Psychology 210 Exam 2

Give the correct formula for Cohen's d for a single-sample t test

(M-u)/s

The correct formula for Cohen's d for a paired-samples t test is:

(M-u)/s

What is the equation for the t statistic for a distribution of means?

(M-um)/sM

The critical cutoffs for a 95 percent confidence interval are -1.96 and 1.96. What are the cutoffs for a 90 percent confidence interval?

-1.65 and 1.65

What are the advantages are using a within subjects design over a between subjects design? (short answer question)

-A type of experiment done with participants being in both parts of the experiment -An example being, if you want to see if a drug supplement would increase speed of an individual you would first test the participants with no drug and then test those same participants after they were given the drug. -ADVANTAGE 1: you do not need a large sample size. You are comparing people to their own results -ADVANTAGE 2: it helps reduce error in individual differences. Because all participants in the design are exposed to each level of treatment, individual differences will not influence the results

What are cohen's conventions for r^2

.01 small effect .06 med effect .14 large affect

Name small,med, and large effect sizes (cohen's d)

.2 small, .5 med, .8 large

Name 6 steps of single sample t test

1. identify assumptions (dependent scale, normal, random sample) 2. state hypotheses 3. characteristics of comparison dist 4. identify crit values, df=N-1 5. calculate 6.decide

What are ways to increase power? Which of them should you never do?

1. increase sample size (this reduces standard eorror, creates less overlap, + you can do this!) 2. recude variability (reduce pop stand dev) 3. increase size of effect (dif between means, can do this!) 4. increase alpha (instead of .05 do .10, + DONT do this) 5. take two-tailed and make it a one tailed test (instead of 2.5 and 2.5 on each side make .5 on one, lose all power on other side, DONT do this)

In calculating pooled variance, why are adjustments made for sample size?

Because estimates from smaller samples tend to be less accurate than those form larger sample.s.

When the population mean falls in our confidence interval, we conclude that the sample:

Comes from the same population

"95% of the time we expect the sample statistic to between these numbers" What is this term?

Confidence interval

Credit card debt for 10 people was compared before a major economic crash and then after economic recovery. What would be the appropriate null hypothesis for this research?

Credit card debt before the economic crash will be approximately the same as that after the economic recovery.

Critical t values ______ as the degrees of freedom ____

Decrease; increase

What standardized value helps measure the size of the difference in relation to the variability of the data rather than to that of sample size?

Effect size

What is the 4th assumption for independent t-tests?

Homogeneity of variance (variance of two populations can't be over 2) standard error of large sample/standard error of smaller sample

A dot plot allows us to _____, while a stem- and -leaf plot does not.

Include more than one sample

Overlap between distributions being compared decreases, the effect size:

Increases

The _______ estimate acknowledges the amount of uncertainty in the _____ estimate by reporting the margin of error

Interval; point

The _____ estimate acknowledges the amount of uncertainty in the ______ estimate by reporting the margin of error.

Interval;point

In Cohen's d, the closer together the means of two distributions, the ____ the effect size, assuming the standard deviation is held constant

Lower

An independent-samples t test compares sample _______ differences

Mean

A Cohen's d of -.49 is what type of effect size?

Medium

How do you find confidence interval?

Msample+/m t(sm)

In Bayesian statistics probability distributions are based on:

Prior beliefs of probablities

Where do you find confidence interval for z statistics on the packet? What statistics do you need to find the confidence interval?

Sample derived on first page (you know sample and parameters) -standard error, tcrict, sample mean

In calculating confidence intervals in a z distribution, start by finding the z scores centered around the:

Sample mean

A researcher calculates a confidence interval for a paired-samples t test. That interval is centered around -8.32, which is the:

Sample mean difference

What does pooled variance take into account?

Sample size

Glen has limited funds left in his grant to pay participants to be in his experiment. He wants to know how many participants he needs to detect a difference between conditions. What statistical technique would help him estimate how many participants he needs?

Statistical power analysis

Can you ever have too much power and would that be a problem? (short answer)

Statistical power is a measure of the likelihood that we will reject the null hypothesis, given that it's actually false (Probability of not making a type II error) Usually researchers want at least .8 or 80% power to ensure true effects. There are times, however, where too much power means the test is very sensitive to true effects. These effects could be relatively small and though you may have technically statistically significant results, the results aren't actually useful or relevant for your study.

T or F: Increasing the size of your sample will have no impact on your effect size.

T

T or F: if the sample data are in the critical region with a=.01, then the same sample data would still be in the critical region if a were changed to .05.

T

T or F: you use estimated sd for effect size (cohen's d)

T (M-popm/s)

T or F: If both studies use an alpha level of .05,, a study with an N of 100 is just as likely to results in a type I error as a study with an N of 20

T (type II not??)

What happens to the effect size if the variability within each distribution becomes smaller?

The effect size increases

What is statistical power?

The measure of our ability to reject the null hypothesis given that the null is false (prob of not making a type 2 error)

In order to conduct a single-sample t test, one needs to know:

The population mean and the propertied of the sample

An effect size of .53 was calculated on data after performing a hypothesis test with a single-smaple t statistic in which the null hypothesis was rejected. What can be concluded about the results based on this information?

The sample data are significantly different from what was expected based on the population, with tthe sample mean .53 standard deviations greater than the population mean.

When determining critical values for a t test, we don't assume to use 1.96. Why is this?

They change based on the sample size

By increasing statistical power, the probability of making a ____error is _____

Type 2; decreased

Dividing ss by "n-1" results in a "sliding" adjustment, yielding an unbiased estimate of sd? Explain the adjustment made by "n-1" especially as it relates to sample size. (short answer)

We use "n-1" as our denominator when computing sd for t-tests. This is because we do not know the parameters. The "n-1" gives us an unbiased variance because if we just divided by n, we tend to underestimate the true variance. IN terms of sample size, any sample is likely to have less spread than the entire population. By subtracting 1 by N, we get a larger and more accurate standard deviation, that in turn, better represents variance.

The process of________ involves researchers at different institutions work together recruiting and sharing methodology online.

crowdsourcing

What is the correct formula for calculating degrees of freedom for an independent-samples t test?

dftotal= dfX+dfY

For a single-sample t test, how do you find estimated population standard deviation?

s=E(x-m)^2/(N-1)

How do you find the z statistic?

sample mean-sample population mean/ standard error

How do you find standard error for a z test?

sd/square root of sample size

How do you find standard error for a single-sample t test?

sm=s/sqrt N

The ____ indicates the distance of a sample mean from a population mean in terms of the standard error.

t statistic

What is the correct way to report results? (APA style)

t(df)=t stat, p<.05 (if reject) (p>.05 if fail to reject)

What test is used when we don't know variability within a population (don't have sd)

t-test

R^2 for independent t tests looks like.....

t^2/t^2+df

A dot plot graphs all the data in a sample, with the range of scores along the _____ and a dot for each data point above the corresponding value.

x-axis