Statistics and Distributions Flashcards
measures of dispersion
descriptive statistics
how tightly clustered are individual data points to the mean
variance
descriptive statistics
difference between each data point and mean
square each value
divide sum by (n-1)
Standard deviation
the square root of the variance
the extent of deviation for a group as a whole
probability distributions of a discrete variable
identifies the probability of each value
probability distributions of a continuous variable
identifies the probability of the value falling within a particular interval
all probability distributions add up to ____
1 or 100
percentile
1% cut off
decile
10% cut off
tertile
1/3 cut off
quartile
1/4 cut off
quantile
1/5 cut off
positive skew
heavy right tail
negative skew
heavy left tail
are +/- 1 SD away from mean
68%
area +/- 2 SD away from mean
95%
are +/- 3 SD away from mean
99.7%
mean IQ is
100
z score is
number of SDs away from the mean
conceptual hypothesis
texts that concern a topic familiar to a reader are easier to comprehend
research hypothesis
operationalizes conceptual hypothesis
statistical hypothesis
whether selected hypothesis about whether selected parameters of data population are are equal or different
null hypothesis
baseline, statement of equality
alternative hypothesis
statement of inequality typically what you are trying to prove
we use _______ to test the null hypothesis
inferential statistics
to prove null hypothesis is wrong we just need to prove that _______ is right
alternative
statistically significant p=___
0.05 or 0.01
t test
evaluation of difference between 2 variables relative to variability in the data
f test
used to compare means of more than 2 groups, compares between group variability with within group variability
Inferential outcomes: correct
reject null hypothesis which is in agreement with true population value
Inferential outcomes: false positive, type 1 error
reject null hypothesis when there is no true effect
Inferential outcomes: power
do not reject null hypothesis and rightly so
Inferential outcomes: false negative, type 2 error
we do not reject null hypothesis when there is an effect truly there
ho do we decrease type 1 errors
by moving judgment criteria
is 1 SD from mean enough to be signifigant
No
inferential tests
estimate how certain we are about rejecting the null hypothesis
standard error
how close mean of sample is to true population mean
like the SD of inferential statistics
increase power by … (3)
increase sample to decrease standard error
make signifigant threshold less strict
make data less variable
participation observation
enter the world of people
laboratory observation
controlled lab
interviewing
honest answers?
intimidating
get direct info
phone interview
no travel
selection bias
quesstionares
cheap and fast
perfect correlation
correlation of 1 or -1
semipartial correlation
a measurement of the correlation between 2 variables that remains after controlling for the effects of 1 or more predictor variables
standard score
(individual score-mean score) / SD
r^2
squaring correlation
coefficient of determination
calculates effect for correlations
indicates proportion of variance
eta squared
proportion of the total variability of the DV that is accounted for by the IV
cramers Φ
chi squared test for independence with 2 rows and 2 columns
cramers v
chi squared for larger than 2x2
cohens w
chi-squared goodness of fit
cohens d
effect size calculation that looks at the difference between the means of two groups and takes into account the variability using the pooled sample standard deviation
