Stats and Research Design

Question 1

Factorial Design

Answer 1

Research designs that have 2+ IVs. Can analyze main and interaction effects. Main= effect of one IV on the DV. Interaction= effects of one IV on different levels of another IV.

Question 2

Shared Variability

Answer 2

Correlation coefficient for 2+ variables can be squared to obtain a measure of shared variability. So, if correlation between X and Y is .5, 25% of variability in Y is shared with or accounted for by X.

Question 3

Scales of Measurement

Answer 3

Nominal: name! have a number of groups and frequency count in each group (marital status)

Ordinal: also named categories but have a natural sequence or order (like income, likert). differences between values are not consistent

Interval: often continuous values (dates on a calendar) but no true zero

Ratio: can be continuous or discrete, have a true zero (weight)

Question 4

Standard Deviation

Answer 4

measure of variability of scores around the mean of a distribution.

Question 5

Regression Analysis

Answer 5

Helps predict a score on one criterion based on the person’s score on a predictor. Line of best fit helps make predictions.

Question 6

Statistical Power

Answer 6

Probability of rejecting a false null hypothesis. Can’t be controlled directly but is increased by large sample, large alpha, reducing error, and maximizing effects of IV

Question 7

Factorial ANOVA

Answer 7

2+ IVs and a single DV, interval and ratio scales.

Question 8

Randomized Block ANOVA

Answer 8

Use when blocking is done to control for an extraneous variable.

Question 9

Experimentwise Error Rate

Answer 9

Probability of making Type 1 error. As number of comparisons in a study increases, error rate goes up.

Question 10

MANOVA

Answer 10

1+ IVS and 2+ DVs. Use of this helps reduce experimentwise error and increases power.

Question 11

F-Ratio

Answer 11

Indicator from ANOVA of significant difference between means. A large effect is larger than 1.0

Question 12

Chi-Square Test

Answer 12

Nonparametric test used with nominal data. Single sample has one variable. Count IV and DVs together.

Question 13

Threats to External Validity

Answer 13

External Validity = generalizability

Pretest sensitization–pretest affects reaction to treatment
Reactivity–respond differently b/c they know they’re in a study
Multiple treatment interference–when they receive more than one level of IV

Question 14

Trend analysis

Answer 14

Type of ANOVA used to assess linear and nonlinear trends when IV is quantitative.

Question 15

Discriminant Function Analysis

Answer 15

Analysis used when 2+ predictors will be used to predict one’s status on a single discrete (nominal) criterion.

Question 16

Effect Size

Answer 16

Measures the magnitude of the relationship between IV and DV. Helpful for establishing clinical or practical significance.

Cohen’s d and eta squared.

Question 17

Threats to Internal Validity

Answer 17

Internal Validity= degree to which a research study allows investigator to conclude that the observed variability is due to the IV.
Maturation, History, stat regression, selection (groups are diff at start)

Question 18

Multiple Regression

Answer 18

Used to predict a score on a continuous criterion based on performance on 2+ continuous or discrete predictors.

Question 19

Multicollinearity

Answer 19

Issue in multiple regression when there are high correlations between predictors (don’t want this!)

Question 20

Skewed distributions

Answer 20

Positively skewed: most scores on low side, with a few scores on the high side and mean>median>mode

Negatively skewed: most scores on high side, with a few on low side mode>median>mean

Question 21

ANCOVA

Answer 21

Used to increase efficiency of analysis. Removes variability in DV due to extraneous variable. DV score is adjusted for each person.