cueFlash

Glossary of Psych Statistics

Deck Info

Other Decks By This User

Descriptive Statistics
method for organizing and describing experimental data
Inferential Statistics
method for making decisions about the significance of experimental results. were the results caused by the IV or just by chance
cumf
the cumulative frequency. the number of scores at or below any given score
cumprop
the proportion of scores at or below any given score ( take the cumf and divide by the total number of scores)
%ile
percentage of scores at or below any given score
x
the data (scores) arranged from lowest to highest
f
the frequency with which each score occurs in the data
mean
mathematical average of a group of scores, the sum of the scores divided by the number of scores
(most vulnerable to extreme scores)
median
the middle score, at the 50%ile where 1/2 the scores are above and 1/2 the scores are below(best used for skewed distributions)
range
the simplist measure. the difference between the highest and lowest scores in the distribution
Deviation
the difference between a given score and the mean of the distribution D=x - xbar
Variance
the square of a scores deviation from the mean d2=variance.
standard deviation
the avergae deviation of all scores from the mean
calculating standard dev.
1. find mean of distributon
2. create freq. dist. chart
-x(score)
-f(freq.)
-d(deviation, the score-mean)
-d2(variance, deviation squared)
-fd2(variance * freq)

1 add up all the numbers in fd2
2 divide dum of fd2 by N
3 take square root of result
Normal frequency distribution
symmetrical all around
mean=median=mode
Non-Normal
(Skewed Curves)
skewed to right or left. either few extremly high scores or a few extremly low scores
Non-Normal
(Ogive Curve)
s-shaped curve. either all high scores or all low scores.
Non-Normal
(Bimodal Curve)
m-shaped. has two modes.
Non-Normal
(U-shaped Curve)
U-shaped. extreme scores on both ends.
Normal
(Leptokurtic Curve)
small standard dev. (scores all bunched up around mean)
Normal
(Platykurtic)
large standard deviation (scores very spread out)

((very LOOOOONG)
Normal
(Mesokurtic)
classic shape. Standard normal curve.
xbar
+1, +2, +3 numbers
50%tile
34.13%
13.59%
2.15%
.14%
percentiles can also be seen as...((13.59%))
probabilities ((P=.1359))
z-score defintion
allows us to determine the exact distance of any given score from mean (in standard deviations). All you need is the mean and stand. dev.
z-score formula
z= (x-xbar)/ standard dev.
example of z-score
IQs=140 xbar= 200 SD=30
IQmj=60 xbar=30 SD=10
s=-2
mj=+3
correlation
refers to the type and the strength of the relationship between two variables. Correlation data does not indicate a cause and effect relationship.
correlation coefficient
Pearson's r
+1=perfect/direct + rel.(Aup B up)
0.00=no relationship
-1=perfect/indei2rec/inverse/- relationship (Aup B down)
Logic of inferential statistics
to decide if the difference between the experimental group mean and the control group mean was caused by the IV or just chance
Null Hypothesis (Ho)
Nearly always define the Ho as follows "the difference between the xbar"e" and the xbar"c" is caused due to chance" (Assumption/Theory)
(skepticism)
Ho is correct
Fail to reject the Ho
assert difference is due to chance
Ho is incorrect
Reject the Ho
assert difference is caused by IV
alpha level
the probability of an event occuring by chance at which one is willing to reject the Ho
alpha=.01
p=1 of 100 times (by chance)
alpha=.05
p= 5 of 100 times (by chance)
if Xbar"e" < alpha...
reject Ho
caused by IV
if xbar"e" > alpha...
fail to reject the Ho
caused by chance
diff between descriptive/inferential
make decisions with inferential andnot with descriptive. describe with descriptive
if r=+1.00 and variable A is decreasing what is happening to variable b
also decreasing
mean is 10. sd is 2. what is score at z=+2
14
cumulative freq.
number of scores @ or below or @ and above
type 1 error
increases as alpha increases=rejecting the ho when it is true=stating IV caused the result when it was really due to chance
type 2 error
increases as alpha gets smaller=failing to reject t he Ho when it is really false=stating the result was due to chance when was due to IV
ANOVA
stat. use for multiple IV's
students t
(t-test)
used when control group represents the entire pop.
omega-squared
correlation between IV and DV= helps detect interactions between variables
chi-squared
compares observed outcomes(data) with expected outcomes (Ho)