Exam 2

a t-test result of t(30) = 2.87 with a critical value of 1.96 indicates:

a significant result

An ANOVA result with an F-value of 1.00 indicates:

Between group variance = Within

group variance

The degrees of freedom for a t-test comparing two groups of 10 participants would be?

18

Your independent samples t-test was significant! Next, you should:

Check the group means to conclude

Your ANOVA comparing 3 groups was significant! Next, you should:

Run a post hoc analysis to conclude how the groups differ

Counterbalancing is a method used for

Preventing ordering effects

A researcher is looking at whether listening to music improves exam scores. The music group tested in the morning, and the no-music group is tested at night. What is the confoun variable?

time of day

Which of the following z-scores would most likely be used to determine outliers?

a) +.50

c) +3.00

b) - 1.00

d) +1.25

+3.00

A researcher tells participants that they are completing a study on memory, but the measuring prejudice. What is the ethical issue that is most relevant?

deception

If a student has a z-score of +2.00. What does this mean?

2 standard deviations above the mean

What happens to the shape of a distribution when the scores are converted to z-score

shape stays the same

Which of the following would use an independent samples t-test?

comparing men and women on depression levels

A t-test is written as t (30) = -2.77, p = 0.01. What does 30 mean?

degrees of freedom

After a study which includes deception, the researcher explains why the true purpose of the study and why deception was necessary. What is this process called?

debriefing

Statistical significance usually requires a p-value that is:

less then .05

this is the formula for

z-scores

In a z-score distribution...

mean = 0, SD = 1

“There is a difference between older workers and younger workers on job satisfaction”

non-directional alternative hypothesis

19. “Studying will increase test performance” is a...

directional alternative hypothesis

An independent variable is the variable that is:

manipulated by the experimenter

what is a primary strength of experimental research designs?

greater internal validity

descriptive statistics

summarize, organize, and simply data

inferential statistics

statistical procedures the allow us to make generalizations about the population based on sample data

research questions

a broad question you’re interested in answering via a research study
for correlational studies (association claims)
→ relationships between variables
for experiments and quasi-experiments (casual claims)
→ group differences or change

hypothesis

states what you think will happen in your research study (i.e, your results)

hypothesis testing

using data from samples to evaluate a research question about the population
1) state the hypothesis
2) obtain a random sample from the population
3) test the hypothesis using sample data
4) compare the results with the hypothesis

null hypothesis (Ho)

no effect, difference, relationship,etc

alternate hypothesis (or research) (H1)

what difference do you expect to see
→ there is a difference, change, or relationship

non-directional alternate hypothesis

reflects a difference between groups, but the direction Is not specified

directional alternate hypothesis

reflects a difference between groups AND the direction is specified

why do we have a null hypothesis?

assume Ho is true until evidence is strong enough to suggest otherwise
→ easier to disprove than to prove
however, rejecting the null hypothesis does NOT prove the alternative hypothesis true

difficultly with samples

no sample will be exactly the same as the population

sample size and sampling error

smaller samples have more sampling error
→ chance, random factors are just as likely to increase a sample mean as they are to decrease it
→ with large enough sample size, the random factors cancel each other out
→ the smaller the sample size, the easier it is for a single outlier to influence the mean

central limit theorem

1) sampling distributions are approximately normally distributed
2) the mean of sampling distribution is equal to the population mean
3) the standard deviation of the sampling error distribution is a function of the sample sizes and standard deviations with samples

(CLT) approximately normal distributed

sample means can be too high / too low compared to population mean
most sample means close to the population means, only a few being extremely different from the population mean

(CLT) mean of the distribution

the mean of the sampling distribution is equal to the population mean
a lot of sample means will be equal to the population mean

(CLT) standard deviation of the distribution

the standard deviation of the sampling error distribution is called standard error (can be calculated)

standard error

quantifies the sampling error

standard error of the mean

average (I,e standard) distance between sample mean and population mean

translation

measures how accurately, on average, a sample mean represents its corresponding population mean

standard error decreases as…

the sample sizes within the sample increases
standard deviation in the population decreases
→ if the null is false the sample, we infer it is likely false in the population and “reject Ho”

statistical signifiance

indirect support for H1 since we can’t test it directly
→ using a sample to make inferences about a population
→ when we get a result from our sample, we calculate the probability
PROBABILITY of Ho being true is really low

z-score

how many standard deviations away from a particular score is from its sample mean
→ you can turn any score into a z-score
→ z = x-m/s (z = z-score, x = individual raw score, M = mean of the sample, s = standard deviation of the sample)
→ mean of the z-distribution is ALWAYS 0, standard deviation of z-distribution is ALWAYS 1
→ sampling error distributions are normal curves

critical values (for rejecting Ho)

test statistics that are unlikely to be obtained if Ho is true. this might be a mean, or things we haven’t talked about yet life t-scores

alpha level

probability value used to define “unlikely” score if Ho is true

alpha (a)

significance level (commonly .05)
probability of making a type 1 error

type I error

you reject Ho when it’s true
you conclude that the study has significant results when it actually doesn’t

type II error

fail to reject Ho, when it was false
you conclude that the study didn’t have significant results when it really did

beta (b)

probability of making a type II error

independent variable (IV)

manipulated by experimenter

conditions

levels of an independent variable

dependent variable (DV)

measured, outcome variable

control variable

any variable that an experimenter holds constant

confounding variable

a variable that varies with the IV, and could hide it’s effect

why experiments support casual claims

establish covariance
→ is the cause variable related to the effect variable
establish temporal precedence
→ does the cause variable come before the effect variable
establish internal validity
→ are the alternative explanations for the results

comparison group

comparison condition

control groups

no treatment condition

treatment group(s)

one or more treatment conditions

placebo group

placebo control group

design confound

a flaw in how an experiment is set up where a hidden, extra factor changes along with the main variable you are testing, making it impossible to know what actually caused the results

selection effect

participants in one level of the IV are systematically different than the participants in the other level (or levels) of the iv
→ avoid selection effects with random assignment

matched groups

matched based on category, to even out groups

raw score

observed score, x

positive z-scores

above mean

negative z-scores

below mean

within-group design

each participant is presented with all levels of the IV

repeated-measures design

a type of within-groups design in which participants are measured on the DV more than once (after exposure to each level of the IV)

concurrent-measures design

not covered

advantages of within groups design

1) particpants in groups are equal because they are the same participants and serve as the same participants and serve as their own controls
2) within-groups design require fewer participants than other designs

order effects

a confounding effect caused experiencing one condition than another

→ avoiding order effects by counter balancing

disadvantages of within-groups designs

1) potential for order effects
→ solution: counterbalancing
2) might not be practical/possible for research question
3) experiencing all levels of the IV changes the way participants act (demand characteristics)

construct validity

how well were all the variables measured / manipulated

external validity

to whom or what can the causal claims generalize

statistical validity

how much? how precise? what else is known?

internal validity

are there alternative explanations for the results?

the three fundamental questions about internal validity

1) any design confounds?
2) if independent-design group is used, did they control for the selection effects
3) order effects? within-groups

t-tests for independent means (aka independent samples)

comparing the groups of scores from two independent samples
→ mean difference between groups
→ groups are independent
→ only 2 groups
→ same dependent variables is measured in each group

anova

analysis of variance
focusing on one-way ANOVA (simple ANOVA)

one-way anova

inferential statistic
determining the significance of the difference between two or more independent means
one independent variable
→ two or more groups
→ independent
→ same dependent variable is measured in each group
compares the variance between the groups with the variance within groups
variance between means

F

large when mean difference between groups is large
variance between treatments/variance within treatments
treatment effects + random error/random error
“omnibus test”

T

difference between z means

homogeneity of variance

each level of IV has similar amounts of within-group variance
observations in each sample are independent of each other
data is randomly sampled
→ no systematic bais

the Belmont report

IRBS evaluate the ethics of human subjects research with guidance from the key ideas outlined in the Belmont report
product of an intense four day discussion that was held of the Smithsonian Institution’s Belmont Conference Center in 1976

respect for persons

individual’s should be treated as autonomous agents
→ diminished autonomy > entitled to protection
→ privacy

beneficence

maximize possible benefits and minimize possible harms
→ risk/benefit analysis
mental and physical stress

justice

fair distribution of research benefits and burdens
→ research shouldn't be limited to one population

autonimity

participants identifying information is never shared
→ researcher doesn’t record identifying information

confidentiality

individual identifying information is not published
→ records, but protects information
limits to confidentiality
→ danger to participants or others
→ mandated reported issues (abuse)

deception

deliberately providing incorrect information
→ no reasonable alternative
→ drawback: may increase suspicion about research broadly

debriefing

informing participation about full nature and rational of study
→ reverse any negative influence of participation

physical stress

mild electric shock, cold presser test