Upgrade to remove ads
PSYCO 532 Final Vocab
Terms in this set (54)
design consists of two or more factors, each with discrete possible values or "levels", and who experimental units take on all possible combinations of these levels across all such factors; efficient but can be expensive due to the need of a large amount of participants
simple main effect
When there is an interaction, this is the difference in means at a particular level of a factor in the design; effect of one independent variable within one level of a second independent variable.
can arise when considering the relationship between 3 or more variables, and describes a situation in which the simultaneous influence of two variables on a third is not additive; most often considered in the context of regression analyses.
statistical model that represents the observed quantities in terms of explanatory variables that are treated as if the quantities were non-random; effect will be the same if you do the experiment over and over again
assumes that the data being analysed are drawn from a hierarchy of different populations whose differences relate to that hierarchy; comparing differences between a large population
completely crossed & balanced design
if n's are equal and same number of rows and columns (i.e. every level of B occurs at each level of A)
a linear combination of variables (parameters or statistics) whose coefficients add up to zero, allowing comparison of different treatments.
square matrix whose determinant is not equal to zero.
square matrix whose determinant is equal to zero.
Using a qualitative regression to test the null that the beta coefficeints are equal to zero to find a difference betweem the means of the treatment. You use p-1 variables in the regression along with the Bnot to remove the effect of linear dependency. Each treatment level has 1's for that treatment in the matrix and zero otherwise. For the Pth treatment level you only have zeroes .
For the Pth treatment level you only have -1's in the data matrix.
a measure of the degree to which a null hypothesis is false; tells you something about the statistical power of a test; an F-distribution with a
of zero means that the F-distribution is a central F-distribution; measured in standard deviations; how far the actual mean is from the null mean
the probability that a test of statistical significnance will reject a false null hypothesis; inversely related to beta (probability of making a Type I error).
intraclass correlation coefficient
measure of the reliability of measurements or ratings; what you use for power of randomized treatment levels
randomized block design
experimenter divides subjects into subgroups, such that the variability within these subgroups is less than the variability between the subgroups. Then, subjects within each subgroup are randomly assigned to treatments conditions; reduce noise of nuisance variable by blocking on it
data model in which the effects of individual factors are differentiated and added together to model the data; model has only main effects; when you have predictors that are main effects
model with interaction term as well as main effects
repeated measures design
uses the same subjects with every branch of research, including the control
Tukey's test for non-additivity
approach used in two-way ANOVA to assess whether the factor variables are additively related to the expected value of the response variable; can be applied when there are no replicated values in the data set.
refers to the condition where the variances of the differences between all possible pairs of within-subject conditions (i.e., levels of the independent variable) are equal; condition where the variances of the differences between all combinations of related groups (levels) are equal; an important assumption of a repeated measures ANOVA.
when sphericity is not met, use epsilon-correction for F-test; reduces power
procedure that estimates epsilon in order to correct the degrees of freedom of the F-distribution; tends to underestimate epsilon when epsilon is close to 1 (i.e., it is a conservative correction).
estimagtes epsilon in order to correct the degrees of freedom of the F-distribution; tends to overestimate epsilon (i.e., it is a more liberal correction).
uses 1 and n-1 df (step 2 in sequential)
most commonly used in two cases: 1) comparing the same dependent variables between groups over several time-points, and 2) when there are several measures of the same dependent variable; uses plots of the data to visually compare across groups; multivariate equivalent of repeated measures or mixed ANOVA
in profile analysis, means do not differ across time within the group
In profile analysis,testing if the overall means are different between the groups. Testing if there is seperation between the groups.
design that contains both fixed and random effects.
design involving assigment of levels of one factor to main plots in a CRD, RBD(random block design), or a Latin-Square and then assigning the levels of a second factor to subplots within each main plot.
compares all possible pairs of means; compares the means of every treatment to the means of every other treatment; applies simultaneously to the set of all pairwise comparisons.
Latin square design
a square array which contains n different elements with each element occurring n times but with no element occurring twice in the same column or row; emphasis is that treatments only appear once in a column or row; often the blocking factors are time
repeated measures design in which each experimental unit (subject) receives different treatments during the different time periods; important feature: time variable/effect are the column means, ignoring pattern of treatments; (i.e. subjects "cross-over" from one treatment to another during the course of the trial.
way of reducing your SSerror; controls for nuisance variables; subjects are often blocked; if a blocking effect, SSresid< SSerror.
design in which every level of a given factor appears with only a single level of any other factor; model has no interaction
levels of one factor appear at levels of another factor; if every level of one appears with every level of others, factors are said to be completely crossed; if not, they are partly crossed.
analysis of covariance (ANCOVA)
used to test the main and interaction effects of categorical variables on a continuous dependent variable, controlling for the effects of selected other continuous variables, which co-vary with the dependent. The control variables are called the "covariates."
report after you complete ANOCVA to remove effect of covariate
Two contrasts are
if the sum of the products of corresponding coefficients (i.e. coefficients for the same means) adds to zero.
multivariate analysis of variance (MANOVA)
an ANOVA with several dependent variables; tests for the difference in two or more vectors of means.
strategy for finding the local maxima and minima of a function subject to equality constraints; can do a statistical test of a simple null hypothesis that a parameter of interest is equal to some particular value.
Roy's Greatest Characteristic Root Criterion
used in MANOVA to measure statistical significance; get critical value from Heck's Charts using dfs of s, m, and n. Multivariate F-test; ideal to use when one eigenvalue is much larger than the other eigenvalues.
Wilk's lambda distribution
a probability distribution used in multivariate hypothesis testing, especially with regard to the likelihood-ratio test and multivariate analysis of variance (MANOVA); ideal to use when eigenvalues are close to uniform
used as a test statistic in MANOVA and MANCOVA. This is a positive valued statistic ranging from 0 to 1. Increasing values means that effects are contributing more to the model; you should reject the null hypothesis for large values.
a function of several variates used to assign items into one of two or more groups. The function for a particular set of items is obtained from measurements of the variates of items that belong to a known group.
independent and identically distributed (i.i.d.)
a sequence or other collection of random variables is
if each random variable has the same probability distribution as the others and all are mutually independent.
in Latin squares, if the first row and column are in a logical order sequence (1,2,3,4...)
self-conjugating Latin square
when treatment sequence in the rows matches treatment sequence in columns; if you transpose, it is the same as the original
general linear model approach to ANOVA
procedure in which the calculations are performed using a least squares regression approach to describe the statistical relationship between one or more predictors and a continuous response variable.
replication of the experiment; if all cells have one observation, you have one of
; repetition of an experimental condition so that the variability associated with the phenomenon can be estimated; the repetition of the set of all the treatment combinations to be compared in an experiment.
a random variable that is fundamental to the probabilistic model, but that is of no particular interest in itself or is no longer of interest; often dealt through blocking; controlling for influence on the dependent variable
coding used to compute contrasts.
YOU MIGHT ALSO LIKE...
331 unit 2 uwl
Exam 4 stats
PSYC 305: Statics for Experimental Design
OTHER SETS BY THIS CREATOR
Candidacy - Methodology
Psyco 532 Midterm Vocab
Psyco 531 In-class Vocab