LECTURE 5 Flashcards

1
Q

Parametric stats are used to analyze ____data

A

quantitative

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

true/false: data needs to be normalized for parametric stats

A

true
based on t, F, chi-square distribution

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

-test, ANOVA, Pearson correlation, linear regression
examples of

A

Parametric statistics

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Spearman rho, Mann-Whitney U, Friedman’s ANOVA, Wilcoxon-signed ranks
examples of…

A

non-parametric stats

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

This is our option when we have violated assumptions, or we have nominal or ordinal data.

A

non-parametric stats

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Parametric Assumptions for t-test/one way ANOVA

A
  1. IR data
  2. normality
  3. homogeneity of variance
  4. free of extreme outliers
  5. independence of observations
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

if you knew that the population is normally distributed…even small samples (n<30) will meet this assumption.
In practical terms, as long as your sample is fairly large, outliers are a much more pressing concern than ______

A

normality! (need 30 bc of CLT)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

how do you check for normality?

A

histograms
look at skew/kurtosis (greater than 2 or smaller than -2)
shapiro-wilk test (want it to be non-significant at 0.05-greater than 0.05)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

for skewness and kurtosis, we are looking for

A

Is data more than 2, less than -2?
IF SO, NOT NORMAL

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

How would you test for HOV?

A

Levene’s Test

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is Levene’s Test?

A

Tests if variances in different groups are the same.
Want this test to be not significant. You want them to show “no difference.”
Set alpha at .05.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

In t-test/ANOVA, what will influential outlier do to data?

A

pull mean to outlier

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

For regression tests, what will influential outliers do?

A

pull best fit line towards outliers

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

How can you look for influential outliers?

A
  1. histograms
  2. skewness/kurtosis
  3. boxplots
  4. regression: COOK’S DISTANCE GREATER THAN 1
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

-Data must be Independent
-scores must not follow a pattern over time
-scores from one participant can’t influence another participant’s scores.

What assumption for parametric stats is being met?

A

independence of observations
(scores shouldn’t trend over time
data can’t be from same body parts of same person…)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

regression assumptions

A
  1. linearity
  2. homoscedasticity
  3. outlier testing in regression
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

In correlational/relationship analyses (ex/ regression), the variance of the outcome variable must be about the same at all levels of the predictor variable.
If the variance is not evenly distributed, it’s called having

A

heteroscedasticity

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

what is a residual?

A

distance between observed score and predicted score in linear model

19
Q

An outlier will have a large _____

A

residual

20
Q

to have _____, all of the residuals in a scatterplot should be similar

A

homoscedasticity

21
Q

Whats the goal of a hypothesis test?

A

rule out sampling error (chance)

22
Q

The null hypothesis states there is -_____in the pop before and after intervention

A

NO CHANGE IN POP

23
Q

The alt hypothesis states there is _______ in population following an intervention

A

A CHANGE

24
Q

ALPHA LEVEL DETERMINES RISK OF

A

type 1 error!

25
Q

test statistic t is calculated as

A

difference between pre and post means / amount of diff one would expect without any treatment effect (STAND ERROR)

26
Q

If alpha is 0.05, there is a 5% chance of committing a type ___ error

A

type 1 ERROR

27
Q

actual probability that the results occurred just because of sampling error.

A

p-value

28
Q

when values are in tails, it is

A

significant! p-value is smaller or equal to alpha level

29
Q

HOW TO INCREASE POWER

A
  1. Increase the ES….increase the difference between the groups or decrease the variability.
  2. Increase the sample size
  3. Increase the alpha (.01 to .05)
    OR
  4. Use a 1-tail test
30
Q

independent t-test

A

compares two means, based on independent data
*data from diff groups of people

31
Q

dependent t-test

A

compares 2 means based on related data
-data from same ppl measured diff times (pre/post)
-data from matched samples/twins

32
Q

df is smaller if ___ is not met

A

HOV is not met
Levene’s is smaller than 0.05 = significant

33
Q

when do we use Repeated Measures t-tests?

A

Comparing twins or matched pairs
Studies that measure the same participants twice (ex/ pre/post)

34
Q

You are more likely to find significance with a RM test. Why?

A

Less error when you use the same person and test them twice. However, this isn’t always the case.

This is more likely to be true if the degrees of freedom (number of participants) for the two t-tests are approximately the same… RM will be more powerful.

35
Q

Mann-Whitney U data is based on

A

rankings! look at mean ranking

36
Q

if t (test stat) is in critical region, then difference is ___

A

significant (reject null!)

37
Q

effect size tells you

A

how meaningful the relationship between variables or the difference between groups is. PRACTICAL SIGNIFICANCE

38
Q

small effect size can be significant if sample is

A

very large (hypothesis test is dependent on sample size)

39
Q

effect size
r=0.1, d=0.2
r=0.3, d=0.5
r=0.5, d=0.8

A

small
medium
large

40
Q

r=0.3 explains

A

9% of total variance between two samples (medium effect size)

r=used for correlational data

41
Q

power is defined as

A

probability that stat test will reject the null when treatment does have an effect!

42
Q
A
43
Q
A