← Introduction to Psychology

Unit 2: Research Methods and Statistics

How psychologists turn a hunch into a testable claim, the designs that separate correlation from causation, and the numbers behind a real result.

14 topics · 52 questions

Each topic is a path, not a menu: work down it in order and every question builds on the one above. Click any of them to watch the answer draw itself.

2.1 Why Psychology Needs the Scientific Method

  1. How it worksHow does a theory generate a hypothesis that a study can actually go out and test?
  2. Common mistakeWhy doesn't a theory ever get proven true, no matter how many studies support it?
  3. Going furtherWhat happens to a widely believed theory when a single well-designed study contradicts it?
  4. Worked exampleFind why the claim everything happens for a reason fails to count as a testable scientific hypothesis.
Your turn1 of 2
A friend argues that 'everything happens for a reason' explains why a coworker got laid off right before a big personal setback. A psychologist studying this friend's reasoning points out that the claim can never actually be tested. Why can't it be tested?

2.2 Case Studies, Naturalistic Observation, and Surveys

  1. How it worksHow does naturalistic observation try to capture behavior without disturbing it?
  2. How it worksWalk me through how a poorly worded survey question can quietly bias every answer it gets.
  3. Common mistakeWhy isn't a case study's finding automatically true of people in general?
  4. Going furtherWhat changes about a survey's results when the sample answering it isn't who you think it is?
Your turn1 of 2
A psychologist spends three years intensively studying a single patient with an extremely rare brain injury that destroyed only her ability to recognize faces, leaving every other visual ability intact. A student concludes that this proves all humans process face recognition in a completely separate brain system from other vision. What is the flaw in this conclusion?

2.2 Sampling and Sampling Bias

  1. How it worksHow does random sampling let a small group stand in for an entire population?
  2. Common mistakeWhy isn't a bigger sample the same thing as a representative one?
  3. Worked exampleFind the problem with a study that draws its entire sample of 2000 people from one college's psychology classes.
Your turn1 of 2
A pollster expects a true proportion near 0.618 and wants the margin of error to be no more than 0.0589 at the standard 95 percent confidence level. Using n = (1.96 squared times p times (1 minus p)) divided by margin squared, what sample size n is required?

2.2 Archival Research and Longitudinal vs. Cross-Sectional Designs

  1. How it worksHow does a longitudinal design track the same group of people across decades?
  2. How it worksWalk me through why a cross-sectional design can mistake a generational difference for an aging effect.
  3. Going furtherWhat changes about a longitudinal study's conclusions when a third of the original participants drop out?
Your turn1 of 2
A longitudinal study begins with 509 participants. By the twenty-year follow-up, 151 of the original participants are still being tracked. What percentage of the original sample has been lost to attrition?

2.2 Correlational Research and the Correlation Coefficient

  1. How it worksHow does a correlation coefficient turn a scatterplot into a single number between -1 and 1?
  2. Worked exampleFind what a correlation coefficient of -0.75 between hours of sleep and reported stress actually tells you.
  3. Common mistakeWhy doesn't a correlation of 0.8 mean the relationship holds eighty percent of the time?
Your turn1 of 2
A study reports a correlation coefficient of 0.35 between hours of sleep and next-day reported stress. Using r-squared as the proportion of shared variance, what percentage of the variance in stress is statistically explained by hours of sleep?

2.3 Why Correlation Isn't Causation

  1. How it worksHow does a hidden third variable create a correlation between two things that don't actually affect each other?
  2. How it worksWalk me through the three explanations you have to rule out before calling a correlation causal.
  3. Going furtherWhat would it take to turn the ice-cream-and-drowning correlation into an actual experiment?
  4. Worked exampleFind the likely third variable explaining why cities with more churches also tend to have more crime.
Your turn1 of 2
A study finds that towns with more fire trucks dispatched per year also have more fire damage reported per year. A newspaper headline claims 'Fire trucks cause fire damage.' What is the actual explanation for this correlation?

2.3 Reading Scatterplots at Different Correlation Strengths

  1. How it worksHow does a scatterplot's shape change as the correlation climbs from 0 toward 1?
  2. Worked exampleFind how a scatterplot of two variables looks different at r equals 0, 0.3, 0.7, and 0.95.
  3. Going furtherWhat happens to a scatterplot's shape when a correlation is strong but negative instead of positive?
Your turn1 of 2
A scatterplot of two variables shows points clustered so tightly that they nearly form a single straight line sloping downward from upper left to lower right. Which correlation coefficient is most consistent with this pattern?

2.3 Experiments: Independent, Dependent, and Confounding Variables

  1. How it worksHow do the independent and dependent variable split the two halves of an experiment?
  2. How it worksWalk me through a confounding variable riding along with the treatment and stealing the credit for the result.
  3. Worked exampleFind the independent, dependent, and confounding variables in a study testing whether background music raises test scores.
  4. Common mistakeWhy isn't a study with two groups automatically a controlled experiment?
Your turn1 of 2
A researcher wants to know if extra sleep improves memory test performance. Group A sleeps 8 hours before testing; Group B sleeps 5 hours. All testing happens in the same room at the same time of day. What is the independent variable in this experiment?

2.3 Random Assignment, Control Groups, and Blinding

  1. How it worksHow does random assignment scatter every unmeasured difference evenly across the groups?
  2. How it worksWalk me through what a double-blind design hides from the participant and from the researcher.
  3. Common mistakeWhy isn't random assignment the same procedure as random sampling?
  4. Going furtherWhat happens to a drug trial's results when the placebo group also gets better?
Your turn1 of 2
In a drug trial, 79 percent of the drug group's symptoms improved, while 29 percent of the placebo group's symptoms also improved from expectation alone. What percentage of improvement can be attributed to the drug's actual effect, beyond the placebo response?

2.2 Operational Definitions, Reliability, and Validity

  1. How it worksHow does an operational definition turn a fuzzy concept like aggression into something a second lab can measure the same way?
  2. How it worksWalk me through the difference between a measure that's reliable and a measure that's valid.
  3. Common mistakeWhy can a measure be perfectly reliable and still not be valid at all?
  4. Worked exampleFind an operational definition of stress that a second lab could measure exactly the same way.
Your turn1 of 2
A researcher splits a 20-item personality scale in half and correlates the two halves, getting a split-half correlation of 0.648. Using the Spearman-Brown formula, full-scale reliability equals 2 times r divided by (1 plus r), what is the estimated reliability of the complete scale?

2.3 Central Tendency and Variability

  1. How it worksHow do the mean, median, and mode each try to summarize a distribution with one number?
  2. Worked exampleFind the mean, median, and mode of the quiz scores 60, 70, 70, 80, 95.
  3. How it worksHow do the range, variance, and standard deviation each capture how spread out a set of scores is?
  4. Worked exampleFind the standard deviation of the scores 2, 4, 4, 4, 6 for a data set with a mean of 4.
Your turn1 of 2
A small study records five participants' scores on a stress questionnaire: 31, 34, 48, 33, and 32. What is the mean score?

2.3 The Normal Distribution and Standard Scores

  1. How it worksHow does a normal distribution split up predictably into percentages by standard deviation?
  2. Worked exampleFind what percentage of scores fall between one standard deviation below and one above the mean of a normal distribution.
  3. Worked exampleFind where an IQ score of 130 falls on a normal distribution with a mean of 100 and a standard deviation of 15.
  4. Going furtherWhat happens to the shape of the curve when a distribution is skewed instead of normal?
Your turn1 of 2
A participant scores 116 on an anxiety inventory with a population mean of 104 and a standard deviation of 19. How many standard deviations above or below the mean is this score? (A negative value means below the mean.)

2.3 Statistical Significance and Effect Size

  1. How it worksHow does a p-value estimate the odds that a result happened by chance alone?
  2. Common mistakeWhy doesn't a statistically significant result automatically mean a large or important one?
  3. How it worksHow does an effect size measure how big a difference actually is, separately from whether it's significant?
  4. Going furtherWhat happens to a tiny effect size's p-value when the sample size gets enormous?
Your turn1 of 2
Two therapy techniques are compared on reducing anxiety scores: Technique A produces a mean reduction of 25 points, while Technique B produces a mean reduction of 11 points. If the pooled standard deviation is 10, what is Cohen's d effect size for this comparison?