Chapter 4 Flashcards
variability due to chance
error variance, the fact that for any given pop. with standardized mean of 50, the true mean may vary some (for example, one group’s mean may be 49.7 and another group’s mean may be 52.3)
sampling error
numerical value of a sample statistic probably will be slightly in error as a result of of the particular observations that happened to be included in the sample (like one unusually obnoxious child in a group of children)
hypothesis testing
where data is ambiguous, attempting to determine if difference is small enough to dismiss as chance or large enough to entail significance
sampling distribution of differences between means
distribution of differences between means of an infinite number of pairs of random samples drawn under certain specified conditions
research hypothesis
idea researchers are hoping to prove, such as “Kids who eat more sugar behave more wildly.”
null hypothesis
idea opposite to what researchers are hoping to prove, like “Kids who eat more sugar don’t behave more wildly.”
Fisher-We cannot prove that something is true, but we can prove something to be false. One can never claim to have “proved” the null hypothesis, but failure to reject the null hypothesis often means we have not collected enough data.
alternative hypothesis
To prove the truth of a hypothesis, one would need an absolute value (which cannot be provided). For example, what is the absolute level one could assign to “wildness”? Instead, we must use the null hypothesis method for behavioral sciences.
sample statistics
describe characteristics of samples
test statistics
associated with specific statistical procedures and have their own sampling distributions
decision-making
using arbitrary conventions to determine whether we should accept or reject our data
rejection level
significance level-arbitrary number at which we choose to reject our findings (either less than or equal to .05 or .01)
critical value
actual score that cuts off the highest 5%
Type I error
erroneously rejecting the null hypothesis
alpha
the probability of reaching a Type I error, erroneously rejecting a true null hypothesis
Type II error
erroneously failing to reject the null hypothesis
beta
the probability of reaching a Type I error, erroneously failing to reject the null hypothesis
power
probability of correctly rejecting the null hypothesis when it is truly false, power=1-beta
one-tailed test
directional test-the situation in which we reject the null hypothesis for only the lowest or only the highest mean differences; predicting that the individual will differ from the mean in only one direction of the mean (either above-positively or below-negatively the mean)
two-tailed test
nondirectional test-the situation in which we reject the extremes at both the positive and the negative ends, used either in exploratory testing where the experimenter has no idea as to how outcomes will be affected or when an experimenter does have a good idea but wants to CYA, or when more than two groups are being tested and a one-tailed test wouldn’t work
conditional probabilities
confusing the probability of the hypothesis given the data with the probability of the data given the hypothesis
Tukey & Jones method
Instead of predicting an hypothesis (and a null hypothesis) before beginning, the experimenter simply gathers data and analyzes it before trying to draw conclusions. This is incredibly helpful because in most experiments, proving the null hypothesis perfectly is impossible and therefore moot.
