Skip to content
AP Psychology · Cognition

Inferential statistics

Inferential statistics appears in 10 reviewed AP Psychology questions and sits in Cognition. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

Inferential statistics are tests that tell researchers whether a result from a sample is likely to reflect a real effect in the population or is likely due to chance. A result is usually called statistically significant when p is less than .05.

Why it matters

Every experiment in the course depends on this step. Without it, you cannot tell whether a difference between groups means anything.

How the exam tests it

Questions often give a nonsignificant result, like p = .12 in a 20-person study, and a student who says it proves no effect. That is wrong, because failing to reject the null hypothesis does not prove it true. A small sample may lack the power to detect a real effect, which risks a Type II error.

Don’t confuse it with

Students mix it up with descriptive statistics, but descriptive statistics such as the mean and standard deviation summarize the data you have, while inferential statistics judge whether you can generalize beyond that sample.

Inferential statistics in real questions

Example 1Harder
A researcher hypothesizes that people conform to a unanimous group's incorrect judgment more when they must answer aloud in front of the group than when they answer privately in writing. In a study of 20 participants, the public-response group conformed slightly more often than the private-response group, but the difference was not statistically significant (p = .12). A student concludes that the study proves response privacy has no effect on conformity. Why is the student's conclusion incorrect?
  1. AFailing to reject the null hypothesis does not prove it true; a sample this small may lack the power to detect a real effectCorrect
  2. BA p-value of .12 means there is an 88 percent chance that privacy really does affect conformity, so a real effect is quite likely
  3. CBecause the public group conformed more often, the raw difference shows that privacy does affect conformity
  4. DA p-value of .12 is below the .15 cutoff that psychologists use, so the result counts as significant

Why A is right

The research-methods idea is that a nonsignificant result means the researcher fails to reject the null hypothesis. It does not prove the null is true. With only 20 participants split between two conditions, the study has low statistical power. A real situational effect on conformity, such as more conformity under public responding (the normative social influence covered in topic 4.3, Psychology of Social Situations), could exist and still go undetected. That would be a Type II error. Absence of evidence is not evidence of absence, so the student cannot conclude that privacy has no effect.
Why the other 3 choices miss
B
Incorrect because a p-value is not the probability that the hypothesis is true or false. It is the probability of getting results at least this extreme if the null hypothesis were true, so 1 - p does not give the chance of a real effect.
C
Incorrect because a raw difference between group means can easily come from chance (sampling error). Inferential statistics exist to judge whether such a difference is reliable, and this one was not statistically significant, so it cannot show that privacy affects conformity.
D
Incorrect because the conventional alpha level in psychology is .05, not .15. Since p = .12 is above .05, the result is not statistically significant.

Studied alongside

Other concepts from Cognition.