Final

Question 1

Null Hypothesis

Answer 1

• H0
• original hypothesis, no change
• there is no wolf

Question 2

Alternative Hypothesis

Answer 2

• H1
• there is an effect
• there is a wolf

Question 3

Type I Error

Answer 3

• false alarm
• rejecting the null hypothesis when null is true
• thinking there was effect when there wasnt

Question 4

Type II Error

Answer 4

• miss
• failing to reject null when the null is false
• missing a real effect

Question 5

Power

Answer 5

• rejecting the null hypothesis when the alternative hypothesis is true
• probability of a Type II error
• P = 1 - B

Question 6

Low Power

Answer 6

• less likely to detect a real effect
• high risk of a type II error

Question 7

High Power

Answer 7

• more likely to detect a real effect
• lower risk of a type II error

Question 8

Alpha α

Answer 8

• level of significance
• = 0.05
• probability of a type I error

Question 9

Central Limit Theorem

Answer 9

• as sample size increases, sampling distribution of the mean is normal
• large sample = smaller standard error
• n > or equal to 30

Question 10

What do we mean when we say p-value?

Answer 10

how likely is it to get a result like this if the null hypothesis is true?

if the null is true, how likely is it to get a result this extreme?

Question 11

What p-value is considered significant in psych?

Answer 11

p < 0.05:
- these results unlikely to occur by chance alone
- question the null
- this result is statistically significant

p > 0.05
- these rules could’ve happened by chance
- stick with the null
- this result is not statistically significant

Question 12

Ways to Increase Power

Answer 12

1. If directional hypothesis, use one tailed test
   - when using one tail tests, the CV of you test statistic (like z-value) is smaller
2. Increase sample size
   - bigger sample = smaller standard error -> easier to detect effects
3. Increase the dose or exposure
   - levels you choose for your independent variable can change effect size
4. Decrease variability
   - reduces extraneous factors

Question 13

Probability

Answer 13

= number of ways it can happen / total number of outcomes
ex: probability of drawing a 10 from deck of cards

Question 14

Mutually Exclusive Events

Answer 14

• P(A or B) = P(A) + P(B)
  ex: probability of drawing a 10 or a face from cards
• just add together the probs of both because they don’t relate

Question 15

Non-Mutually Exclusive Events

Answer 15

• P(A or B) = P(A) + P(B) - P(A and B)
• probability of both then subtract the amount of times theres an intersection

Question 16

Series of Outcomes of Independent Events

Answer 16

(probability of heads)flip 1 x (probability of heads)flip 2
.5 x .5 = .25 – used a product of their individual probabilities

Question 17

Correlation Coefficient

Answer 17

• measure of the degree of relationship between 2 sets of scores
• varies between -1.0 and +1.0

+/- .00 - .29 = none (.00) to weak correlation
+/-.30 - .69 = moderate correlation
+/-.70 - 1.00 = strong correlation

Question 18

Scatterplot

Answer 18

graphically represents the relationship between 2 variables

Question 19

Why do we say we cannot assume causal relations when we talk about correlations?

Answer 19

1. we assume the correlation is causal & one variable causes changes in the other (not true)
2. fail to recognize that other variables could be responsible for the observed correlation (the third variable)

• when 2 variables are correlated, it means one variable is present at a certain level, the other variable also tends to be present at a certain level
• not saying its a prediction guaranteed or causal, but the variables occur tg at specific levels

Question 20

What is restricted range? Why is it a problem?

Answer 20

• truncating the variable limits variability

ex: BMI & heart risk disease
- looking @ everyone from underweight to obese, the correlation might be strong, r = .82

• but if you look at the ppl in normal BMI (restricted range) now r = .22, it looks weak but only as you cut the extremes

Question 21

When we calculate Pearson’s r, we make certain assumptions about the nature of the relationship. What do we assume?

Answer 21

• most common correlation coefficient
• we assume: linearity, variables measures of interval or ratio scales

Question 22

What is the difference between a one-tailed and two-tailed test?

Answer 22

One-tailed test:
- had a hypothesis in mind
- directional

Two-tailed:
- no predictions/hypothesis
- null hypothesis: correlation between ___ and ___ is zero in the population. H0 = 0
(no linear relationship)
- alt hypothesis: correlation between ___ and ___ is not zero. H1 not equal to 0.

Question 23

Know how to calculate the df for r and use the table of r-values to determine if a relationship is significant.

Question 24

Know how to report an r value in APA style.

Question 25

What is r2? What does it tell you?

Answer 25

- measure of the proportion of variance in one variable accounted for by another variable - how well can one predict the other ex: r = .764 r2 = .5837 so, 58% of variance in state scores can be explained by variance in practice test scores

Question 26

What is a regression line? If you are given the formula for a regression equation, you should be able to calculate the regression line if you are given r, the means and standard deviations.

Answer 26

best fitting straight line down thorugh the center of the scatterplot - indicates relationship between variables Y1 = bX + a

Question 27

In regression, we talk about predictors and criterion. Which are sometimes referred to as the independent variables? Is X or Y the predictor in a regression equation?

Answer 27

X = predictor Y = criterion

Question 28

What is the difference between simple regression and multiple regression?

Answer 28

Simple Regression: - one variable predicts another - r^2 Multiple Regression: - 2 more more variables are used as predictors ex: predict first grade state testing using - kindergarten results AND teacher assessment of readiness - R^2

Question 29

What is R2?

Answer 29

- coefficient of determination - the proportion of variance in the Y that can be accounted by the variation of the combined predictor variables - ranges 0 - 1.00 - when R2 increases, the overlap between criterion and predictor variable is larger - basically r^2 but using multiple variables

Question 30

What is the difference between rAB.C and rBC.A?

Answer 30

rAB.C = correlation between A&B, controlling C rBC.A = correlation between B&C, controlling A

Question 31

Why would we compute a partial correlation?

Answer 31

- tells us the relationship betwen A&B variables after removing the effect of C - "if we remove C, are A & B still related?"

Question 32

How do the assumptions for Pearson's r and Spearmans rho differ?

Answer 32

Pearsons R: - use for linear relationships - interval/ratio - sensitive to outliers Spearman's Rho: - numerical and odinal variables - monotonic (not a straight line) - less sensitive to outliers - no normality assumptions - ordinal, outliers, unequal variance, non-linear

Question 33

Explain what a confidence interval is. When we talk about confidence intervals, what assumptions do we make about μ?

Answer 33

gives you a range of values that you beleive is likely to contain Mu (population mean) - estimates where Mu might fall

Question 34

Be able to calculate a confidence interval.

Question 35

Understand the difference between the standard deviation, standard error of the mean and a confidence interval. What can we say about their relative width?

Answer 35

Standard Deviation: - tells us how spread out individual scores are in your sample - measures variability in individual data points - how are people in my sample different than each other width: narrowest Standard Error of the Mean: - how much sample mean varies from the population mean. - small n = bigger SEM - bigger n = less SEM width: medium Confidence Interval: - range of values we think true population lies width: widest