RUTGERS quantitative methods in psych final

standard deviation

average distance from the mean

variance

SD squared

SEm

SD for sampling distribution of the mean

est. SEm

estimation of SEm when population standard deviation is not available

SEd

SD for infinite number of differences between pairs of sample means

z score

determines what percentage of cases lie between that score and the mean in a normal distribution; only used when population SD is known; Ho: sample mean = population mean

single sample t test

how many SEm a sample mean is from the pop mean; Ho: sample mean = population mean

independent sample t test

compares 2 means when there are no population parameters

chi squared

when DV is nominal; Ho: fo = fe

reliability

consistency; similar results when retested

validity

measuring what it intends to measure

Type 1 error

reject Ho when you shouldn't

Type 2 error

accept Ho when you shouldn't

Pearson's r

correlation coefficient; relationship is linear; Ho: rpop = 0

coefficient of determination

effect size in Pearson's r; r squared

paired samples t test

used for repeated measures experiments

one way ANOVA (F test)

differences between 3 or more means; Ho: popmean1 = popmean2 = popmean3; looks at variance in DV and it's sources

why the F test instead of multiple t tests?

multiple t tests would increase the chance of type 1 error

Tukey's HSD

post hoc test to find out which means have a significant difference

SSw

affected by levels of IV and each score

SSb

affected by grand mean, condition means, and the sample size of each condition

factorial ANOVA

more than one IV

main effect of factorial ANOVA

effects of each IV

interactions of factorial ANOVA

effects of combos of IV

Fisher's z

used to compare two Pearson's r correlations; null is that there is no difference

simple linear regression

to predict a value of Y (criterion) from a given value of X (predictor)

residual

errors of prediction

least squares regression line

minimizes the distance between each point in the scatterplot and the line; minimizes residuals

slope

for every unit change in X, what is the unit change in Y?

y-intercept

what is the value of Y where the regression line crosses the Y axis? (x = 0)

in linear regression, as r gets larger,

residuals get smaller

in simple linear regression, if r is insignificant,

Ypred = My

in linear regression, as r gets smaller,

Ypred gets closer to My

multiple linear regression

when there is more than one predictor (X)

3 parts of multiple R

1. Pearson's r

2. multiple R and F test

3. making predictions using linear regression equation

in multiple linear regression, if no correlation between a predictor and Y,

use a simple linear regression

in multiple linear regression, if no correlation between both predictors and Y,

the prediction would be My

in multiple linear regression, if no correlation between the predictors,

substitute 0 anywhere you see rx1,rx2 in formulas

there will always be residuals unless

Pearson's r is 1.0 because that means there is a perfect relationship (100% accuracy)

In the t and F ratios, the numerator will always represent variability due to _______ , and the denominator will always represent ________.

the IV and sampling error; everything else

df in single sample t test

N - 1

df in ind. sample t test

(n1 + n2) - 2

df in chi squared

# of cells - 1 OR (# of rows - 1) x (# of columns - 1)

df in Pearson's r

N - 2

df in paired samples t test

N - 1

dfb ANOVA

# of levels of IV - 1

dfw ANOVA

N - # of levels of IV

dfb multiple R

# of x and y variables - 1

dfw multiple R

N - 3

The numerator in the F ratio represents the variation in the dependent variable due to chance and ____.

the independent variable

The larger the F ratio,

the more likely we are to reject the null

The in 2x2 factorial ANOVA, how many null hypotheses would there be?

3

In a 2x2x3 factorial ANOVA, there are

3 independent and 1 dependent variable

In an F test, the SS within represents

The squared amount of variation between each individual score and the mean in the group in which the individual score is located

Which of the following is FALSE with respect to SS within?

It includes variation in the DV due to the effect of the IV

A researcher conducts a one way anova and reports F(4,95)=13.82, p<0.01. The 4 represents ______ and the 95 represents _______.

Degrees of freedom between; degrees of freedom within

A study is being conducted by the US Agricultural service on the effects of fertilization and pesticide use on plant growth. Six randomly selected groups of corn plants are exposed to one of the combinations of conditions for 3 months, and then plant height is measured. We should analyze this data with

A 2x3 factorial anova

In the study described in #11, lets say it was found that the plants exposed to high nitrogen grew significantly taller than the others, but only if they were not exposed to pesticide. When they were exposed to pesticide, they grew to the same height as the plants in the other conditions. The best way to describe this result would be

There was a significant interaction between pesticide exposure and fertilizer type

In a 2x3x4 factorial anova, which of the following would be true

three different independent variables (with 2,3, and 4 levels)

In a 2x4 factorial anova, how many individual treatment groups are there?

8

Assume for a 2x4 factorial anova that both main effects were significant. Which of the following would be true?

Chi Square

How many Fs would we need to examine in the 2x4 anova?

2

In a hypothetical study, a researcher wanted to see if there was any effect of different drugs on whether or not people with asthma experienced improvement. The appropriate test would be

Chi Test

A researcher wants to find out if watching TV has an effect on students' academic performance. For one semester she has them watch 20 hrs of tv per week; for another semester, she has the same group of students watch no tv. She then compares their GPAs from both semesters. Which statistical test should she use?

two tailed paired samples t

The denominator in the within-subjects t test would usually be much smaller than the denominator in the independent samples t test. Why is that true?

Because r >0 for paired samples, but r =0 for independent samples

When we have to compare the means of more than two groups, we use the 1 way anova. Why don't we just use t tests to compare each pair of mean?

Using repeated t tests would increase the probability of rejecting the null hypothesis when it is in fact true to greater than 0.05.

In an anova design, as the scores in each cell vary more from the cell mean, the ____ will _____.

The ssw; increases

In a repeated measures t test, as the correlation between the before and after scores decreases, the probability of rejecting the null hypothesis _____.

Decreases

If we use an alpha of 0.01, the probability of rejecting the null hypothesis when it is in fact true ______.

Decreases

The null hypothesis in a Pearson's r test is _______.

R pop = 0

In a 1 way anova where 3 means are being compared, the null hypothesis for a non-directional research prediction would be _______.

U sample a= U sample b = U sample c