Final Exam Flashcards
real limits
0.5 above the highest number and 0.05 below the lowest number
frequency distribution, why is it useful
Present score values and their frequency of occurrence
shows the entire data
ungrouped vs grouped freq. dist.
ungrouped: raw scores can be pulled out, can find individual score
grouped: raw scores cannot be pulled out, no individual score, information is lost
relative frequency
proportion of total in interval
relative f = f/N
* should add up to 1
cumulative frequency
number of scores that fall bellow the upper real limit
* start at the bottom then add up
* final answer should equal to N
cumulative percentage frequency
Percentage of scores that fall below upper real limit of each interval
Cumulative % f= (f/N) x 100
percentile point
% that falls bellow a specific percentage (P30 = 30% of scores fall below this point)
percentile rank
percentage of scores with values lower than the score in question
ex: 10% of scores fall below 59
bar graph vs histogram vs frequency polygon
bar graph: nominal and ordinal data, bars do not touch
histogram: interval and ratio data, bars do touch
polygon: interval and ratio data, point places above midpoint of each interval
symmetrical vs positive vs negative
symmetrical: mean = median
positive: mean > median
negative: mean < median
mean vs median vs mode
mean: is affected by outliers
median: is not affected by outliers
mode: used for nominal data
correlation coefficient (pearson r)
specific measure of correlation, -1 to 1
coefficient of determination (r^2) EFFECTSIZE
% of variability in one variable with is determined by its relationship with the other variable
spearman rho
used for ordinary scaling
assumptions for pearson r
only for interval and ratio scales
regression
predicting
regression line
to predict a score of one variable based on our knowledge of another
regression constant
where the regression line crosses the y-axis
conditions to use linear regression
- has to be linear
- random sampling
- predictions that lie within the range not outside
- not interest in the individuals that are used in the linear regression
standard error of estimate
how far away, on average, a point will be from the regression line
random sampling, why is it important?
equal chance of being selected, decreases error
two ways to conduct random sampling
- sampling with replacement
- sampling without replacement
sampling with replacement vs without
with: the selected is returned, doesn’t change the probability
without: the selected is not returned, changes the probability
a priori vs a posterior (probability)
priori: before hand, based on rationalism
posteriori: after the fact, based on empiricism
addition probability rule
p(A or B)= p(A) + p(B) - p(A and B)
mutually exclusive: p(A) = p(B)
multiplication probability rule
INDEPENDENT: p(A and B) = p(A) x p(B)
DEPENDENT: p(A and B) = p(A) x p(B|A)
Mutually exclusive: p(A and B) = 0
independent vs mutually exclusive vs mutually exhaustive events
independent: does not influence e/o
mutually exclusive: cannot happen at the same time, has to be OR
mutually exhaustive: includes all possible events
sampling distribution of F
used to evaluate mean differences between two or more treatments
F distribution characteristics
- F-ratio is always positive
- treatment effect is no greater than the variance expected by chance alone
- for statistical significance obtained > critical
Ho and H1 for one-way, independent groups ANOVA
Ho is true: effect is near zero, F is near 1.00
H1 is true: effect is more than 0, F > than 1.00`
Why is MS between sensitive to the IV but not the MS within?
MS between uses df between which subtracts the IV by 1 while MS within uses df within which N is subtracted by IV
assumptions for one-way ANOVA test
- 3 or more groups being compared
- independent groups
- IV is nominal and DV is interval or ratio
- homogeneity of variance
- N > 30
omega^2 vs eta^2
omega^2 is less biased than eta^2
- both provide variance
planned vs post hoc comparisons
planned (a priori): comparisons that are hypothesized before the experiment, based on theory and previous research
post hoc (posteriori): comparison after seeing the data or to explore differences
do planned or post hoc have more power?
planned comparisons have more bower but has a higher chance of type 1 error