Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
43 Cards in this Set
- Front
- Back
frequency
|
number of participants or cases; N=a population, n=a sample
|
|
proportion
|
part of 1
|
|
frequency distribution/polygon
|
a table/drawing that shows how many participants have each score
|
|
positive skew
|
distribution curve with longer tail to the right
|
|
negative skew
|
distribution curve with longer tail to the left
|
|
bimodal distribution
|
has two high points, most likely to emerge when human intervention or a rare event has changed the composition of a population
|
|
mean
|
balance point in a distribution of scores, or, the point around which all the deviations sum to zero
|
|
computation of mean
|
sum scores and divide by number of scores, symbolized by M, or m, or (x-bar)
|
|
notes on mean
|
pulled in direction of extreme scores therefore inappropriate for skewed distributions; should only be used with interval and ratio scales
|
|
median
|
middle point of a distribution, 50% of cases above and 50% below it.
|
|
notes on the median
|
insensitive to extreme scores therefore good for skewed distributions; cannot be used on nominal data
|
|
mode
|
most frequently occurring score; can be used for nominal data, though percentages may be more informative
|
|
variability
|
differences among scores; also called spread or dispersion
|
|
range
|
difference between highest and lowest score (but being based on the two extreme scores can be a weakness)
|
|
outliers
|
scores that lie far outside the range of most other scores
|
|
interquartile range
|
range of the middle 50% of the participants (ignores outliers)
|
|
standard deviation
|
most frequently used measure of variability; provides an overall measurement of how much scores differ from the mean
|
|
standard dev. and normal curve
|
about 68% or two-thirds of cases will lie within one standard deviation of the mean
|
|
correlation
|
extent to which two variables are related
|
|
census
|
study in which all members of a population are included
|
|
direct relationship is also called
|
positive relationship (high or low on both variables)
|
|
inverse relationship is also called
|
negative relationship (high on one variable, low on the other)
|
|
correlation is not
|
causation
|
|
what is needed to study cause and effect?
|
a controlled experiment
|
|
Pearson correlation coefficient (Pearson r, Pearson product-moment correlation coefficient)
|
describes relationship between 2 variables: -1 (perfect inverse) to 1 (perfect direct relationship); 0 is absence of relationship (this is NOT a proportion)
|
|
coefficient of determination (r squared)
|
when converted to a percentage (by multiplying by 100) tells us how effective one variable is in predicting another
|
|
null hypothesis
|
for the difference between 2 sample means, the true difference between the means is zero; or, there is no true difference between the means
|
|
research hypothesis
|
a researchers "expectation" or personal hypothesis
|
|
directional hypothesis
|
that one group's average will be higher than another's
|
|
nondirectional hypothesis
|
that one group's average will be higher but there's insufficient info to say which
|
|
significance test (p-italicized)
|
tests the null hypothesis and yields a probability that it is true
|
|
what kinds of tests are used to test the null hypothesis?
|
inferential tests
|
|
probability level
|
(also called alpha level) the level for rejecting the null hypothesis, that is, that it is not true; commonly .05
|
|
Type I error
|
the error of rejecting the null hypothesis when it is correct
|
|
synonym for "rejecting the null hypothesis"
|
results are statistically significant
|
|
Type II error
|
error of failing to reject the null hypothesis when it is false
|
|
t-test
|
tests the difference between two sample means to determine statistical significance and yields a probability that the null hypothesis is correct
|
|
3 factors resulting in a low probability of a correct null hypothesis (ie, that it will be rejected)
|
1) larger samples (reduces sampling error)
2) larger difference between means (random sampling produces few large differences) 3) smaller variance (less sampling error) |
|
chi-square
|
tests for differences among frequencies, ie, for nominal data where number of cases and percentages are reported (can't compute mean or standard deviation for nominal data)
|
|
is significant difference (ie, reliable) the same as a large difference?
|
no
|
|
limitations of significance testing
|
1) fails to indicate size of difference
2) does not assess the practical significance of a difference |
|
effect size
|
standardizes the size of the difference between two means
|
|
Cohen's d
|
measure of effect size, in standard deviation units (there are only about 3 standard deviation units above and below the mean).
|