Adv Correlat Methods Quiz 2

Front 1

These types of variables are independent of each other and are not associated with one another (r=0).

Back 1

orthogonal variables

Front 2

If two variables are orthogonal, knowing the value of one variable provides what information about the value of the other variable?

Back 2

If two variables are orthogonal, knowing the value of one variable gives no info about the value of the other.

Front 3

Factoral designs create what kind of variables?

Back 3

orthogonal

Front 4

ANOVA requires what type of variables?

Back 4

orthogonal IVs

Front 5

What advantages are associated with orthogonality of variables?

Back 5

each IV adds to the prediction of DV in a simple fashion (but variables that influence a DV are often correlated).

Front 6

ANOVA requires _____ IVs, but regression can handle _____ IVs.

Back 6

ANOVA requires orthogonal IVs, but regression can handle nonorthogonal IVs.

Front 7

This analytic technique can examine unique and overlapping variance accounted for in the DV.

Back 7

Regression

Front 8

The assignment of numerals to objects or events according to rules

Back 8

measurement

Front 9

Distinguish between continuous and discrete/dichotomous types of measurement.

Back 9

Continuous measures are quantitative (e.g. age, time). Continuous measures can take on any value within the range of the scale. The size of the number reflects the amount of the variable.

Dichotomous/discrete measurements are nominal, categorical, qualitative (e.g. sex, religion).

Front 10

What are the four scales of measurement?

Back 10

nominal, ordinal, interval, ratio

Front 11

This scale represents the lowest level of measurement. Mutually exclusive categories are present, but numbers are arbitrary and act only as labels. There is no difference in magnitude among the numbers, so it is relatively meaningless to add, average, etc.

Back 11

Nominal scale

Front 12

Examples of this type of scale include: sex of participant, political party affiliation.

Back 12

nominal scale

Front 13

In this scale, there are mutually exclusive categories. The numbers reflect order, but the differences between adjacent numbers are not equal across the scale. This scale does not have an absolute zero.

Back 13

Ordinal scale

Front 14

Examples of this type of scale include: birth order, relative finish in a race

Back 14

ordinal scale

Front 15

In this scale, mutually exlusive categories exist and the numbers are ordered. There are equal distances between adjacent numbers, but there is no absolute zero (complete absence of variable). Negative numbers are possible.

Back 15

Interval scale

Front 16

Examples of this type of scale include: temperatures in F or C, psych scales.

Back 16

interval scale

Front 17

In this scale, you can''t compute ratios, but you can add a constant.

Back 17

interval scale

Front 18

negative numbers are possible with this scale.

Back 18

interval scale

Front 19

In this scale, there are mutually exclusive categories that are numerically ordered. There are equal distances between adjacent numbers and a true zero that indicates the absence of the measured characteristic. Also, in this scale, the arithmatic ratio of scores is meaningful.

Back 19

ratio

Front 20

Examples of this type of scale include: height (e.g., km, mi), weight (e.g. kg, lb).

Back 20

Ratio

Front 21

Why is random sampling important in nonexperimental research?

Back 21

Random sampling is important if you want results from your sample to be reflective of a larger population. You need to get a representative sample from a predefined population so that you can generalize to the population.

Front 22

Why is random assignment not as important in experimental research?

Back 22

In experimental research you take a homogenous group, randomly assign treatment to groups, and see if the groups seem to come from different populations after your treatment.

Front 23

What is important to keep in mind about generalization?

Back 23

You want to generalize to the type of people who participated. If (instead) you want to generalize about a theory you ask, what can happen?

Front 24

What are some important things to keep in mind when considering which variables to include in your analysis?

Back 24

1. best soln with fewest variables
2. reliability, cost, availability, meaning, validity, theory, correlations among variables

Front 25

An index of the strength of an association between the iv and dv.

Back 25

effect size

Front 26

What is the difference between "statistical significance" and "effect size."

Back 26

Statistical significance tells you whether or not an effect exist, but effect sizes tell you how much of an effect there is (i.e., the size of the effect).

Front 27

Is F a good index of effect size? Why or why not?

Back 27

F is not a good index of effect size, because it is influenced by sample size.

Front 28

Are eta squared and omega squared influenced by sample size?

Back 28

no

Front 29

Indicate the proportion of variance in dv that can be attributed to variation in IV.

Back 29

Eta squared and omega squared

(Eta squared and omega squared are both measures of effect size. Eta squared is the proportion of the total variance that. is attributed to an effect).

Front 30

Eta squared and omega squared range from zero to 1. What does it mean to have an eta or omega squared of zero? one?

Back 30

An eta/omega squared of zero indicates that the variables are unrelated/orthogonal.

An eta/omega squared of 1 indicates that the variables are related.

Front 31

Distinguish between small, medium, and large effect sizes.

Back 31

small = .01
medium = .06
large = .15

Front 32

Squared multiple correlation coeficient

Back 32

R squared

Front 33

What is power?

Back 33

The ability to detect an effect if one is present.

Front 34

Probability of rejecting the null hypothesis when it is false

Probability that true effects will produce statistical significance in your data analysis.

Back 34

Power

Front 35

Is equal to 1-B

Back 35

Power

Front 36

Beta refers to the likelihood of making what kind of error?

Back 36

Beta refers to the likelihood of making a type II error. A type II error involves assuming that there is no effect when their really is.

Front 37

What four things does power depend upon?

Back 37

1. size of treatment effects
2. degree of error variance
3. significance level
4. Sample size

Front 38

The greater the effect size, the more likely you are to detect an effect. This has what effect on power.

Back 38

Greater effect size = greater power.

Front 39

The ___ the error variance the ____ the power.

Back 39

The smaller the error variance the greater the power.

Front 40

Which alpha level ( .05 or .10 ) will give us greater power? Why?

Back 40

The .10 alpha will give us greater power because we are being less stringent.

Front 41

The ___ the sample size, the ___ the power.

Back 41

The larger the sample size, the greater the power.

Front 42

What are some strategies for increasing power?

Back 42

1. increase size of tx effects
2. decrease error variance by influencing study design or reliability
3. use a less stringent alpha level
4. increase your sample size.

Front 43

A more reliable measure will have less error variance. What will this do to power?

Back 43

increase power

Front 44

How do we estimate effect size for a power analysis?

Back 44

we make a realistic guess based on pilot studies and previously published research. (You can also specify relative size of tx effects?)

Front 45

What term do we use to specify the relative size of tx effects?

Back 45

Cohen's d. Can be small (corelat .10), medium (.30), or large (.50).

Front 46

How much power do we need?

Back 46

.80 is realistic/reasonable for behav sciences.

Front 47

Index of the amt of observed variability in participant's behavior.

Back 47

variance

Front 48

Indicates how tightly or loosely a set of scores cluster around the mean.

Back 48

variance

Front 49

Scores that are tightly clustered around the mean have ____ variance.

Back 49

small

Front 50

equivalent to the average squared difference from the mean

Back 50

variance

Front 51

To get the standard deviation from the variance, take the ...

Back 51

square root of the variance.

Front 52

______ is in the same unit of measure as the original scores.

Back 52

standard deviation

Front 53

This propoerty is dependent upon the scales of the variables, which can make it difficult to make comparisons.

Back 53

covariance

Front 54

is an index of the relationship between two variables

Back 54

pearson's r

Front 55

?Provides a general index of the relationship between two variables by taking into account the different variabilities of the 2 original sets of scores.

Back 55

z score

Front 56

When you take raw scores to standard scores, this is a ...

Back 56

linear transformation

Front 57

z scores are an example of ..

Back 57

standardized scores

Front 58

Every score is changed by multiplying or dividing by a constant and or adding or subtracting a constant.

Back 58

linear transformation

Front 59

when you do linear transformation, __ and __ change, but the relationship between scores is ____.

Back 59

When you do linear transformation, means and standard dev change, but the relationship between scores is retained.

Front 60

The sum of a set of z scores equals..

Back 60

0

Front 61

The mean of a set of z scores equals..

Back 61

0

Front 62

The variance of a set of z scores equals...

Back 62

1

Front 63

The standard deviation of a set of z scores equals..

Back 63

1

Front 64

Is the shape of a distrubution of x affected by transforming it to Zx?

Back 64

No

Front 65

Is the absolute value of the correlation of X with any other variable affected by transforming it to Zx?

Back 65

No

Front 66

Used to describe our data from the sample

Back 66

sample statistics

Front 67

M and r are examples of...

Back 67

sample statistics

Front 68

summarize population distribution

Back 68

population parameters

Front 69

mu for mean and p for correlation. These are associated with what types of stats?

Back 69

population parameters

Front 70

How are population parameters typically determined?

Back 70

Population parameters are usually estimated from sample statistics

Front 71

A sample contains info about what two things?

Back 71

1. population distribution
2. parameters of population distribution

Front 72

the probability distribution of a sample statistic for all random samples of a given size from the same population

Back 72

sampling distribution

Front 73

What are some examples of things for which we can create a sampling distribution.

Back 73

We can create a sampling distribution for any stat: variance, median, range, mode, correlation

Front 74

How would you go about creating a sampling distribution of the mean?

Back 74

We would draw many samples of n cases independently and at random from some population. We would then calculate the mean for each sample and create a distribution of those means.

Front 75

There are two types of estimation of sample stats. What are they?

Back 75

biased and unbiased estimates of sample stats.

Front 76

We find a single value that is our best guess of the population parameter. What is this type of estimation called?

Back 76

Point estimation

Front 77

Create confidence intervals defining the range of good estimates to either side of the point estimate. What is this type of estimation called?

Back 77

Interval estimation

Front 78

The best estimate of mu is M.
What does this mean?

Back 78

?The best estimate of the population mean is the sample mean

Front 79

An estimated range of values with a known high probability of covering the true population value.

Back 79

Confidence interval

Front 80

The upper and lower limits of the CI tell us what about the effect?

Back 80

The upper and lower limits of the confidence interval show how small or large the effect might be in the population.

Front 81

If the value specified in the null hypothesis does not fall within the CI, do we accept or reject the null?

Back 81

reject

Front 82

If we collected many samples and computed the mean and 90% CI for each sample, then 90% of those CIs would contain...

Back 82

the true population mean.

Front 83

The width of the confidence interval depends on 3 things. What are they?

Back 83

1. chosen level of CI (i.e., 95 will be narrower than 99).
2. population strd dev (the larger the variance, the wider the CI)
3. sample size (the larger the sample, the narrower the CI).

Front 84

Which will be wider - a 95% CI or a 99% CI ?

Back 84

99%

Front 85

The ____ the variance, the _____ the CI.

Back 85

The higher the variance, the wider the CI.

Front 86

The ______ the sample, the _____ the CI.

Back 86

the larger the sample, the narrower the CI.

Front 87

The sampling distribution of the mean will approach a normal distribution as sample size gets larger.

Back 87

central limit theorem

Front 88

A guess concerning the value of a parameter.

Back 88

statistical hypothesis

Front 89

Describe the hypothesis testing procedure.

Back 89

1. define null and alternative hypothesis
2. test null hypothesis
3. make decision about null hypothesis
4. alternative hypothesis: a guess concerning the value of a parameter that suggests the presence of some pattern or relationship of interest; null hypothesis: a guess concerning the value of a parameter that suggests absence of some pattern or relationship of interest.

Front 90

A guess concerning the value of a parameter that suggests the presence of some pattern or relationship of interest

Back 90

alternative hypothesis

Front 91

a guess concerning the value of a parameter that suggests absence of some pattern or relationship of interest.

Back 91

null hypothesis

Front 92

Describe the steps associated with hypothesis testing

Back 92

step 1: Set up H0 and H1
2: find critical value of t
3: set alpha and set up decision. Rule to reject null if t value is greater than or equal to the critical t.
4: compute t
5: if observed t value is greater than or equal to the critical value of t, than we can reject the null (there is a significant diff).

Front 93

you reject the null if the t value is ____________ the critical t. Rejection means that ___.

Back 93

greater than than or equal to,
Rejection means that there are significant differences between groups.

Front 94

If we obtain a siginificant result, we've observed something relatively unlikely (given the hypothetical situation) but ____ ____ given some alternative situation.

Back 94

more likely

Front 95

What does it mean to find a significant mean difference between two groups (p<.05)?

Back 95

If we collected lots of samples and if the null hypothesis is actually true, then we'd find a mean difference this big or bigger only 5 out of 100 times.