When to use what stats test

?
Want to estimate the mean of a population and compare it to a target/reference when you know the standard deviation
one-sample z test
1 of 40
compare the mean of an unknown population with a known/specified population mean
one-sample t-test
2 of 40
e.g. of a one-sample t-test
5 years ago, HKU GPA average was 3. have recorded 40 students GPAs from this year. See if there has been a change
3 of 40
DF for a one-sample t-test
n - 1
4 of 40
Compare means of two sets of scores from the same sample/related samples
Paired-samples t-test
5 of 40
e.g. of a paired samples t-test
Studying a method of leaning. Ppts tested before and after using the method. Is there a mean difference?
6 of 40
DF for paired samples t-test
n-1
7 of 40
Compare means of two different groups
Independent samples t-test
8 of 40
e.g of an independent samples t-test
Two different drugs, compare the effects of the two drugs
9 of 40
DF for independent samples t-test
(n1-1) + (n2-1) ....
10 of 40
Want to see if there is an association between two variables. Data for both variables is continuous and paramentric
Pearson's correlation
11 of 40
Want to see if there is an association between two variables. Data is ranked
Spearman's correlation
12 of 40
How to carry out a spearman's ranked correlation
rank x, rank y (keep pairs together). Rx-Ry = D. Square so all positive. rs = 1- (6ΣD² )/n(n² -1). n = no. of pairs
13 of 40
Want to find an association. One variable is continuous, one is dichotomous
Point biserial correlation
14 of 40
Want to find an association, both variables are dichotomous
Phi correlation
15 of 40
When would you use a regression analysis?
To find equations that may fit your data
16 of 40
What is the linear regression equation?
y=bx+a
17 of 40
What is the multiple regression equation?
y = β1x1 + β2x2 + βpxp + β0 + E (β0 = intercept, E = error)
18 of 40
When do you use a multiple regression?
When there is more than 1 x
19 of 40
Want to see if data matches a hypothesised distribution
Chi² goodness of fit test.
20 of 40
How to carry out a chi² goodness of fit test.
Must have an idea of how data should be distributed e.g.percentages. Convert % to actual scores(frequencies). Collect data, does it fit the model? (do equation)
21 of 40
Want to find a relationship between two discrete variables. e.g. is there a relationship between gender and owning a dog or cat?
Chi² test of independence/contingency table.
22 of 40
How to carry out a chi² test of independence
Observed table (2x2). Expected table - find grand total, work out marginal totals, E = (Ri x Ci)/N (for each cell) THEN X² = Σ(O-E)² /E
23 of 40
Find differences - One IV, 2+ levels, One DV. e.g. investigate the effect of violent movies on children's aggressiveness - 3 conditions - violent movie, non-violent movie, no movie
One-way between-subject ANOVA
24 of 40
What do you need to find for a one-way between-subjects ANOVA
** Between(g)/Within(err)/Total DF Between (no. of groups -1) Within (sum of df's of each group) Total (one less than total sample size) MS=**/DF. F = MS between/MS within
25 of 40
Find differences - Two+ IVs, One DV - detect interactions
Factorial ANOVA
26 of 40
What is a main effect?
The effect of one IV independent of any other IV
27 of 40
What is a simple effect?
The effect of one level of IV at a level of another IV (effect of gender at different age groups)
28 of 40
What is interaction?
When the effects of one IV differ according to levels of another IV
29 of 40
How do you carry out a factorial ANOVA?
divide data into cells e.g.3 x 2. Find mean for each cell THEN mean for each column/row. Calculate grand mean (sum of all observations/no. of observations). SSIV(s), SScells, SSint, SSerror -- MSIV(s), MSint, MSerror.
30 of 40
How to calculate SSint
SScells - SSIV1 - SSIV2 etc.
31 of 40
How to calculate SSerror
SStotal - SScells
32 of 40
Find differences within subjects - one group of ppts, measured multiple times
Repeated measures ANOVA
33 of 40
How to calculate a repeated measures ANOVA
SStotal, SSgroup(no.of ppts in each group, SSsubjects (no. of conditions), SSerror
34 of 40
DF for repeated measures ANOVA
treatment DF = k-1, error DF = (k-1)(n-1)
35 of 40
Ordinal/nominal data that violates parametric assumptions
nonparametric tests
36 of 40
A nonparametric test to compare central tendencies of two groups
Mann-Whitney
37 of 40
a nonparametric test to test distribution for paired data
Wilcoxon signed ranks.
38 of 40
a nonparametric test for one IV with 3 + levels, DV = ordinal data -
Kruskall Wallace
39 of 40
a nonparametric test with one IV with 3+ levels, DV = ordinal data, repeated measures
Friedman
40 of 40

Other cards in this set

Card 2

Front

compare the mean of an unknown population with a known/specified population mean

Back

one-sample t-test

Card 3

Front

e.g. of a one-sample t-test

Back

Preview of the front of card 3

Card 4

Front

DF for a one-sample t-test

Back

Preview of the front of card 4

Card 5

Front

Compare means of two sets of scores from the same sample/related samples

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all stats resources »