Terms  Definitions 

mesokurtic 
mid level height

sample 
a subset of the population

Reliability 
How stable the measure is

Domainreferenced tests 
Attempt to measure lessdefined properties (like intelligence) and need to be checked for reliability and validity

Inferential statistics 
Prevent unreasonable conclusions from data

hypothesis testing 
decisions are made concerning the values of parameters. key to inferential stats. it takes the data results and compares it to the population which allows you to make predictions based on the results.

Construct validity 
External validity Whether the test really taps the abstract concept being measured

Splithalf reliability 
Measured by comparing an individual's performance on two halves of the same test (odd vs. even questions)  reveals internal consistency. Can also increase internal consistency of a measure by performing an item analysis

error factors of alpha 
set by the experimenter

Parametric statistics like the ttest assume that the data collected is... 
Normally distributed

Interval variables 
Shows order and spacing, equal space lie between the values, do not include a real zero

median 
the point on the scale of measures where the population is centered. The measure of a population is the point that divides the distribution of scores in half.

Random Sample 
Requires that each individual in the population has an equal chance of being selected. Probablities must stay constant from one selection to the next if more than one individual is selected.

the two possible sampling situations: 
sampling with replacement, sampling without replacement

Standard deviation 
Square root of the variance, tells you the average extent to which scores were different from the mean (large SD= scores were highly dispersed, SD small= scores were very close together)

nominal scale 
names, eg. list of students in alphabetical order

convicting an innocent woman of a crime is an example of what error? 
Type I error

Typer II Error 
Occurs when a researcher fails to reject a null hypothesis that is really false. Person has cancer, tell them they dont

directional (one tailed) test 
designed to detect extreme outcomes in only one specified direction of the distribution

Standardized tests 
Test that are tried out on huge groups of people in order to create norms

obstacles to using dependent 
if impossible to participate in both conditions (normal vs parkinsons, men vs women)if participation in one condition affects other conditions (conjunction fallacy, solving a reasoning problem)

Typical application of hypothesis testing with two related samples 
Within subjects design with two conditions

type I error 
You have committed a this error if you have rejected the hypothesis tested when it was true. More seriousmore serious

two tail  non directional 
no prior evidenceH0: u=aH1: u NOT =a

four components of statistical power 
1. sample size: # of units studied2. effect size: strength of the relationship between 2 variables.3. alpha level: odds that the observed result is due to chance (normall is .05)4. power: the odds that youll observe a treatment effect when it occurs

When do you accept your null? 
if probability of sample result is above alpha

Spearman r correlation coefficient 
Used when the data is in the form of ranks

why must you subtract the joint probabilities of a and b when doing an adding probabilities problem? 
So that you do not doublecount your data

Leave a Comment ({[ getComments().length ]})
Comments ({[ getComments().length ]})
{[ comment.comment ]}