psych 2811 final

*made this for my OWN study use, and I highly suggest not using this as your only resource

DeMére paradox

a probability puzzle that asks if it's more likely to get at least one six in four rolls of a die or at least one double six in 24 rolls of two dice (gambler’s fallacy)

gambling

where did the study of probability originate?

p(a) + p(b) - p(a and b)

p(a or b) for independent events

p(a) x p(b)

p(a and b) for independent events

marginal probability

when calculating probability, looking at only one variable in the study (e.g. probability of getting an A)

joint probability

when looking at a contingency table, calculating probability for two factors (e.g. being a woman and getting an A on an exam)

point estimate

single point to represent the population (e.g. mean or median of the sample)

p(a and b) / p(b)

p(a | b) for a contingency table

right skewed

skew of a distribution if median > mean

left skewed

skew of a distribution of median < mean

1.5 x IQR

calculation for if something is an outlier

highest value - lowest value

calculating range of a dataset

IQR

middle 50% of your data (Q1-Q3)

varience

how variable is this data from the mean?

standard deviation

square root of variance; average that a single data point will deviate from the mean

mean, variance, standard deviation

statistical measures most vulnerable to outliers

uniform

a distribution where each point appears with same frequency

p(b|a) = [p(a and b) x p(b)] / p(a)

Bayes’ Theorem

leptokurtic

a distribution with long tails

platykurtic

a distribution with short tails

ad hoc

sample type based on convenience; likely to be biased and potentially not-representative

sampling error

difference between true population parameter (theoretical) and result from sample; not a mistake, just a thing that happens

measurement error, calculation error, misinterpretation

types of non-sampling error

sample is random, normally distributed, and scores are independent from each other

assumptions made about a sample with a confidence interval

bootstap sample

re-takes your sample multiple times with replacement to estimate the population. does not require assumptions about underlying distribution, but can have issues if sample is too small

confidence interval

range of results to expect if we repeated the exact experiment an infinite amount of times

null hypothesis

assumes the opposite of your hypothesis is true

p-value

shows how rare your data is. higher = less likely your data is result of random chance `

0.05

common P value

type i

error in which you reject the null hypothesis, even if it is true

type ii

error in which you fail to reject the null hypothesis, even if it is false

p

lowering this value increases chance of type ii errors, but increases chance of type i errors. increasing it does the opposite

single sample tests

help compare sample data to a common parameter (e.g. normalized IQ tests)

z score

scores on the standardized normal that describe how many standard deviations a result is from the mean

Cohen’s D

measure of how many standard deviations the mean of a sample is from the population score (calculated in units of SD)

t statistic

difference between a parameter estimate and either a hypothesized parameter or another study. normalized by standard error. leptokurtic

Gasset

published the original T distribution under psudonym “student”

IV (categorical with 2 levels), DV

what do you need for a 2 sample t test?

normal distribution, no outliers, similar variance in both samples, independence

assumptions you make with a two sample t-test

if samples came from same pop or different ones; group differences vs sampling error

what does a two-sample t-test tell you?

variance of DOSM of mean 1 + variance of DOSM of mean 2

variance of distribution of differences between means

validity

How much a measurement or study actually co-responds to the real world

random sample, normally distributed, independent

assumptions we make when making a confidence interval

bootstrap

Re-takes your sample with replacement to create a DOSM. May be noisy if sample has a small N

Karl Pearson

formerly introduced the P statistic