Adv Corelat Methods Quiz 3

Front 1

Type I error is represented by...

Back 1

alpha

Front 2

Type II error is represented by...

Back 2

Beta

Front 3

This term represents the power to detect a true effect.

Back 3

1-Beta

Front 4

If the null is true and you reject it, this is...

Back 4

alpha (type I error)

Front 5

If the null is true and you do not reject it, this is...

Back 5

1-alpha (correct decision)

Front 6

1-alpha and 1-beta both represent correct decisions. What specific decision is attributed to each?

Back 6

1-alpha = the null is true and you don't reject it

1-beta= the null is not true and you reject it

Front 7

If the null is true and you reject it, this is...

Back 7

alpha (type I error)

Front 8

If the null is true and you do not reject it, this is...

Back 8

1-alpha (correct decision)

Front 9

1-alpha and 1-beta both represent correct decisions. What specific decision is attributed to each?

Back 9

1-alpha = the null is true and you don't reject it

1-beta= the null is not true and you reject it

Front 10

If the null is not true and you reject it, this is ...

Back 10

1-Beta (correct decision)

Front 11

If the null is not true and you do not reject it, this is ...

Back 11

Beta (type II error)

Front 12

In addition to being a correct decision, what term is used to describe 1-Beta?

Back 12

Power

Front 13

If power = .80, what is the likelihood that you'll make a type II error?

Back 13

.20

Front 14

Researchers are willing to accept higher type ___ error than type _.

Back 14

Researchers are willing to accept higher type II error than type I.

Front 15

Is the variation in y related to the variation in x?
This question is reflective of what concept?

Back 15

correlation

Front 16

Discuss the three pieces of information that a correlation provides.

Back 16

1. form of the relationship (i.e., is it linear)
2. direction of the relationship (i.e., if the plot is leaning down and to left (neg) or up and to right (pos)).
3. strength of the relationship (i.e., points that are closely clustered together would suggest a strong relationship).

Front 17

Correlation is designed only to detect what type of form of the relationship between 2 variables.

Back 17

linear relationship

Front 18

What are some of the limitations of scatterplots?

Back 18

subjective, data is non-numerical

Front 19

index of the tendency for 2 variables to go together

Back 19

covariance

Front 20

reflects how much the variation in one variable is related to variation in a second variable.

Back 20

correlation

Front 21

What two things are true about the manner in which covariance will be expressed?

Back 21

covariance is expressed
1. in terms of deviation from the mean
2. in the original unit of measure

Front 22

Covariance depends on the scale of variables. To remove this property, what do we do?

Back 22

We make variables equivalent in terms of mean and strd dev before comparing the two variables. We do this by transforming them into z scores.

Front 23

In order to gauge the relationship between the two sets of numbers, what do we examine?

Back 23

the cross product of z scores

Front 24

The term "r" is known as...

Back 24

the pearson product moment correlation coefficient.

Front 25

The r from cross products is a ...

Back 25

pure measure

Front 26

What is the range associated with r, the correlation coeficient?

Back 26

-1 to 1

Front 27

If there is a perfect relationship between height and weight, what does this mean?

Back 27

If you know height you can difinitively determine weight.

Front 28

r squared is equal to...

Back 28

the coeficient of determination

Front 29

If r squared is equal to .25, what does this mean?

Back 29

This means that 25% of the variability (variance) in one measure can be explained by the other measure.

Front 30

can be used for 2 continuous variables

Back 30

pearson's r

Front 31

used for skewed variables

Back 31

spearman's r

Front 32

used for two dichotomous variables; it does the same thing as pearson's but is easier to compute.

Back 32

phi coefficient. (represented by circle with a vertical line going thru it)

Front 33

used for 2 dichotomized variables

Back 33

tetrachoric coeficient

Front 34

used for one continuous and one dichotomous variable

Back 34

point-biserial (r bis)

Front 35

used for one continuous and one dichotomized variable

Back 35

biserial (r bis)

Front 36

the quantity is independent of the unit of measure, meaning that we can directly compare two of these from different studies.

Back 36

correlation coeficient

Front 37

What does the sign of r tell you about the relationship between two variables?

Back 37

sign of r tells you about the direction

Front 38

What does the absolute value of r tell you about the relationship between two variables?

Back 38

It tells you about the magnitude of the relationship.

Front 39

What is simple linear regression?

Back 39

a method for representing the dv as a linear function of the iv

Front 40

What are the two ways to conceptualize SLR?

Back 40

1. geometrically as a line
2. algebreically as an equation

Front 41

What are the 2 purposes of SLR?

Back 41

1. predicition
2. explanation (i.e., use one variable to explain variance in another).

Front 42

What are the two questions asnswered by SLR?

Back 42

1. In what direction is the relationship going?
2. By how much does this variable increase when I
increase another one?

In other words,
1. Does Y increase or decrease as X increases?
2. By how much does Y increase or decrease as X increases?

Front 43

Discuss dependent and independent variables.

Back 43

"if x is given, then y occurs", where x represents the independent variable and y represents the dependent variable.

Front 44

Does optimisim pre-surgery predict recovery post-surgery?

What is the dv? iv?

Back 44

dv = y = recovery

iv = x = optimisim

"if x is given, then y occurs", where x represents the independent variable and y represents the dependent variable.

Front 45

y hat = a + bx

Identify what each term means

Back 45

y hat = predicted y (a specific point on the line)

a = y intercept; where line touches y axis when x = 0

b = slope

Front 46

We should assume a linear relationship unless..

Back 46

1. you see something else in scatterplot
2. you have a theory that predicts diff relationship

Front 47

Discuss the various terms that can be used to describe x and y.

Back 47

x = iv = predictor variable

y = dv = criterion variable

"if x is given, then y occurs", where x represents the independent variables and y represents the dependent variables.

Front 48

How do we determine how well the observed data fit the line?

Back 48

find SS error; to do this we take the distance of each outlying point (the residuals) and square them.

Front 49

this is a measure of badness of fit: the higher it is, the worse the fit. Your goal is to have this be as smalll as possilbe.

Back 49

SS error

Front 50

the best line will have the smallest sum of squared errors

Back 50

least squares criterion

Front 51

Our goal in regression is to find values of a and b such that:

Back 51

The ss error of prediction is minimized. Finding a and b values that minimize ss error allow us to find the best fitting line.

Front 52

Alpha and Beta are...

Back 52

population parameters

Front 53

*When you have just 2 variables, the standardized b is equal to.....

Back 53

*When you have just 2 variables, the standardized b is equal to the correlation (r).

Front 54

the a and b that are calculated from the computational formula are....

Back 54

least squares estimates of the population parameters alpha and beta

Front 55

This term is the average estimated value of y when x =0.

Back 55

a

Front 56

This term is the point at which the regression line intercepts the y axis.

Back 56

a

Front 57

This term is called the y intercept and the "constant" ( in spss output).

Back 57

a

Front 58

This term represents the average estimated change in y associated with each one-unit change in x.

Back 58

b

Front 59

This term is called the slope and the regression coeficient.

Back 59

b

Front 60

Here, 34.75 is the average estimated value of recovery scores when optimisim is zero. However, range for optimisim doesn't go that low. What should be done?

Back 60

Interpret it as a constant

Front 61

Talk about the ways to refer to a and b.

Back 61

a = y intercept = constant
b = slope = regression coeficient

Front 62

when x goes up by one unit, this is how much y will increase or decrease

Back 62

b

Front 63

when x goes up by one unit, this is how much y will increase or decrease

In the statement above, one unit should be replaced with something. what is it?

Back 63

the "one unit" piece refers to the value of scale. For instance, one unit = 1 standard deviation

Front 64

What type of regression coeficient does the following equation yield:

y hat = a + bx

Back 64

y formula yields unstandardized regression coeficient. To obtain standardized regression coefficient, you must use the following formula:
zy = b * zx + a

Front 65

when using the equation z = b*Zx + a, what is the value of a? Why? Having an a of this value will do what to the graph?

Back 65

The mean for any set of z scores is zero, so a will be zero. Therefore, the graphed line will go thru (0,0).

Front 66

when using the equation z = b*Zx + a, what is equal to the slope?

Back 66

for standardized variables, r = slope

(I believe that this is only true with standardized variables and one predictor).

Front 67

represents what proportion of the variance in the dv I can explain by variation in my iv

Back 67

R square

Front 68

Interpret the following:
R square = 40

Back 68

This means that 40% of the variance in dv is accounted for by variance in the iv.

**In the homework, it is important to substitute the actual values for the dv and iv - ie, optimisim and recovery

Front 69

t squared is equal to

Back 69

F

Front 70

Interpret the following:
Standardized B = .636

Back 70

Basic formula: for every one unit increase in x, y will increase by .636 units.

For every 1 SD increase in optimisim, there will be a corresponding strd dev incr of .636 in recovery scores.

Front 71

?If your data have been standardized, then the 0,0 point of the graph represents..

Back 71

the mean of both variables (x and y)

In other words, the line and points have stayed the same (following standardization) but the axis has changed.

Front 72

The standardized regression line will always go through which point on the graph?

Back 72

(0,0)

Front 73

The variance of y from the mean can be partitioned into 2 parts. What are they?

Back 73

1. explained variance (predicted y - mean)
2. unexplained variance (y - predicted y).

Front 74

The total variance in y can be represented by which term?

Back 74

total SS

Front 75

The part of the variance that we are able to explain.

Back 75

SS regression

Front 76

The deviation in what we predicted they'd get for y and what they actually got.

Back 76

SS error

Front 77

predicted y - mean =

Back 77

explained variance

Front 78

y - predicted y =

Back 78

unexplained variance

Front 79

standardized regression coefficient is represented by

Back 79

b*

Front 80

With standardized variables and one predictor, r = ...

Back 80

r = b* = slope of best fitting line.

Front 81

What is relationship between b and beta?

Back 81

you can calculate one from other. see formula in notes

Front 82

Our best guess about a particular y score without the regression line would be..

Back 82

the mean of y

Front 83

The equation for the regression line tells us that the variance of our actual y score from the mean of y is equal to....

Back 83

variance of our actual y score from the mean of y = variance accounted for by regression line + error variance