Skills

Research skills you need: the one-page sheet

Research Methods and Variables

Psychologists test hypotheses with operationally defined variables. Only an experiment, with a manipulated independent variable and random assignment, supports a cause-and-effect conclusion; correlational studies, case studies, naturalistic observation and surveys describe and predict. Control groups, placebos and blinding control confounding variables and expectations, and peer review, replication and meta-analysis decide which findings hold up.

Name the method exactly. An experiment needs a manipulated independent variable and random assignment. Otherwise it is a correlational study, a case study, naturalistic observation or a survey.
An operational definition says how this study measured a variable: "the number of words written down in two minutes", not "memory".
Controls: a control group (often given a placebo), and single-blind or double-blind procedures against expectations and experimenter bias.
Surveys are open to wording effects and social desirability bias. Science corrects itself through peer review, replication and meta-analysis.

the groups start out alike, on average, in everything else a difference in the dependent variable can be traced to the treatment: cause and effect a third variable or the reverse direction could explain the result, so no cause can be claimed expectations of participants and researchers cannot push the results

scientific method: A way of answering questions with evidence: ask a question, form a testable hypothesis, collect data systematically, and accept or revise ideas based on what the data show. Evidence gathered by observation or measurement is called empirical evidence.
psychological perspective: A broad way of explaining behavior and mental processes: biological (the body and brain), behavioral (learning from the environment and consequences), cognitive (thinking and memory), psychodynamic (inner conflicts and motives outside awareness), humanistic (growth and free choice), sociocultural (culture and other people) and evolutionary (traits that helped survival and reproduction).
theory (science): In science, a broad explanation that organizes many observations and predicts new ones. A theory is tested and revised, not a guess.
hypothesis: A specific, testable prediction about how variables are related, stated so that a study’s results could support or refute it.
experiment: A research method in which the researcher manipulates an independent variable and randomly assigns participants to conditions, then measures a dependent variable. It is the only method that supports cause-and-effect conclusions.
independent variable: In an experiment, the variable the researcher manipulates (for example, which drink a participant gets).
dependent variable: In an experiment, the variable that is measured to see whether the independent variable changed it (for example, reaction time).
operational definition: A statement of exactly how a variable is measured or manipulated in a particular study, such as "the number of words recalled in two minutes".
experimental group: The group in an experiment that receives the treatment, the level of the independent variable being tested.
control group: The group in an experiment that does not receive the treatment (or receives a placebo), giving a baseline for comparison.
random assignment: Using chance, such as a coin flip, to decide which condition each participant is in, so the groups start out alike on average.
confounding variable: A variable other than the independent variable that differs between conditions and could explain the results.
placebo: An inactive substance or fake treatment that looks like the real one. The placebo effect is a change caused by expecting a treatment to work.
single-blind study: A procedure in which participants do not know which condition they are in.
double-blind study: A procedure in which neither the participants nor the researchers who collect the data know who is in which condition.
experimenter bias: When a researcher’s expectations influence how a study is run or how results are recorded; double-blind procedures guard against it.
correlation: A relationship between two variables, so that knowing one helps predict the other. A correlation alone does not show that one causes the other.
correlational study: A research method that measures two or more variables as they naturally occur, to see whether and how strongly they are related. It can predict but cannot show cause and effect.
third variable: An outside variable that affects both of two correlated variables and may explain why they are related.
case study: An in-depth study of one person or a small group, often someone with a rare condition. It can suggest hypotheses, but the case may not be typical.
naturalistic observation: Watching and recording behavior in its natural setting without trying to change it.
survey: A research method that asks many people to report their own attitudes, beliefs or behaviors, through questionnaires or interviews.
social desirability bias: The tendency to answer questions in a way that makes oneself look good (social desirability bias), the main form of self-report bias, which distorts what people say about themselves.
wording effect: A change in people’s answers caused by how a question is phrased.
Likert scale: A rating scale that asks how strongly a person agrees or disagrees with a statement, for example from 1 (strongly disagree) to 5 (strongly agree).
quantitative data: Data recorded as numbers, such as scores, counts or times.
qualitative data: Data recorded as words or descriptions, such as interview answers, later sorted into themes.
meta-analysis: A statistical method that combines the results of many studies on the same question to estimate the overall effect.
replication: Repeating a study, often with new participants, to see whether its result holds. Findings that fail to replicate are treated as doubtful; the replication crisis is the discovery that many published psychology findings did not.
peer review: The process in which other experts evaluate a study’s methods and conclusions before a journal publishes it.

Sampling and Generalizability

A sample stands in for a population. Random sampling makes a representative sample likely, while convenience and self-selected samples are biased however large they are. Random sampling supports generalizing to the population; random assignment supports a causal claim. Whether results generalize to a group depends on how much the participants resemble that group in age, culture and setting.

The population is everyone a study wants to describe; the sample is the people actually studied.
Random sampling gives everyone the same chance of being chosen, so a representative sample is likely. A convenience sample or volunteers create sampling bias.
Random sampling decides who is studied and supports generalizability. Random assignment decides which condition each person gets and supports cause and effect.
Judge generalizability from participant characteristics (age, culture, setting), not from the sample size alone.

the sample is likely to be representative: it mirrors the population the sample is biased, and a larger biased sample is still biased results may not generalize to that group; argue from those characteristics

population (research): In research, the whole group a study wants to draw conclusions about, such as all U.S. high-school seniors.
sample (research): The part of a population that is actually studied.
random sampling: Choosing participants so that every member of the population has the same chance of being selected.
representative sample: A sample whose makeup (age, gender, background and so on) matches the population it stands for.
sampling bias: A flaw in how a sample is chosen that makes it unrepresentative, as with a convenience sample of whoever is easy to reach or a self-selected sample of volunteers.
generalizability: How far a study’s results can be applied to people beyond its participants; judged by how similar the participants are to the group in question.

Correlation and Scatterplots

A correlation describes how two variables are related. On a scatterplot, points rising to the right show a positive correlation and points falling to the right a negative one; the correlation coefficient r gives the direction by its sign and the strength by its size, from 0 to 1. A correlation allows prediction but never shows cause: a third variable or the reverse direction may explain it, and some correlations people believe in are illusory.

The correlation coefficient $r$ runs from $-1$ to $+1$ . The sign is the direction; the size (ignoring the sign) is the strength. $r = -0.8$ is stronger than $r = +0.5$ .
A positive correlation: both variables rise and fall together. A negative correlation: as one rises, the other tends to fall.
A scatterplot shows each person as a point; the line of best fit shows the trend and allows prediction.
Correlation is not causation: a third variable or the directionality problem can explain it. An illusory correlation is a relationship people believe in that the data do not show.

a scatterplot shows whether they rise together, move in opposite directions or show no pattern its sign gives the direction and its size, from 0 to 1, gives the strength a third variable or the reverse direction may explain the relationship, so no cause can be claimed

correlation coefficient: A number, r, from -1 to +1 that describes the direction (its sign) and strength (its size) of the relationship between two variables.
positive correlation: A relationship in which two variables tend to rise and fall together; r is above 0.
negative correlation: A relationship in which one variable tends to fall as the other rises; r is below 0.
scatterplot: A graph of paired data in which each point is one person or case, placed by its values on two variables.
line of best fit: The straight line that best fits the points on a scatterplot, showing the trend and allowing predictions. Also called a trend line.
directionality problem: In a correlation, not knowing which variable affects the other: A may lead to B, or B to A. Also called reverse causation.
illusory correlation: A relationship that people believe exists between two things when the data show none, often because memorable cases stand out.

Ethics in Research

Psychologists follow ethical guidelines that protect participants. A review board approves the plan; people give informed consent (minors give assent with a parent’s consent), may withdraw at any time, have their data kept confidential and are protected from harm. Deception is allowed only when necessary and low in risk, and it is always followed by debriefing. Animal research is reviewed for humane care. On the exam, name a safeguard the study describes and say how the researchers applied it.

An institutional review board (IRB) approves research with people before it starts; animal studies need an animal care committee.
Informed consent before taking part; informed assent from minors, with a parent’s consent; the right to withdraw at any time without penalty.
Confidentiality keeps identities and data private. Protection from harm keeps physical and emotional risk minimal.
Deception is allowed only when needed and low in risk, followed by debriefing. On the exam, name a safeguard the summary describes and say how it was applied.

an institutional review board (or an animal care committee) checks that risks are minimal and justified their informed consent (or a parent’s consent plus a minor’s assent) is real, and they may withdraw at any time it is allowed only with small risk and must end with debriefing that tells the truth

ethical guidelines: Rules that protect the rights and welfare of research participants, such as the American Psychological Association’s ethics code.
institutional review board: A committee at a university or other institution that reviews research plans involving people and approves them only if risks are kept as low as possible and justified by what the study can learn.
informed consent: A participant’s agreement to take part in a study after being told what it involves, including its risks and the right to withdraw.
informed assent: A minor’s own agreement to take part in a study after it is explained in terms they understand, given along with a parent’s or guardian’s consent.
right to withdraw: A participant’s right to stop taking part in a study at any time without penalty.
confidentiality: Keeping participants’ identities and data private, for example by using code numbers instead of names. Anonymity goes further: no names are collected at all.
protection from harm: The guideline that researchers keep physical and emotional risks to participants minimal, no greater than in everyday life when possible.
deception (research): In research, misleading participants about a study’s true purpose or procedures; allowed only when the study needs it, the risk is small and participants are debriefed.
debriefing: Explaining a study’s true purpose to participants after they take part, including any deception and the reason for it.
animal research: Research with animals, which must be approved by an animal care and use committee that checks for humane treatment and justified use.

Descriptive Statistics

Descriptive statistics summarize a set of numbers. The mean, median and mode describe the center; the range and standard deviation describe the spread. Outliers pull the mean toward them, so in a skewed distribution the median is the better summary of a typical value. A frequency distribution or histogram shows the shape: symmetric, skewed toward a tail, or bimodal with two peaks.

Central tendency: the mean ( $\bar{x}$ , the total divided by the count), the median (the middle value in order) and the mode (the most frequent value). Two peaks make a bimodal distribution.
An outlier pulls the mean toward it. In a skewed distribution the mean sits toward the tail, so the median describes the typical case better.
Variation: the range is highest minus lowest; the standard deviation is roughly the typical distance of scores from the mean.
A frequency distribution counts how often each value or range occurs; a histogram shows it as touching bars.

the mean, which uses every value and so moves toward any outlier the median, which stays with the typical values even when there are outliers the distribution is skewed and the mean moves toward the tail, so the median is the better summary the range and the standard deviation are large: more variation

descriptive statistics: Numbers that summarize a set of data, such as the mean, median, mode, range and standard deviation.
frequency distribution: A table or graph showing how often each value, or each range of values, occurs in a set of data. A histogram shows it as touching bars.
central tendency: A single value that describes the center of a set of data: the mean, the median or the mode.
mean (statistics): The arithmetic average: the sum of the values divided by how many values there are, written $\bar{x}$ .
median: The middle value when the data are put in order; with an even number of values, halfway between the two middle ones.
mode (statistics): The value that occurs most often in a set of data.
bimodal distribution: A distribution with two peaks, or two modes, often because two different groups are mixed together.
variation (statistics): How spread out the values in a set of data are; measured by the range and the standard deviation.
range (statistics): The highest value minus the lowest value in a set of data.
standard deviation: A measure of variation: roughly the typical distance of the values from the mean. A larger standard deviation means more spread.
outlier: A value far from the rest of the data. An outlier pulls the mean toward it but barely moves the median.
skewed distribution: A lopsided distribution with a long tail on one side: positively skewed when the tail stretches toward high values, negatively skewed when it stretches toward low values. The mean is pulled toward the tail.

The Normal Curve and Percentiles

In a normal distribution, scores form a symmetric bell curve around the mean. By the 68-95-99.7 rule, about 68% of scores lie within one standard deviation of the mean, 95% within two and 99.7% within three. A z score gives a score’s distance from the mean in standard deviations, which lets you compare scores from different tests, and a percentile rank gives the percent of scores below it.

A normal curve is symmetric and bell-shaped: mean, median and mode are equal, in the middle.
The 68-95-99.7 rule: about 68% of scores lie within 1 standard deviation of the mean, 95% within 2 and 99.7% within 3. Bands: 34%, 13.5%, 2.35% on each side.
A z score: $z = \frac{x - \bar{x}}{\text{SD}}$ . Above the mean, positive; below, negative.
A percentile rank is the percent of scores below a score, not the percent correct. One SD above the mean is the 84th percentile.

their frequency distribution is bell-shaped and symmetric: the normal curve the 68-95-99.7 rule gives the share of scores between the marks the z score says how many standard deviations the score is from the mean the percentile rank: 50 at the mean, 84 at one SD above, 16 at one SD below

normal curve: A symmetric, bell-shaped frequency distribution in which most scores fall near the mean and fewer fall farther out; the mean, median and mode are equal.
68-95-99.7 rule: In a normal distribution, about 68% of scores lie within one standard deviation of the mean, 95% within two and 99.7% within three. Also called the empirical rule.
z score: How many standard deviations a score is above (positive) or below (negative) the mean: $z = \frac{x - \bar{x}}{\text{SD}}$ . Also called a standard score.
percentile rank: The percent of scores in a distribution that fall below a given score.

Statistical Significance and Effect Size

Inferential statistics let researchers judge whether a sample’s result reflects the population. A statistically significant difference ( p < .05 ) is unlikely to be due to chance alone; a difference that is not significant could plausibly be chance. Significance does not measure size: effect size tells how big a difference is, and with very large samples even tiny effects can be significant.

Descriptive statistics summarize a sample; inferential statistics use the sample to draw conclusions about the population.
Statistically significant ( $p < .05$ ): a difference this large would be unlikely if only chance were at work. Not significant: the difference could plausibly be chance.
Effect size says how big the difference or relationship is. Significance is not size: big samples make small effects significant.
Always read a result for its groups: which group was higher, by how much, whether it was significant, and how large the effect was.

researchers need to judge whether a difference is bigger than chance usually produces below 5% ( p < .05 ), the result is called statistically significant a significant result can still be small, so effect size is reported to say how big it is

inferential statistics: Statistics that use data from a sample to draw conclusions about the population it came from, with a stated chance of being wrong.
statistical significance: A result is statistically significant when it would be unlikely to occur by chance alone if there were no real difference or relationship; psychologists usually require a probability (p value) below .05, written $p < .05$ .
effect size: A measure of how large a difference or relationship is, such as how many standard deviations apart two group means are, regardless of sample size.