Research and Program Evaluation Flashcards
parsimony
interpreting results in the simplest way
occams razor
parsimony = interpreting results in the simplest way
RA Fisher
pioneered hypothesis testing
P
probability
type i
reject the null when its true
alpha
type ii
accept the null when its false
beta
increased sample size means…
decrease of type i and ii errors
t - test
two sample groups, use t-table, similar to z table
t score = (Z x 10) + 50
t score M=50 SD=10
positively skewed
tail to the right
negatively skewed
tail to the left
abscissa
x-axis
ordinate
y-axis
range
highest - lowest (sometimes plus 1)
how does range react to sample size
range increases with sample size increase
important variance/sd vaues
1 SD = 68
2 SD = 95
3 SD = 99
platykurtic distribution
flatter curve
required response rate for surveys
50-75%
hawthorne effect
when subjects perform better just because they are part of an experiment, reacting to the presence of the investigation
rosenthal effect
experimenter expectancy effect, the experimenter beliefs about the subject or the test may cause special or different treatment for those subjects
halo effect
when a trait not being measured influences the rating of a trait that is being measured. attractiveness and likability
analysis of covariance (ANCOVA)
allows the researcher to correct for differences in covariates between various samples.
ways to test differences between groups or ANOVA’s
duncans multiple range, tukeys or scheffes test
types of sampling for a true experiment
random sampling and systemic sampling
indicative vs deductive reasoning
inductive = goes from one case to a generalization across a population deductive = takes many cases and applies to one case