PSY 1001 Chapter 2: Research Methods - Part 3 Notes

Experimental Research

Allows researchers to identify causal relationships between variables.
Involves manipulating one variable (Independent Variable, IV) and measuring another (Dependent Variable, DV).
Correlational research utilizes two measured variables.
General question: Do different levels of the IV lead to different behaviors in the DV?

Research question: Does listening to music while studying affect learning ability?
IV: Students studying with music (Experimental group) vs. in silence (Control group).
DV: Score on a test of the material.

Start with two identical "things".
Treat them differently in only one way (the IV).
If the things differ in terms of the DV, then differences in the IV caused differences in the DV.

Fill two beakers with the same chemical substance.
Heat one beaker, refrigerate the other.
If the contents of the beakers are different afterward (e.g., different colors), then the temperature affected the substance.

It's difficult to start with identical groups of people.
Instead, create groups that are statistically equivalent (on average) through random assignment.
Random assignment: Randomly assign people to one group or another to create equivalent groups.
Important distinction: Random assignment differs from random sampling, where a sample is chosen randomly from a population.

One goal of experimental research is to identify causes of behavior.
Three conditions must be met to say that changes in the IV CAUSE changes in the DV:
1. Correlation between the IV and DV: When the value of the IV changes, the value of the DV changes.
2. Temporal precedence: Changes in the IV must precede changes in the DV.
3. Rule out plausible alternative explanations: Ensure nothing else could have caused changes in the value of the DV.

Research question: Do students learn Intro Psych material better when the lectures include fancy colorful slides?
IV: Fancy colorful slides vs. boring slides.
DV: Score on a test of the material.
Random assignment is used to create equivalent groups.
- Flip a coin for each student: Heads = attend lecture with fancy slides, Tails = attend lecture with boring slides.
Hypothetical Results:
- Students in the fancy-slides lecture score higher.
- Possible conclusion: Students learn better with fancy slides.
- Differences in the IV (fancy colorful slides or boring slides) caused a difference in test scores.

Any other factor that systematically differs between the groups might have caused the difference in test scores.
Examples: different lecture topics, different instructors, different rooms, different times of day.
Experimental control: Hold those factors constant for the two groups to eliminate alternative explanations.

Confound: Any factor that covaries perfectly with the IV.
Confounds provide alternative explanations for differences in the DV.
Use experimental control to eliminate the confounds (same instructor, same topic, etc.).

Reliability: When we make a specific measurement, do we get the same answer every time?
Validity: Are we really measuring what we think we’re measuring? (Also called "construct validity".)

Measuring the length of a table with a tape measure.
If you get the same length every time, the measurement is reliable.
However, if the tape measure is missing the first two inches, the measurement is not valid.
A measurement can be reliable even though it might not be valid.

Giving people a survey intended to measure extraversion.
If you get the same extraversion score every time, the survey is reliable.
However, if the survey is made up of questions asking about a person’s happiness, the survey is not a valid measure of extraversion.
To have validity, the survey should ask questions about sociability, talkativeness, assertiveness, etc.

Why do we need statistics?
Different people behave differently in identical situations.
There is variability in the world.
Even if students were taught identically, different students would get different test scores.
And different groups probably will have different average test scores.

If the average scores of the groups are different, it might be because:
- (1) Students learn better with fancy slides.
- (2) Just by chance, the “fancy slides” group scored higher.
Statistics help us decide which is correct.

Statistical tests determine if the difference between groups is (1) just a chance difference due to variability (individual differences) OR (2) a real difference representing an effect of the IV.
Most statistical tests produce a p-value.
p-value = the probability that you would get the observed difference just by chance, if the IV really had no effect.
If the p-value is very small (usually less than 0.05), conclude that the difference is a real effect of the IV. = the result is “statistically significant”.

Beneficence: Maximize the benefits to society while minimizing harm to research participants.
Autonomy: Allow research participants to give consent to participate in research. Don’t force or coerce people to participate.
Justice: Don’t conduct research on a small segment of the population (one type of person). Be sure that the people who are bearing the burden of participating are representative of those who can benefit from the research.
Researchers also have ethical obligations:
Role of the IRB (Institutional Review Board) in overseeing research.

Psychologists are scientists who gather and examine evidence to gain knowledge about behavior.
Psychologists apply treatments that are supported by evidence of their effectiveness.