test z-test single-t ind-1 chi-square

when to use: - 1 sample mean - no pop. SD - compare 2 sample means - nominal IV

- pop. parameters - compare 1 sample mean to - no pop. parameters - frequency data
- pop. SD another value

variables: pop parameters - 𝝁 = other value - M = sample mean - N = total # of responses

- 𝝁 = pop. mean - M = sample mean - k = # of columns, cells,
- 𝝈 = pop. SD - there is SD choices
- M = mean - fo = freq. obs.
- SD = standard dev. - fe = freq. exp.
(# of pp. by chance)

formula: 𝑀−µ 𝑀−µ 𝑀1−𝑀2 fe = (# row) + (# col) ÷

𝑧 = 𝑆𝐸𝑚
𝑡= 𝑒𝑠𝑡. 𝑆𝐸𝑚
𝑡= 𝑒𝑠𝑡. 𝑆𝐸𝑑𝑖𝑓𝑓 total #
σ 𝑆𝐷 2
2
𝑆𝐸𝑚 = 𝑒𝑠𝑡. 𝑆𝐸𝑀 = 𝑆𝐷𝑎 2 𝑆𝐷𝑏 2
𝑥 =
(𝑓𝑜−𝑓𝑒)
𝑁 𝑁−1 𝑆𝐸𝑑 = ( ) +( ) 𝑓𝑒
𝑁𝑎−1 𝑁𝑏−1

df n/a 𝑑𝑓 = 𝑛 − 1 𝑑𝑓 = (𝑁𝑎 + 𝑁𝑏) − 2 𝑑𝑓 = 𝑘 − 1

effect size (𝑀−µ) 𝑡 𝑑 = 𝑡( 𝑁𝑎 + 𝑁𝑏/(𝑁𝑎 * 𝑁𝑏)

𝑑= σ
𝑑= 𝑥
2
𝑁 𝑑= 2
s: 0.2 𝑁+𝑥
m: 0.5
l: 0.8

conf. inter. C.I = M ± Zcrit. (SEm) C.I = M ± Tcrit. (Est. SEm) C.I = (Ma-Mb) ± Tcrit. (Est. SEdiff) n/a

steps 1. calc SEm 1. calc est. SEm 1. calc SEdiff 1. draw contingency table
2. calc z 2. calc t 2. calc t 2. calc fe for each cell
3. compare z and z critical 3. calc df 3. calc df 2
3. calc 𝑥 for each cell
0.05 or 95% +- 1.96 4. compare t and t critical 4. compare t and t critical and add everything
0.01 or 99% +- 2.58 5. calc effect size (d) (if t beats 5. calc effect size (if t beats critical) 4. calc df
4. calc effect size (if z beats critical) 2 2
critical) 5. compare 𝑥 and 𝑥
critical
6. calc effect size

keywords - population SD - random sample - is there a sig. diff. between ____ and - yes or no outcome
- parameter mean _____ - counts not scores
- compared on continuous outcome

concepts - difference between M and 𝝁 - difference between M and 𝝁 or - difference between 2 sample means - differences between
when you have 𝝈 between M and another value observed and expected
frequencies
 test pearson’s r fisher’s test paired t - test 1 - way ANOVA factorial ANOVA

when to if you are given a zx compare two zr values compare two means and standard deviation (r given ssb and ssw given multiple ssb’s and
use: and zy chart is given) (1x1 table) one ssw (2x2 table)

variables: n = sample zr = critical value of r m = mean b = between b = between

s1 = est. sem 1 w = within w = within
s2 = est. sem 2 levels = cells levels = cells under that iv

formula: Σ (𝑍𝑥)(𝑍𝑦) 𝑍𝑟1−𝑍𝑟2 𝑀1−𝑀2 𝑚𝑠𝑏 𝑚𝑠𝑏

𝑟= 𝑧= 𝑡= 𝑓 = 𝑓 =
𝑛 1 1 𝑆𝐷 2 𝑆𝐷 2 𝑚𝑠𝑤 𝑚𝑠𝑤
+ ( ) +( ) −2(𝑟)( 𝑠1) (𝑠2)
𝑁1−3 𝑁2−3 𝑁1−1 𝑁2−1 𝑠𝑠𝑏 𝑠𝑠𝑏
𝑚𝑠𝑏 = 𝑑𝑓𝑏 𝑚𝑠𝑏 = 𝑑𝑓𝑏
𝑆𝐷 𝑠𝑠𝑤 𝑠𝑠𝑤
𝑒𝑠𝑡. 𝑆𝐸𝑀 = 𝑚𝑠𝑤 = 𝑑𝑓𝑤 𝑚𝑠𝑤 = 𝑑𝑓𝑤
𝑁−1

df 𝑛 − 2 n/a 𝑛 − 1 dfb = # of levels -1 dfw = n-# of levels

dfw = n - # of levels dfb iv = # of levels -1
dfb interaction = dfb1
(dfb2)

effect size n/a n/a 𝑡 𝑜𝑏𝑡𝑎𝑖𝑛𝑒𝑑 2 𝑠𝑠𝑏 2 𝑠𝑠𝑏

η = 𝑠𝑠𝑏 + 𝑠𝑠𝑤
η = 𝑠𝑠𝑏 + 𝑠𝑠𝑤
𝑁

table C D B E E

steps 1. multiply zx and 1. use the table to find 1. plug all the given values into the formula 1. find dfb 1. find msw
zy the critical values of r and solve 2. find dfw (denominator for all
2. add up the values (zr) 2. find df 3. find msb f-statistics)
3. divide the value 2. plug all the numbers 3. find the critical value using df 4. find msw 2. find msb (find for each
by n (sample) to get into the formula and 4. compare obtained value to t critical 5. find f obtained iv + interaction)
r solve 5. if r did not beat the critical, use 0 for r 5. find f critical 3. find f-statistics
4. find df 3. find z critical at 0.05 (dfb, dfw) 4. find critical value
5. find the critical or 0.01 6. compare obtained [f(dfb, dfw)]
value using df 0.05 or 95% +- 1.96 value to f critical 5. find effect size if
6. compare obtained 0.01 or 99% +- 2.58 7. use hsd table if applicable
value to r critical 4. compare obtained you reject two IV 6. find hsd if applicable
value to z critical levels 7. explain results

other interaction only: there is

probably no difference
ebtween ____ and ____ or
____ and ____. the effect
of ____ and ____ depends
on both

concepts reject: we can use reject: one group has a reject: there is a correlation between both reject: they are not reject: there was a
the r value stronger relationship values equal (there is a difference between the
accept: no than the other accept: no correlation between both values difference) values
correlation between accept: two correlations accept: all are equal accept: no difference
the two variables are not meaningfully (Ma = Mb = Mc) between the values
different
(compare each value to
each other)