Stats Flashcards Preview

EPPP - select domains > Stats > Flashcards

Flashcards in Stats Deck (31):
1

Nominal 1 IV - 2+ indep groups

Chi square

2

>1 nom IV, 2+ indep grps per

Multiple sample chi-square

3

1+ nom IV, 2+ CORR grps per

McNemar

4

1 IV & 1 DV - Interval or ratio data:

1 indep group

2 indep group

1 - t test single sample

2 - t test independent samples

5

1 IV & 1 DV - Ordinal data

1 indep group -

2 indep group -

1 - kolmogorov (single vodka)

2 - Mann-Whitney (indep man) or median test or Kolmogorov-smirnov (double vodka)

6

1 IV & 1 DV - interval/ratio

2 correlated groups

T test matched samples

7

1 IV & 1 DV - Ordinal Data

2 correlated groups

Wilcoxon (sign rank) or sign test

(Ordinal oxen matches rank)

8

1 IV & 1 DV - interval/ratio

>2 correlated groups

One way repeated meas Anova

9

1 IV & 1 DV - Ordinal

>2 correlated groups

Friedman

10

1 IV & 1 DV - interval/ratio

>2 indep groups

One way ANOVA

11

1 IV & 1 DV - Ordinal

>2 indep groups

Kruskall Wallis

12

2 IV
2+ groups per IV

All independent

2 way ANOVA

13

2 IV
2+ groups per IV

One IV w corr groups,
Other w indep groups

Split plot ANOVA or
mixed ANOVA

14

2 IV
2+ groups per IV

Both IVs w corr groups

Repeated measures factorial ANOVA

15

2 IV
2+ groups per IV

One IV blocked,
indep groups

Randomized block ANOVA

16

2 IV
2+ groups per IV

W covariate, indep/corr groups

Ancova

17

>1 DV
2+ groups per IV
Indep and corr groups

Manova - multivariate ANOVA

18

The coefficient of determination

% of variability in Y shared/accounted for by X.
Generated by squaring the correlation coefficient.

19

Line of best fit

pred line through scatter plot; made from least squares criterion. Y = a + bX (performance = y-intercept + slope(yrsEdu)

20

Assumptions of bivariate correlations

a. linear relationship b. homoscedasticity (consistent spread) c. unrestricted range

21

Correlation co-eff: X, Y = int/ratio

Pearson r

22

Correlation co-eff: X, Y = ordinal

spearman's rho / Kendall's Tau

23

Correlation co-eff: X = int/ratio, Y = true dichotomy

point-biserial

24

Correlation co-eff: X = int/ratio, Y = artificial dichotomy

Biserial

25

Correlation co-eff: X, Y = true dichotomy

Phi

26

Correlation co-eff: X, Y = artificial dichotomies

Tetrachoric

27

Correlation co-eff: XY = curviliniar relationship

Eta

28

Slope and beta values

"b" values can be standardized partial slopes; "b" used when comparing variables btw studies or samples, but beta only compares within a study/sample.

29

Discriminant function analysis

Used when Y is nominal (not interv/ratio; allows prediction of membership based on kwg of a set of predictor variables: likelihood of passing EPPP based on time spent studying and practice tests completed.

30

Logit

Loglinear analysis: predict category based on categorical var(s). [b/g pred pass/fail]

31

Standard error of measurement

SD x sqrt(1-reliability) e.g., r=0.84; 1-0.84 = 0.16; sqrt0.16 = 0.4; If SD is 15, then 15x0.4 = 6