Exam Flashcards
3 kinds of error
Sampling error
Measurement error
Standard error
What do the 95% confidence intervals tell us?
95% confident
Mean in population for each condition will fall within values
What is the non-parametric, between participants ANOVA called?
Kruskall Wallis
What is the non-parametric, within participants ANOVA called?
Friedman’s
Non parametric test for:
2 conditions
Between participants
Mann-Whitney
Non parametric test for:
Within participants
Wilcoxon
What is standard deviation?
Standardised measure of deviation between data points and the mean
Why is standard deviation useful?
Tells us about the spread of data in our sample
What is standard error?
Amount of deviation between population and sample
Standard deviation tells us about our ______, whereas standard error tells us about
Sample
Population
What is variance?
Spread of data around the mean
What is a confidence interval?
Range of plausible values for the mean
A confidence interval with a 95% confidence level has a…
95% chance of capturing the population mean
3 types of hypothesis
Null
Alternate
Directional
An uncertain CI would be quite ________ whereas a more precise CI would be…..
Large
Smaller
What are the assumptions for a between participant ANOVA?
1) Continuous DV
2) At least 1 factor with 2 or more levels
3) No significant outliers
4) Normal distribution
5) Homogeneity of variances
What test do we use to test for homogeneity of variances?
Levene’s
Levene’s Test is what we use to test for…
Homogeneity of variances
We test for Levene’s homogeneity of variances in a between/within participants ANOVA
Between
Levene’s test is likeable to…
Mauchly’s sphericity test
ANOVA is generally robust to _____, provided the
Outliers
Sample sizes are equal
Within participants design
Repeated measures - same participants
Mauchly’s test of Sphericity
A P value of 0.01 would indicate…
Assumption of sphericity has been violated
Mauchly’s test of Sphericity
A P value of 0.15 would indicate…
Sphericity of variances
Why do we use the Bonferroni correction?
Eliminate the chance of making a Type 1 error
How is the grand mean worked out?
Working out the mean of each group mean
How is the F statistic calculated for a one-way ANOVA?
Between groups variance/within groups variance
Between groups VARIANCE is…
The difference between groups
Within groups VARIANCE is…
The difference within the groups, between each participants scores
What is one difference in assumptions between the one way ANOVA and the one way related ANOVA?
Levenes vs Mauchly’s
If the results of Mauchly’s test are significant and sphericity is violated, which section of SPSS should be reported?
Greenhouse-Geisser
When reporting main effects, which statistic do we mention?
Marginal means
What do you begin an ANOVA output with?
A … ANOVA was conducted with 1 factor of… and a second factor of…
What is a two way ANOVA?
An ANOVA with 2 or more factors
Partial eta squared provides a measure of…
How much of the total variance is accounted for by the treatment effect
How much of the total variance is accounted for by the treatment effect
^ What is this description referring to?
Partial eta squared
If error is larger than the effect, it is likely that…
The outcome will not be significant
A P value higher than 0.5 indicates a high probability that the effect
Observed is due to chance
Why 0.5 cutoff for significance?
Strikes a balance between making a Type 1 and a Type 2 error
Why do we run post-hoc tests?
Tells us WHERE the differences lie between levels
A two way mixed (split plot) ANOVA has one…
Within subjects variable
One between subjects variable
This type of ANOVA compares mean differences between groups split into two variables - within subjects variable and between subjects variable
Split plot (mixed) ANOVA
In a split plot ANOVA experiment
The effect of one of the two factors on the dependent variable to be…
Dependent on the value of the other (usually within) factors variable
Do blood cholesterol levels change over time naturally?
Or, if one embarks on an exercise training programme of either low, moderate or high intensity?
Which type of ANOVA needs to be run…
Two way mixed ANOVA
Profile plots
If the lines are not parallel or have different patterns…
May have an interaction effect
A profile plot is based on …… data rather than …….
Sample
Population
Why can’t we determine an interaction effect from profile plots?
Based on sample data
We’re interested in the POPULATION
In a two way ANOVA, we are interested in seeing how…
One factor is MODERATED by the other
A researcher was interested in determining whether effect of level of education on political interest was affected by gender.
What type of ANOVA is this?
Two way independent ANOVA
What is a null hypothesis?
The idea that there is no effect between the variables
If Mauchly’s test is above 0.5…
The assumption of sphericity has not been violated
The ANOVA is generally robust to violations of assumptions, as long as…
Group sizes are equal
ANOVA
If the group sizes are unequal/there is no homogeneity of variances, which statistic will be affected?
The F statistic
A P value of less than 0.5 in Levene’s test indicates…
A violation of the assumption
Why do we test at the 0.5 level?
It strikes a balance between making a Type 1 and a Type 2 error
What points would a full ANOVA write up include? (6)
1) Type of ANOVA
2) Main effects
3) Interaction
4) Bonferroni post-hoc tests
5) Interpret effect size
6) Summary statement
What 2 things might you look at to see if distribution of data is normal?
Histogram
Shapiro-Wilk
Profile plots indicate…
Whether there is an interaction in the SAMPLE
Profile plots indicate whether there is an interaction in the ________, not the __________
Sample
Population
A __________ with a 95% level has a 95% chance of capturing the population mean
Confidence interval
As year group increases, so will anxiety scores.
^What kind of hypothesis is this?^
Directional hypothesis
The less confidence interval bars overlap, the more likely…
We are to find an effect in the population
Non overlapping confidence interval bars between groups indicate there may be…
A significant difference
If you were to generate an error bar graph, which would would you use to assess where the mean is likely to fall in the population?
95% confidence interval error bar graph
Error bars for year 3 do not overlap with years 1 and 2, suggesting that there…
Will be a DIFFERENCE
Between year 3 and years 1 & 2
In what order would you report the difference between levels from a T-test?
M, SD, T, df, P
What values do you look at from the T Test?
M, SD, T, df, sig (p)
M, SD, T, df and P are all values you look at from the ….
T-test
In which ANOVA do you report the confidence intervals?
One way
Which values do you look at on an ANOVA table?
df, F, sig
A ONE WAY ANOVA showed that there was a ….
Significant/non significant difference between the levels
A one way between participants ANOVA showed that there was a significant difference in the levels of statistics anxiety for the 3 year groups. What values would you report?
F (df) P
In what order do you report P, F, df in an ANOVA writeup
F (df) P
P value .334
The contrast between (IV 1) and (IV 2) was…
Not significant
In a non-parametric ANOVA, do you report the mean or the median?
Median
In a non-parametric ANOVA, which measure of central tendency is reported?
Median
Mean, median, range etc are all measures of
Central tendency
Which 2 values are reported in place of T and F, in a non parametric ANOVA?
X2
Z
What central tendency value do we report for the MAIN EFFECT of a factor?
Marginal mean
Which (4) values do we report for the MAIN EFFECT of a factor?
F, df, p, marginal mean
In what order (sentence order) would you report F, df, P and the marginal mean?
1) F, df, P (significant main effect)
2) Marginal mean (evidence of higher score)
There was a significant main effect of level of depression on depression scores (np2 = 0.88). This shows that
88% of the variance in depression scores can be attributed to the level of depression
Which ANOVA do we report NP2
Split plot
What does NP2 indicate
The amount of variance in data that can be attributed to that level
We need to run post-hoc tests to examine main effects. True or false?
False
If Mauchly’s in non significant, indicating that there is sphericity, report…
The Sphericity assumed line
If Mauchly’s is significant, indicating there is NOT sphericity, report…
Greenhouse-Geisser
For a mixed ANOVA, why do we split the data in SPSS?
Because it uses both within and between subjects data
How is the F ratio calculated?
Effect/Error
F value tells you if you have…
More effect than error
How is confidence interval worked out using SD and SE?
M +/- (2SD x SE)
How do you work out the T value?
Effect/SE
What is effect?
Mean difference
Do you report marginal means for a 1 way ANOVA?
No
Write up for a One-Way ANOVA
1) 1 way ANOVA showed a significant difference in (DV) for the IVs
2) Bonferroni corrected post hoc tests
3) We can say with 95% certainty…
Write up for a Two Way ANOVA
1) Main effects
2) Interaction
3) Bonferroni corrected post hoc tests
4) These results suggest…(no stats)
Within groups variance will capture…
a) individual differences
b) sampling error
F value less than 1 indicates
More error than effect
A P value of 0.01 you would report as…
P = 0.01
What hypothesis does the ANOVA test?
The null hypothesis that there is no differences between groups
What does an ANOVA NOT tell us?
WHERE the differences lie
This is why we conduct T test
Are post-hoc tests required if the omnibus ANOVA is not significant?
No
You would use a Bonferroni correction if…(2)
Imperative to avoid a Type 1 error
Large number of tests are being carried out
