Descriptive + inferential stats Flashcards
(18 cards)
cumulative frequency
scores at or below each raw score
should be total eg 200 students in test
percentile
cumulative/total number of scores x 100
displays relative standing
should be 100
central tendency
outliers
what scores are most frequent
- mean
- mode
- median
outliers make mean unrepresentative of data set
variability
how different scores are from each other
- range, IQR
- standard deviation
Z-score standardised score
scores relative standing by how much it deviates from the mean of its distribution
- score less than mean = negative z-score
- score above mean = positive z-score
normal distribution
- symmetrical about the mean
- 50% of scores above and below
- tails approach x-axis asymtotically
z-score probability
think - normal distribution
normal distribution low freq scores less probable, high freq more probable
- z-score of 0 most frequent
- higher z-score = lower probability
probability is
the relative frequency of an event compared to all possible events
- proportion of 0-1
z-score normal distribution
TABLE….
mean = 0
SD = 1
frequency gets smaller away from the mean
CHECK IF ITS POSITIVE OR NEGATIVE SCORE WHEN USING TABLE
z-score intervals
intervals –> area of scores
-1to1 = 68%
-2to2 = 95%
-3to3= 99%
normal distribution types
leptokurtic = skinny up high like leaping
platykurtic = wide like playtpus
all z=0 to z=1 are 34.13% of raw scores
How to find exact probability of particular score in distribution?
- need mean + SD
- convert raw to z-score
- probability in z-table
STEPS
IQ mean 100, SD 15, what is probability of having IQ of 116 or lower?
- find area of curve thats at or below 116
- calculate z score 116-100/15 = 1.07
- find probability on POSITIVE z-table = .85769
- this means probability is .85769
- percentile = 85.769%
if m 100, SD 15, what is IQ score above which 5% of popn fall?
- what score –> its at or below 95%
- find in z-table
- get the z-score numbers (1.64)
- what IQ corresponds to this (Rearrange formula)
–> 100 + 1.64x15 =124.6 IQ
standard deviation of a distribution of sample means =
standard error of the mean
population SD = sigma ó, M = u(mew)
sample means SD = sigma xbar, M = u(mew)xbar
z-tests
central limit theorem
using z-scores to tell aabout standing of a sample within a group
- increasing samples gives normal distribution (sample means)
theorem = samples always norm dist as long as sample is large enough
Is a sample representative of the popn?
if probability is high (p>.05) sample is representative
- if probability is low (p<.05) sample is NOT representative (likelihood getting it by change is low)
z= 2.5
z= 1.07
CHECK POSITIVE OR NEGATIVE
- use left side of table
- across top row then column (meet to find value)