Research & Stats EXAM 3 (FINAL)

descriptive statistics

procedures for describing a sample

inferential statistics

procedures for drawing conclusions about a population based on data collected from a sample

z test

test of the null hypothesis for a single sample when the population is known (compare a sample mean to a distribution of sample means)

1. construct HA & H0

2. set decision rule (obtained p must be less than value associated with a p of 0.5)

3. collect data from a sample & calculate test statistic (Z, r, t)

4. apply decision rule (compare obtained probability of stat observed with stat for p< 0.5; if p < 0.5 we reject the null hypothesis)

how do you conduct the 4 steps of hypothesis testing with a Z test?

nondirectional (z test)

does my sample mean differ significantly from the population? (ex: Are IQ scores for students in 3010 different than the population?)

directional (z test)

is my sample mean significantly greater or lesser than the population? (ex: Are IQ scores for students in 3010 higher than the population?)

if p < 0.5, we reject the null hypothesis, thus finding support for our HA

how do you know when to reject or fail to reject the null hypothesis?

Pearson’s r

just like t test, interval/ratio data, H0: population from which the sample was drawn has a correlation coefficient of 0, directional vs non-directional

1) State the null and alternative hypotheses, 2) Calculate the Pearson correlation coefficient (r), 3) Determine the p-value or critical value based on the sample size and significance level, and 4) Compare the p-value (or critical value) to the chosen significance level and make a decision to reject or fail to reject the null hypothesis

how do you conduct the 4 steps of hypothesis testing with a Pearson’s r?

the critical value acts as a cut-off point that separates the rejection region from the area where the null hypothesis is accepted

what is the relationship between critical values & region of rejection?

statistical power

the likelihood of a hypothesis test detecting a true effect if there is one

1. cause must be related to the effect

2. cause must precede the effect

3. no other explanation(s) must exist for the effect

what are the 3 criteria for causation?

independent variable

variable that changes or controls

dependent variable

variable that measures

between-subject desgins

characteristics

• mutually exclusive membership; different subjects assigned to groups

• at least one variable is manipulated: independent variable

• control & experimental groups

• at least one variable is measured: dependent variable

random sampling

refers to how you select individuals from the population to participate in your study

random assignment

refers to how you place those participants into groups

it helps strengthen the internal validity of an experiment and avoid biases

why are random sampling & random assignment important for research?

confounds (inadequate control)

uncontrolled extraneous variable(s) or flaw(s) in the experiment; makes it impossible to say whether changes in the DV are due to the level received in the IV

the temperature in the experimental classroom is higher than the control; the time of day differs in the control & experimental classrooms

example of confounds:

internal validity

the extent to which results of an experiment can be attributed to the manipulation of the IV rather than to confounding variables (the extent to which a research study establishes a trustworthy cause-and-effect relationship)

nonequivalent control group, instrumentation issues, diffusion of treatment, experimenter effects, history, maturation, testing, regression to the mean, subject effects, floor and ceiling effects,

what are common threats to internal validity?

to ensure the validity of the results

what are the reasons why one would use single or double blind procedures?

external validity

refers to how well the outcome of a research study can be expected to apply to other settings

generalization to populations & generalizationn from lab settings

what are common threats to external validity?

1) State the null and alternative hypotheses, 2) Select a significance level, 3) Calculate the test statistic (t-value), and 4) Make a decision based on the p-value, comparing it to the significance level

what are the 4 steps of hypothesis testing with a t test for independent-samples?

t test

we DON’T know the standard deviation for the population, _ distributions don’t fit a normal distribution; cant use the area under the normal curve that applies for z tests

nonequivalent control group

(threat to interval validity) control & experimental groups must be the same in every way at the start

• to minimize: random sampling & random assignment

• ex: self-selection into a study is a big issue; participants vs waiting list studies

instrumentation issues

(threat to interval validity) measuring device is faulty; test may not be reliable

• changes in the DV may be due to measurement issues & not due to changes in your IV

  • to minimize: piloting your instrument to ensure it’s reliable, an equivalent control group helps to ensure your instrument is consistent

diffusion of treatment

(threat to interval validity) information “contamination” from other participants who completed the experiment already

• ex: participants talk with a friend about the experiment, reactivity issues

• to minimize: run large groups in short time frames, asking participants not to talk about the study in debriefing

experimenter effects

(threat to interval validity) bias from experimenter’s expectations

• ex: nicer, more positive feedback to experimental group; changes in DV may be due to experimenter bias not due to changes in your IV

• to minimize: single or double blind experiments

history

(threat to interval validity) an outside event not part of the manipulation could be responsible for results

• ex: doing a stress study during exam week

maturation

(threat to interval validity) naturally occurring changes in subjects could be responsible for results

• maturing physically, cognitively, emotionally

testing

(threat to interval validity) repeated __ could lead to better/worse performance

regression to the mean

(threat to interval validity) extreme scores, upon resting, tend to be less extreme, i.e. they move toward the mean

• some initially high scores may be due to chance/luck, lower scores are more probable

  • ex: sports; after outstanding performance likely to ___

mortality/attrition (dropout)

(threat to interval validity) a problem especially if different rates between groups

• ex: smoking study & heaviest smokers drop out of treatment group

• experimental vs control comparison are no longer meaningful

• to minimize: an equivalent control group allows us to see whether there’s a differential attrition due to our manipulation

floor and ceiling effects

(threat to interval validity) an issue of sensitivity of dependent variable; not sensitive enough & too sensitive

generalization to populations

(threat to external validity) “college sophomore“ problem

• counterarguments:

  • the study’s sound; just needs to be replicated with other populations

  • for some topics, irrelevant if they’re college students

  • college populations are fairly diverse

• to minimize: randomly sampling from entire population, replications

generalization from lab settings

(threat to external validity) counterargument: control is key

• to minimize: replication outside of lab; replicating using various materials