Statistics Exam 3

PSYC 211

What is descriptive research?

• summarizes and describes behavior

• describes the prevalence of a behavior

• cannot determine causality

• only one variable is being measured

What is correlational research?

• discovers the relationship between two variables

• cannot determine causality

• no manipulation of an independent variable

• cannot rule out confounding variables

What is experimental research?

• how one variable affects another variable

• manipulation of an independent variable

• random assignments to group

• attempt to control confounding variables

What is quasi-experimental research?

• how one variable affects another when random assignment is not possible

• can provide low-moderate support for a causal relationship among variables

• passive manipulation of an independent variable, no random assignment

Correlation does not mean causation.

true

When should you use the independent samples t-test?

• one dichotomous independent variable

• one continuous dependent variable

• between-subjects design

What is the purpose of independent samples t-test?

to determine whether the means of two groups are different than one another (i.e., are the means of two groups significantly different)

What is a null hypothesis?

when two groups are not different from one another

What is an alternative hypothesis (non-directional)?

when two groups are different from one another

What results in a higher t-score?

a larger difference between groups

What results in a lower t-score?

a greater deal of variability between groups (i.e., less likely to find significant difference)

What does degrees of freedom tell us?

how many values are free to vary

What is the critical t-value?

the cutoff t-value

What do you use to find the critical value of t?

alpha (.05) and degrees of freedom

When do you reject the null hypothesis?

when there is less than a 5% chance that the null hypothesis is true (i.e., difference is not due to chance)

When do you fail to reject the null hypothesis?

when there is a greater than 5% chance that the null hypothesis is true (i.e., difference is due to chance)

What is Cohen’s d?

standardized difference between two means

How do you calculate Cohen’s d?

by dividing the differences between the means by the pooled standard deviation

What does correlation test?

covariance between variables

What is required to establish causality?

a true independent variable

What is an independent variable?

a variable that the experimenter manipulates (i.e., a variable that is changed, manipulated, or controlled in a study to see how it influences other variables)

What does a study need to be considered experimental?

at least one independent variable

What is a potential problem when evaluating the independent variable?

faulty manipulation may result in failed experiment (i.e., did we actually manipulate what we thought we were manipulating?)

What is a dependent variable (or measured variable)?

an outcome variable 

You may observe different values on a dependent variable depending on levels of an independent variable.

true

How many dependent variables must an experiment have?

at least one

What is a confounding variable?

a variable other than the iv that differs systematically between conditions

How does a confounding variable invalidate an experiment?

it is unclear whether differences are due to iv or confound

Should confounding variables be eliminated or restricted as much as possible?

yes

What are some common confounding variables?

demographic variables, individual differences, individual habits, and external factors

What are subject variables (or individual differences)?

characteristics of the participant (e.g., age, race, hobbies, income, sex)

Are subject variables true independent variables?

no, but they can be treated the same during an analysis

What does correlation do?

measures and describes the relationship between two variables

When conducting a correlation, what are we interested in?

the covariance of two variables

What is covariance?

how do variables change (vary) in relation to one another

Direction of positive correlation

two variables tend to change in the same direction (i.e., as variable X increases, variable Y also increases & vice versa)

Direction of negative correlation

two variables tend to change in the opposite direction (i.e., as variable X increases, variable Y decreases & vice versa)

What does a sign represent in correlation coefficient?

direction of relationships (e.g., -1 to 1)

What is the magnitude (strength) of a relationship determined by?

the absolute value of the correlation coefficient (e.g., value of 0 is no relationship, value of -1 or 1 is perfect correlation)

How do lines impact the strength of the relationship?

the closer (or straighter) the line the stronger the correlation 

What does statistical significance tell us?

• if there is no correlation between x and y in the population, what is the probability of getting a correlation as large or larger than what we found in our sample

• chance of finding the correlation when the null hypothesis is true

• rejecting the null hypothesis when it’s true (aka type 1 error)

• the probability of finding correlation

What can you use to determine statistical significance?

a critical value table of r

Aspects of restricted range

• interpret with caution

• may differ from full-range correlation

• don’t generalize beyond sample range

What are outliers?

extreme x and/or y values

What can outliers do to correlation?

may inflate or deflate correlation

Correlation is necessary for causation, but not sufficient

true

What is causation determined by?

experimental methodology, not the analysis used

What is the criteria for inferring causation?

• correlation (two variables must vary together)

• temporal precedence (iv before dv)

• no plausible alternative explanations (manipulation of a iv, random assignment to groups)

What can correlations be used to measure?

• validity

• reliability

• prediction

What is pearson’s correlation coefficient?

the most commonly used correlation coefficient (r)

Aspects of Pearson’s correlation coefficient

• both variables must be continuous

  • at least interval scale of measurement

• data must be linear

• pearson’s r is strongly impacted by outliers

Aspects of Spearman’s correlation coefficient

• uses pearson’s formula

  • scores are converted to ranks

• measures relationship consistency independent of form

• ordinal, interval, and ratio scales

• nonlinear relationships

Aspects of Point-Biserial correlation coefficient

• uses pearson's formula

• one continuous variable 

• one dichotomous variable

What is the formula for degrees of freedom?

df = n1 + n2 - 2

What is the p-value?

the probability of something occurring due to chance

Aspects of an experimental group

• exposed to iv manipulation

• can have multiple experimental groups for different iv levels

Aspects of a control group

• not exposed to the iv manipulation

• a “baseline”

When may there be no “groups”?

if the iv is continuous

We want groups to be equal on all variables except?

the iv

What is simple random assignment?

equal chance of being assigned to any group/condition

Aspects of matched/stratified random assignment

• participants are matched into subcategories

• participants within a subcategory are randomly assigned to conditions

• deliberate effort to make sample groups representative of target population

Each participant experiences only ___ level of the independent variable in a between-subject design

one

What is necessary for a between-subject design to be effective?

random assignment

All participants experience ___ levels of the independent variable in within-subject design

all

What are the advantages of within-subject design?

• requires fewer participants

• more statistical power

  • ability to detect the effect of an iv if there is one

• reduces effects of subject variables

What are the disadvantages of within-subject design?

• order effects

  • carryover

  • practice

  • sensitization

  • fatigue

What is carryover?

effect of one iv level is still present when the next level is tested

What are practice effects?

performance improves with repeated testing (e.g., sat)

What are fatigue effects?

latter-session performance is worse because the participant gets tired or bored

What is sensitization?

participant realizes what your hypothesis is

Aspects of internal validity

• determines degree to which we can draw accurate conclusions about effects of iv

• highest when all confounds are eliminated

• achieved through experimental control and sound research design

What is differential attrition?

different numbers or types of people drop out from each group

What is demand characteristics?

participants may perform in the way they believe the experimenter wants them to

What is placebo effect

change as a result of the mere suggestion of change

Aspects of external validity

• degree to which a study’s results would hold true in other situations

• the more internally valid a study is (greater experimental control), the less likely it will be externally valid (generalizable to the real world)