Exam 1 Flashcards
Observational research
- investigator measures relationships between events or conditions
- no manipulation
experimental reserach
involves an investigator that directly manipulates conditions to measure a response
reliable
consistent across repitions
valid
obtained with sound method of measurement
precise
collected values are most nearly accurate
objective
data is collected by an impartial investigator
constant
characteristic that can only assume one value does not need to be measured
variable
any characteristic that can “vary” what is measured and typically makes up data
discrete variable
limited to certain values
- e.g. whole numbers or categories
continuous variables
theoretically assume any value
nomial scale
mutually exclusive categories, no logical order
ordinal scales
logical order, but no indication of size of difference (rank order)
ratio scale
equal intervals and an aboslute zero
interval scale
equal intervals but no absolute zero
what is an absolute zero
indicates an instance where variables dont exist
internal validity
result observed in the DV is entirely due to the treatment of the IV
external validity
the result can be generalized to the wider world
random sample
each member of the population has an equal chance of being selected
stratified sample
ensure representation of subgroups within the population of interest
convenience sample
members are selected based on “ease and proximity”
Percintles
comparison to the range of scores/characteristics obtained from the larger sample or population
what is the equation to calculate percentile
percentile = # of scores at and below the value/total # of scores
what is the equation to calculate raw scores from percentiles with a frequency distribution dataset
(PN-C/f)i +L
probability
long run proportion of a particular outcome
trimmed means
mean that ignores the highest and lowest values
geometric mean
restricted to positive scores (financial investments, returns, growth rate)
harmonic mean
price to earning ratios in finance (also restricted to positive scores)
equation for variance
V = (Sum (x-mean)^2)/(N-1)
standard deviation
the square root or variance
coefficient of variation
CoV = SDx/mean * 100
when to use coefficient variation
when there are two datasets are different in magnitudes to compare to each other
Z-scores
- used to figure out how far away a piece of data is from the rest of the group
- express the raw score in SD units
what is the quation for Z scores
Z = x - mean/SDw
T scores
used to compare a sample population mean to the mean of the population. esp when sample pop is small or SD is not there
- only can be used on normally distributed data
equation for T scores
T = 10Z + 50
positive skew
the tail is in the positive direction
no skewness
a normal distributoin
negative skewness
the tail is in the negative direction
what is the equation for skewness
S = sum Z^2/N then divide by SEs = sqrt (6/N)
kurtosis
the steepness of a distribution
platykurtic
wide range of scores, low concentration around mean
- k<0
leptokurtic
narrow range of scores high concentration around mean
- K > 0
mesokurtic
moderate range, moderate concentration
what is the equation for kurtosis
K = (sumZ^4/N) - 3.0 then difvide by SEk = sqrt (24/N)