Every research method does one of three jobs. Descriptive methods (case studies, naturalistic observation, surveys) describe what is happening. Correlational methods measure whether two variables move together. Only the experiment can establish that one thing causes another. Keep the hierarchy — describe, relate, cause — because a study that can only describe or relate can never prove a cause.
In an experiment, the researcher deliberately manipulates one variable to observe its effect on another while controlling everything else. - The independent variable (IV) is the manipulated suspected cause. - The dependent variable (DV) is the measured outcome — the data collected at the end. (Mnemonic: the DV is the Data.) - The experimental group gets the treatment; the control group does not and serves as the baseline. - A confounding variable is any variable other than the IV that could explain the results. - Random assignment places each participant into a group purely by chance, spreading pre-existing differences evenly so they cancel out. This is what makes an experiment an experiment. - Operationalization states a variable in specific, measurable terms ("aggression" → "number of times the child hits the inflatable doll in five minutes"), which makes a study replicable.
To keep expectations from biasing results, the control group may get a placebo (an inert treatment). In a single-blind study the participants don't know their condition; in a double-blind study neither the participants nor the researchers interacting with them know — the gold standard for drug and therapy trials.
A correlational study measures the relationship between two variables without manipulating anything. The result is a correlation coefficient (r), ranging from −1.00 to +1.00. - The sign shows direction: a positive correlation means variables move together (more studying, higher grades); a negative correlation means they move oppositely (more stress, less sleep). - The absolute value shows strength: r = −.85 is stronger than r = +.30. Sign ≠ strength.
Correlation does not prove causation. When two variables correlate, three explanations survive: A causes B; B causes A (the directionality problem); or a hidden third variable causes both. Only an experiment can rule out the last two.
A survey gathers self-report from many people; it is vulnerable to wording effects and sampling bias. A case study examines one person or small group in depth (great for rare cases, but not generalizable). Naturalistic observation watches behavior in its natural setting without intervening. Random sampling — where every member of the population has an equal chance of selection — is what lets you generalize from a sample to a population. (Note: random sampling = who is in the study; random assignment = which group they land in.)
[GRAPH: three distributions side by side — a symmetrical bell (normal, mean=median=mode), a right-skewed curve with a long tail to the high end (mean > median), and a left-skewed curve with a long tail to the low end (mean < median).]
Statistical significance (often reported as p < .05) means a result is unlikely to have occurred by chance alone — it does not mean the effect is large or important. Reliability is consistency (the same measure gives the same result repeatedly). Validity is accuracy (the measure captures what it claims to). A measure can be reliable without being valid — a scale that always reads five pounds high is perfectly consistent but wrong.
Human research requires informed consent (participants are told the risks and procedures and agree beforehand), protection from harm, confidentiality, and debriefing (explaining any deception and the study's true purpose afterward). Studies are reviewed in advance by an Institutional Review Board (IRB) and must follow APA ethical guidelines; animal research is likewise regulated to minimize suffering.
1. (B) Dependent variable. The DV is the measured outcome that depends on the treatment. (A) the IV is what the researcher manipulates; (C) a confound is an unwanted extra variable, not the planned measure; (D) a control condition is a group, not a variable; (E) an operational definition is how a variable is measured, not the outcome itself. Fix: DV = the Data you collect at the end; IV = what I change.
2. (A) Dependent variable. The quiz score is the measured outcome — the DV. (B) the IV is the manipulated study condition (silence vs. music); (C) no unwanted extra variable is described; (D) no fake treatment is used; (E) the sample is the students, not the score. Fix: In a scenario, the DV is whatever gets measured or scored after the manipulation.
3. (B). A negative sign means the variables move in opposite directions (more sleep → fewer errors), and .72 is close to 1, so the relationship is strong. (A) reverses the direction; (C) misreads the sign as weakness — sign is direction, not strength; (D) overclaims cause from a correlation; (E) contradicts a strong coefficient. Fix: Judge direction by the sign, strength by distance from zero — two separate reads.
4. (E). A third variable (hot weather) plausibly raises both ice cream sales and swimming/drownings, with neither causing the other. (A) and (B) leap to causation a correlation cannot support; (C) "illusory" means a perceived relationship that isn't real — here the relationship is real, just non-causal; (D) invents an implausible mechanism. Fix: Before accepting any causal story for a correlation, hunt for the variable that drives both.
5. (D) Positively skewed. High outliers create a long tail on the high (right) end and pull the mean above the median — a positive (right) skew. (A) and (E) require symmetry (mean = median); (B) negative skew has a low-end tail (mean below median); (C) bimodal means two peaks, not described. Fix: Skew is named for the tail; a few high scores → mean > median → positive skew.
6. (A) Standard deviation. Standard deviation measures how spread out scores are around the mean, so the more dispersed class has the larger SD. (B), (C), and (D) are measures of central tendency, not spread — and the means are equal here anyway; (E) a correlation coefficient describes a relationship between two variables, not one class's spread. Fix: Spread around the mean = standard deviation; center = mean/median/mode.
7. (D) Naturalistic observation. Watching and recording real behavior in its natural setting without intervening is the definition. (A) an experiment manipulates a variable; (B) a survey collects self-report; (C) a case study probes one individual in depth; (E) no correlation coefficient or manipulation is involved. Fix: No intervention + natural setting = naturalistic observation.
8. (C) Informed consent. Telling participants the risks and procedures and getting their voluntary agreement beforehand is informed consent. (A) debriefing happens after the study; (B) random assignment sorts participants into groups; (D) operationalization defines variables; (E) deception is withholding or misstating the purpose. Fix: Consent comes before and requires knowing the risks; debriefing comes after.
9. (E) Reliable but not valid. Reading the same amount every time makes the scale consistent (reliable), but being five pounds off makes it inaccurate (not valid). (A) reverses the terms; (B) denies the clear consistency; (C) claims accuracy it lacks; (D) contradicts the perfectly consistent reading. Fix: Reliable = consistent; valid = accurate. A measure can be reliably wrong.
10. (C) Random sampling. Giving every member of the population an equal chance of selection is random sampling, which supports generalization. (A) random assignment sorts already-chosen participants into groups; (B) a convenience sample uses whoever is easy to reach; (D) a case study examines one individual; (E) blinding concerns who knows the condition. Fix: Sampling = who gets into the Study; Assignment = which Arm they land in.
11. (D). Statistical significance means the result is unlikely to be due to chance alone (p < .05). (A) significance says nothing about effect size or importance; (B) significance alone doesn't prove causation without proper design; (C) it doesn't guarantee a representative sample; (E) it doesn't guarantee generalizability. Fix: "Statistically significant" = "probably not just chance," not "big" or "important."
12. (E) Debriefing. Explaining any deception and the study's true purpose after participation is debriefing. (A) informed consent occurs before the study; (B) confidentiality protects participants' identities; (C) random assignment is a design procedure; (D) a placebo is an inert treatment. Fix: Any deception must be followed by debriefing — the after-the-fact explanation.
1. (B) Dependent variable. The DV is the measured outcome that depends on the treatment. (A) the IV is what the researcher manipulates; (C) a confound is an unwanted extra variable, not the planned measure; (D) a control condition is a group, not a variable; (E) an operational definition is how a variable is measured, not the outcome itself. Fix: DV = the Data you collect at the end; IV = what I change.
2. (A) Dependent variable. The quiz score is the measured outcome — the DV. (B) the IV is the manipulated study condition (silence vs. music); (C) no unwanted extra variable is described; (D) no fake treatment is used; (E) the sample is the students, not the score. Fix: In a scenario, the DV is whatever gets measured or scored after the manipulation.
3. (B). A negative sign means the variables move in opposite directions (more sleep → fewer errors), and .72 is close to 1, so the relationship is strong. (A) reverses the direction; (C) misreads the sign as weakness — sign is direction, not strength; (D) overclaims cause from a correlation; (E) contradicts a strong coefficient. Fix: Judge direction by the sign, strength by distance from zero — two separate reads.
4. (E). A third variable (hot weather) plausibly raises both ice cream sales and swimming/drownings, with neither causing the other. (A) and (B) leap to causation a correlation cannot support; (C) "illusory" means a perceived relationship that isn't real — here the relationship is real, just non-causal; (D) invents an implausible mechanism. Fix: Before accepting any causal story for a correlation, hunt for the variable that drives both.
5. (D) Positively skewed. High outliers create a long tail on the high (right) end and pull the mean above the median — a positive (right) skew. (A) and (E) require symmetry (mean = median); (B) negative skew has a low-end tail (mean below median); (C) bimodal means two peaks, not described. Fix: Skew is named for the tail; a few high scores → mean > median → positive skew.
6. (A) Standard deviation. Standard deviation measures how spread out scores are around the mean, so the more dispersed class has the larger SD. (B), (C), and (D) are measures of central tendency, not spread — and the means are equal here anyway; (E) a correlation coefficient describes a relationship between two variables, not one class's spread. Fix: Spread around the mean = standard deviation; center = mean/median/mode.
7. (D) Naturalistic observation. Watching and recording real behavior in its natural setting without intervening is the definition. (A) an experiment manipulates a variable; (B) a survey collects self-report; (C) a case study probes one individual in depth; (E) no correlation coefficient or manipulation is involved. Fix: No intervention + natural setting = naturalistic observation.
8. (C) Informed consent. Telling participants the risks and procedures and getting their voluntary agreement beforehand is informed consent. (A) debriefing happens after the study; (B) random assignment sorts participants into groups; (D) operationalization defines variables; (E) deception is withholding or misstating the purpose. Fix: Consent comes before and requires knowing the risks; debriefing comes after.
9. (E) Reliable but not valid. Reading the same amount every time makes the scale consistent (reliable), but being five pounds off makes it inaccurate (not valid). (A) reverses the terms; (B) denies the clear consistency; (C) claims accuracy it lacks; (D) contradicts the perfectly consistent reading. Fix: Reliable = consistent; valid = accurate. A measure can be reliably wrong.
10. (C) Random sampling. Giving every member of the population an equal chance of selection is random sampling, which supports generalization. (A) random assignment sorts already-chosen participants into groups; (B) a convenience sample uses whoever is easy to reach; (D) a case study examines one individual; (E) blinding concerns who knows the condition. Fix: Sampling = who gets into the Study; Assignment = which Arm they land in.
11. (D). Statistical significance means the result is unlikely to be due to chance alone (p < .05). (A) significance says nothing about effect size or importance; (B) significance alone doesn't prove causation without proper design; (C) it doesn't guarantee a representative sample; (E) it doesn't guarantee generalizability. Fix: "Statistically significant" = "probably not just chance," not "big" or "important."
12. (E) Debriefing. Explaining any deception and the study's true purpose after participation is debriefing. (A) informed consent occurs before the study; (B) confidentiality protects participants' identities; (C) random assignment is a design procedure; (D) a placebo is an inert treatment. Fix: Any deception must be followed by debriefing — the after-the-fact explanation.