in
Testability in Science
Steps in a Scientific Investigation
1. Formulate a Testable Hypothesis (that is, some reality we claim is true)
Are science faculty subconsciously biased against women?
Hypothesis: Science faculty members ARE subconsciously biased against women.
What could we hypothesize? (and operationally define?)
- Who is a science faculty member?
- What do we mean by bias?
An operational definition describes precisely how to measure or identify the variable under review. Here are examples of operational definitions:
- Depression is equal to the number of symptoms on a depression measurement scale
- Self esteem is equal to the number of positive statements about one's self that a person endorses
- Likability is equal to the average of ratings from 1 ("utterly hateful") to 7 ("completely likable") that a target group gives to someone on a survey scale
- Accuracy is equal to the number of correct responses to some sort of task
- Processing speed is equal to the number of seconds of time that it takes someone to complete a task.
2. Select the Research Method & Design the Study
Researchers have to think of how they are going to carry out the study & test their hypotheses. Some of the options to consider would be
Participants (subjects) = the persons or animals whose behavior is systematically observed in a study
- Experiment? Investigator manipulates some variable(s) under controlled condition(s) and observe changes as a result
- Case Study? Focus upon the experience of one or a few subjects/participants
- Naturalistic Observation? Take careful measurements without intruding yourself in a real-world setting
- Survey? Provide participants with a written set of questions or interview the participants yourself
3. Collect the Data
Get the information which our research study design requires
- Record how each subject completes the same task(s)
- Direct Observation
- Questionnaire (Survey)
- Psychological Test
- Physiological Recording
- Examining Archival Records
4. Analyze the Data & Draw Conclusions
- Statistics analyze numbers: how much of a difference has to be there for it to be a real difference?
- Qualitative approaches analyze non-numerical data (e.g., look for common themes, etc.)
5. Report the Findings
- Scientific journals, for example, in APA Journals
- Books, etc.
Advantages of a Scientific Approach to Research
- Clarity & precision (not vague or ambiguous)
- Intolerance of error
- Scientists are skeptical (even of their own ideas)
- If results of studies conflict, scientists want to know why
B. Looking for Causes: Experimental ResearchWhat is an experiment? A research approach in which the investigator controls the conditions under which research subjects or participants experience variables. In an experiment, the research subjects experience identical (or standard) conditions except for the variable under review.
Hypothesis: Science faculty members ARE subconsciously biased against women
Moss-Racusin et al. (2012)
A. Types of Variables
- Independent = Controlled by Investigator
- Dependent = Depends on what happens to the participant
- Independent => Identical application for a laboratory manager position except for name of applicant which is the independent variable (name = gender)
- Male name = John vs. Female name = Jennifer
- Dependent => Participants would read application and then rate the candidate on these dependent variables:
- Competence: How competent? (on a scale of 1 to 7)
- Hireability: How likely would you be to hire this applicant? (on a scale of 1 to 7)
- Mentoring: How willing would you be to mentor this applicant (on a scale of 1 to 5)
- Salary: What would you suggest be the starting salary (in a range between $15,000 and $50,000)
B. Types of Groups in an Experiment
- Experimental: those who are given the independent variable(s) to be tested
- Control: those who are not given the independent variable(s). They are treated in exactly the same way as the experimental group except for the independent variable.
In this science faculty member experiment, the control group might actually be considered those who received the male name, i.e., if our objective is to look whether females are treated differently, the male name group would be the standard against which to judge. Thus, the experimental group would be those who got the female name, i.e., we want to see if they will rate the candidate differently than the control group.
- On each of these dependent variables, there is a significant difference between ratings given male vs. female applicants. Females are always rated below males. The hypothesis is supported.
- Female & male faculty raters did not differ in their overall pattern of ratings. Thus female faculty were equally likely to rate female-named applicants lower than male-named applicants.
C. Problems in an Experiment
- Extraneous variables = two variables are competing to explain the outcome: one is acknowledged, the other ignored. An extraneous variable is one that might influence the dependent variable but which is not actually being tested as the independent variable. You ultimately can't be sure what the independent variable really was.
- e.g., testing a group of children for motivation under different lighting conditions and all of the children are hungry
- Motivation levels may be due to the hunger level of the children rather than the level of lighting
- e.g., measuring speed of processing information under different lighting conditions in an environment in which there is a lot of noise coming from the street
- Processing speed may reflect the noise rather than the lighting conditions
- Confounding variable = one variable is actually linked to another variable but that link is not recognized. A confounding variable is linked to another variable so that that we cannot sort apart their effects separately
- e.g., testing a group of children for motivation under different lighting conditions but some children are hungry & some are well fed
- Differences in motivation might be due to the differences between the children's hunger level rather than the differences in lighting conditions or may be due to both variables
- e.g., Using a test in English with a group of immigrants from a particular country to measure differences in verbal intelligence when some immigrants know a lot of English & some immigrants know very little English
- Differences in verbal intelligence may reflect real differences or they may simply reflect differences in the knowledge of English.
Choosing an Experimental Group
- Need for random assignment to experimental vs. control group
Random assignment assures that the two groups are essentially equal
D. Types of Design
- One (1) Independent Variable
- Two (2) Independent Variables
E. Advantages & Disadvantages of Experiments
Advantage: Powerful: Cause & Effect can be isolated
Disadvantage: Some or even many variables can't be manipulated for practical or ethical reasons
Moss-Racusin, C. A., Dovidio, J. F., Brescoll, V. L, Graham, M. J., & Handelsman, J. (2012). Science faculty's subtle gender biases favor male students. Proceedings of the National Academic of Science (PNAS). Published online before print September 17, 2012. doi: 10.1073/pnas.1211286109 [link]