Extra Questions Flashcards
What is DFFit a measure of?
The difference between the MODEL WITH and WITHOUT the predicted value for a case
What is Cook’s distance a measure of?
The influence of a case
What is Mahalanobis distance a measure of?
The difference between the case and the mean
What is Leverage a measure of?
The effect of the difference between the case and the mean of the IV
What is DFBeta a measure of?
The difference between the SLOPE WITH and WITHOUT the predicted value for a case
What is the Covariance Ratio a measure of?
The influence of the case on the residuals (error) and variance
How do you calculate Rsq?
1 - SSW / SST
What are the degrees of freedom associated with the Corrected Model?
k
What does per comparison error rate effect?
The probability of a Type I error for a PARTICULAR comparison.
What does Family Wise Error rate effect?
The probability of making one or more Type I errors in a family or SET OF comparisons.
In ANCOVA, what is the problem attached to the situation where the covariate has a significant relationship with the independent variable?
Reduced power
Reduced SS for main effect
What is epsilon a measure of?
Sphericity
Does the within participant variation appear in the SPSS output?
No
The correlation between the obtained distribution of scores and what the scores should be if they are distributed normally is known as the?
Shapiro Wilk
The procedure that estimates the sampling distribution by estimating the properties of sampling distribution from sample data is known as?
Bootstrapping
The military titles of Lieutenant, Captain and Major represent which ONE of the following scales of measurement?
Ordinal
A relationship study between HSC examination results (Variable X) and how well students perform in their first year of study at ACAP (Variable Y) had some students drop out. You presume that the data are missing completely at random because?
Students who obtained a range of HSC results also obtained a range of grades in their first year at ACAP
Which ONE of the following strategies is the most preferred way of dealing with missing data?
Expectation Maximisation
What type of distribution is the F statistic?
A distribution of the ratio of two standard errors
In a regression model, the error of prediction refers to?
The difference between the value of the DV and the predicted value for the DV
A multiple regression analysis has yielded a Variance Inflation Factor (VIF) of 36. This value tells us that?
The SE (standard error) has been expanded 6 times in comparison with a situation where there is no correlation between our variables
If the distribution of one variable has a moderate negative skew (z > 2.58). What is the most appropriate transformation to use?
Reflex the scores, then use a logarithmic transformation
Some general ethical principles include:
Protection from physical and psychological harm.
The right to self-determination.
Respect for individual dignity
What are the df for the Shapiro Wilk Test?
N
What is the appropriate class of statistics to deal with research questions concerning relationships between or among variables?
Correlation and regression
In what type of research design does a researcher control and/or manipulate the environment to observe the effects on behaviour?
An experiment
In a normal distribution, what percentage of scores can have the absolute value of z exceeding 2.58?
1%
A point biserial correlation is a measure of association between two variables that are?
Interval and dichotomous
Two variables have a correlation of .4 (r=.4). The amount of variance accounted for by the predictor in the criterion is?
16% (moderate effect)
The strongest type of data transformation is?
Inverse
In linear regression analysis, how is the R squared is calculated?
SSM / SST
In what way is a multiple linear regression different from a factorial analysis of variance?
The predictors must be unrelated to each other in factorial ANOVA, but may be related to each other in multiple regression analysis.
What would the tolerance value be for a correlation of .81 and what would it mean?
.19
Possible issues with multicollinearity among the predictor variables
What is the problem associated with the forward entry of variables into a stepwise multiple linear regression model?
A forward entry increases the likelihood of missing a variable that has a significant effect on the criterion variable.
How do you calculate t?
Mean Diff / SE
The mathematical assumption of homoscedasticity in a linear regression model is tested by?
The plot of regression standardised predicted values against regression studentised residuals
How is the assumption of linearity tested in an analysis of covariance?
Pearson r between DV and CV
In regression analysis, the covariance ratio of an influential case is?
A measure of of whether a case influences the variance of the REGRESSION PARAMETERS
Df of the Total Model =
N
Which path from the path diagram determines whether the sobel test is sig.
a*b
What is the equation for calculating r from a contrast?
Sq root of
tsq / tsq + df
df = N - k
If in an ANCOVA, Age regressed on Narcissism isn’t sig. but Sex is, what does this mean?
The slopes of Narcissism on Age for males and females are the same
What should you use if homogeneity of variance is not met, in a one way between subjects ANCOVA and you want to explore group differences because of a sig. finding?
Not a post-hoc test (e.g. not Games-Howell because ANCOVA)
Use pairwise comparisons using means adjusted for the effect of the covariate with a Bonferroni correction
In effect sizes of mediation, the ratio of the indirect effect to biggest size it could be is known as the
Preacher Kelly Kappa
A one way analysis of variance has yielded an F statistic that is less than 1. What does this mean?
here is a greater amount of variance not explained by the model than what is explained
A spurious correlation occurs when?
The path b= 0, and
2 endogenous variables (Y and M) are jointly DEPENDENT on a prior variable
Another way of saying the df for F is?
(k, error)
k = number of predictors accounted for in that line of the output (e.g. only one if controlling for others)
The assumption of sphericity in a repeated measures analysis of variance assesses?
If variances of the differences between treatment levels are equal
The dfs in the analysis of variance associated with the Covariate term in the model will be?
number of CVs
What does it mean if the DfBeta is larger than 2SDs?
The case substantially influences the model’s parameters.
If the unstandardised coefficients are negative what does this mean?
As the IV increases, the DV decreases by x points.
In a regression model, the error of prediction refers to?
The difference between the value of the dependent variable and the value predicted for the dependent variable
What are the 3 factors known to affect the magnitude of Pearson r?
Size of the sample
Restriction of the range
Heterogeneous subsamples
How does ordinary least squares regression work?
Residuals for each possible regression line are squared
Squares are summed
The regression line with the smallest SS is best fit
In the discipline of Psychology, data measured using a Likert-type scale are treated traditionally as?
Interval scale
A multiple regression analysis has yielded a Variance Inflation Factor (VIF) of 16. This value tells us that?
The SE (standard error) has been expanded 4 times in comparison with a situation where there is no correlation between our variables
In _________________ linear regression, all predictor variables entered at once into the model, in order to _______________.
Stepwise; Build a preliminary model based on chance.
The situation in which a predictor variable, X, has a significant effect on a criterion variable, Y, when a third variable, Z, is held constant, is known as a?
Suppressor effect
What does EM do?
Determines the optimal amount of trimming required
Which is the least recommended way of dealing with missing data?
Mean substitution
The procedure that estimates the sampling distribution by estimating the properties of sampling distribution from sample data is known as
Bootstrapping
Shrinkage found when Rsq to Rsq adj represents?
Loss of predictive power of the model in the population
When asked “at what error level should the tests of main effects and simple effects be compared” what is this referring to?
one-tailed or two-tailed at p
If the hypotheses for the main effects are directional how should the results be reported?
one-tailed
and
at the conventional error rate of p
Should post-hoc comparisons be one-tailed or two?
2-tailed
What can be used to control for error rate in post-hoc tests?
The different types of tests depending on equality of group sizes and variances
When using pairwise comparisons what should you use to control for error rate?
Bonferroni adjustment
When using simple effects to check a sig. interaction effect, what should you use to control for error rate?
Bonferroni adjustment
