ANOVA quiz points Flashcards
do you ever directly compare whether IV 1 is different to IV2?
No, in two way anova you measure effects on the DV, no whether the IVS are sig different from eachother.
Why do planned contrasts have more power than omnibus anova?
No contrast can ever have more than SSB. However even though ss contrast
what is the maximal comparison?
the difference between the largest and smallest means
If this is not significant, no other pairewise comparison will be
in the special___ do you just put the contrast coefficients?
No, Your coefficients address the questions the researcher is interested in, but SPSS needs you to add something to this as part of the “special” within-subjects contrast command.
is a 2 x (3) one that took dependent variable measures at two different times on three groups of participants?
Incorrect. This is describing a design in which there were two groups of participants measured at three different times.
Consider this mixed design: 2 x (3) x (4) How many times was each person tested?
12
is this normalised? 0.816 0 -0.408 -0.408
correct; the sum of the squared coefficients is 1.
2 rules of normalisation
Normalised coefficients have Σc2 = 1
SS for a (J-1) set of orthogonal contrasts will be additive when rescaled so that each contrast has Σc2 = 1
how do you work out MSE for contrasts if they are ommitted from the table in spss?
you can work out the MSE for each contrast by dividing each MS contrast by F
whats this mean (2 x 3)
this is describing a design in which there is just one group of participants who are measured under 6 different conditions, defined by the combinations of two within-subjects variables, one with 2 levels and the other with 3 levels.
criterion of significance was set at 0.05, the probability of incorrectly rejecting H0 is…
. Since the criterion of significance was set at 0.05, the probability of incorrectly rejecting H0 (i.e., making a Type I error) is 0.05. 0.95 is the probability of correctly accepting H0. The probability of correctly rejecting H0 is the power of the test
does a larger critical F means a higher Type I error rate
This is not correct. A larger critical F makes it harder to reject the null hypothesis, and therefore less likely that we would make a Type I error.
whats v2 and n for contrasts?
ν2 = N-J and n = N/J
when are we sure k = ≤ j-1?
k ≤ J - 1 (as is always the case with trend analysis)
does larger critical F mean a lower Type I error rate
This is not correct. A larger critical F makes it harder to reject the null hypothesis, which corresponds to lower power.
is maximal contrast always sig?
although the maximal pairwise comparison is not significant, the maximal contrast may be.\
The maximal contrast is equivalent to the overall ANOVA, and so accounts for all of the SSB (i.e., = SSB). The coefficients of the maximal contrast are represented by the deviation of each group mean from the grand mean, and so are generally not interpretable (although it is conceivable that they could be).
Df for two way
df(A) = (J-1) df(B) = (K-1) df(AB) = (J-1)(K-1) df(Error) = JK(n-1) df(Total) = N-1
df for one way
dfB= J – 1 dfW = J(n – 1) = N – J dfT = N – 1
df for contrasts
𝜈1 for a contrast is always 1
v2 is dfw from anova = J(n – 1) = N – J
df for mixed designs
dfW for contrasts is the same as mixed designs as it is for within subjects designs. But you can read it straight out of the output.
Using the Bonferroni procedure means
The probablitiy of making at least one type 1 error is inflated above .05.
UNIANOVA meanT BY therapy
/contrast(therapy)=special(1 1 -2) /contrast(therapy)=special(1 -1 0) /METHOD=SSTYPE(3) /INTERCEPT=INCLUDE /PRINT=TEST(LMATRIX) /CRITERIA=ALPHA(.05) /DESIGN=therapy.
Is this between or within?
between
GLM alc mj ctrl /WSFACTOR=substance 3 special(1 1 1, .5 .5 -1, 1 -1 0) /METHOD=SSTYPE(3) /PLOT=PROFILE(substance) /EMMEANS=TABLES(substance) /PRINT=DESCRIPTIVE TEST(MMATRIX) /CRITERIA=ALPHA(.05) /WSDESIGN=substance.
is this between or within?
within
