Research Methods Exam #3

sample

What is subset of a population from which data is collected?

population

What is the entire group of interest?

sampling

What is the process by which a sample is selected from a population?

some populations are too large to study, and need a representative group so one can generalize

What are two reasons that we may use a sample?

Probability sample

What type of sample is selected in a way that the likelihood of participants being selected can be determined?

Non-probability

What type of sample can the probability of individuals being selected not be determined?

base sample on research question and implications of interpretations

What are two things that one should contemplate when deciding a sample type?

single random sampling

What is sampling in which where every member of a population has an equal opportunity to be selected?

obtain sampling frame and select sample randomly

Outline the procedure for single random sampling?

a complete list of members of a population

What is a sampling frame?

should provide a representative sample, may not be able to establish sampling frame due to population size

What are the pros and cons of single random sampling?

selecting every nth person

What is systematic sampling?

no

Is systematic sampling random?

can be done without knowing the sampling frame, produces a sample that is equally as likely as any other sample and does not produce a random sample, may not be representative

What are the pros and cons to systemic sampling?

some combinations are not possible, thus it is not a random sample

Systematic sampling produces a sample that is equally as likely, but?

stratified random sampling

What type of sample, is a variation of SRS that involves first splitting population into strata (aka, groups) and then randomly selecting?

end up with a representative sample

What is the goal of stratified random sampling?

should provide with a representative sample and strata, and one will still need a sampling frame

What are pros and cons of stratified sampling?

Cluster sampling

What type of sampling uses a sampling frame of groups rather than individuals?

data that is nested

Wha type of data is cluster sampling good for?

random selection of groups and random selection of individuals

What are two stages for cluster sampling

easier to attain sampling frame and contact/survey individuals

What are two advantages to cluster sampling?

state, county, district, school, and student

What is a common chain example of cluster sampling?

response rate

What is the percentage of survey returned relative to those sent out?

probability sampling strategy and adaquate response rate

What are two things that representative samples are composed of?

remind participants, test for systematic differences, and evaluation

What are three ways that one can improve response rates?

non-probability sampling

What type of sample's probability of individuals being selected is unknown?

Generalization

What does nonprobability sampling limit?

expensive, time-consuming, and not always needed

What are three of the most major issues with probability sampling?

convenience sampling

What type of NP sampling takes place because people are easy to recruit?

purposive sampling

What type of NP sampling takes place because researcher select representative participants?

sampling error

What type of error takes place because there is a disagreement between sample estimates and population data?

estimation of error or margin of error

What is the concept that measures how much effect does sampling error have on data? There are two names?

larger

What size of samples tend to produce less error in our estimate?

cross-sectional, successive independent sample, and longitudinal

What are three types of research designs discussed in lecture?

cross-sectional

What type of research design entail an assessment that occurs at one point in time, tells us about group differences and the prevalence of a construct, but does not have causal direction?

quick, no follow up, gather important information and all variables are assessed at the same time

What are the pros and cons of cross-sectional research?

successive independent samples

What type of research design occurs at multiple points in time with different groups, tells us about group differences and prevalence of constructs at different times, but does not tell us why there are changes over time

can look at change over time, use new samples and the samples do change

What are the pros and cons of successive independent samples?

longitudinal

What type of research design occurs over time with the same groups, and tells us about group differences, prevalence of construct, and changes over time?

can look at change over time and people may not follow-up

What are the pros and cons of longitudinal research?

attrition

What is the term used to describe the concept that people will drop out of a study, and will not finish?

random and systematic

What are two types of missing values that come with attrition?

complete, and compare it to non-completers

In order to deal with attrition, researchers can create a new variable called \------, and compared it how?

Mean imputation

In order to deal with attrition, researchers can take an average of answered items and insert them for missing values?

answered items

When using mean imputation, one must be careful to only divide by the total of what?

consider all randomized participants regardless of whether they are completers, however one must be conservative

In order to deal with attrition, researchers can use intention to treat analysis, what is this?

last observation

In order to deal with attrition, researchers can impute data by carrying the \---- \------ forward?

Statisitc

What is the "r" value?

tells about the association between two variables

What is the Pearson correlation coefficient?

research design not statistic

What does correlation depend on and not depend on?

direction

The sign on a correlation tells us about the?

strength

The number on a correlation tells us about the?

scatterplot

Often, when looking at a correlation, one will first look at the data on a?

about one individual's score on x and y

Each dot on a scatter plot tells us what?

shows a linear relationship, gives us a beta value, and tells is what the score of y would be when x\=0

What is regression?

the intercept, which is the score of y when x\=0

What is "b" according to y\=mx+b

y\=predicted outcome score, b1\= slope of the line, x\=score of the predictor, and b0\= intercept

What are the names for all the values in regression equation?

b1x1 and b0

In the regression equation, mx becomes? and b becomes?

predict an outcome and establish an association

What are two things that linear regression can do?

fitting a line to data

What is linear regression doing?

Residuals

What is the difference between estimated and observed scores? aka error?

sum of squared residuals

What does linear regression minimize?

stronger relationship

larger "b" values equal what?

coefficeint of determination

What are we interested in when it comes to interpreting r?

squaring r

How do we find the coefficient of determination?

how much systemic variance, and the proportion of variance in one variable accounted for by the other

What does the coefficient of determination tell us?

0%

If all variable are completely unrelated, then how much of the variance in accounted for in x?

100%

If all variables are completely correlated, then how much of the variance is accounted for in x?

increase on y for each 1 unit increase on x in standard deviations

B tells us what?

inability to know which variable came first, and use a cross-lagged design

One problem with correlational research is time-order, what is that? and how would a researcher address it?

cross-lagged

What type of research design helps control for temporal limitations of correlational research?

IV and DV at multiple points overtime, addresses which came first and does not answer causality

What does cross-lagged designs assess and not assess?

assess variable at time one and time two, correlate IV at time 1 with DV at time 2, correlate DV at time 1 with IV at time 2

What is the basic procedure for cross-lagged designs?

argument for a causal relationship

If one correlation is sig. and another is not in a cross-lagged design what does that mean?

there is not an argument for a causal relationship

If one correlation is sig. and another is too in a cross-lagged design what does that mean?

control

What is another problem with correlational research, in which not all variables are controlled for?

partial correlation and multiple regression

What are two things that we can do in order to deal with our c variable (in relation to dealing with control during an a and b relationship)?

multiple and outcome

Researchers will use \---- IV to predict a what?

standard multiple regression

What is the easiest method for conducting multiple regression?

includes all variables at once to get estimate

What procedure does standard multiple regression?

add variables and determines whether these new variables predict any extra variance

What are two thing that other model of multiple regression allows a research to do?

other variables

If IV is correlated with DV after controlling for \---- \---- it's an important factor.

hierarchical multiple regression

What type of regressions stack linear regression on top of each other in specific areas?

controls for variables, makes the case that your IV is important, and can determine unique contributions

What is the benefit for HMR?

Mediation

What are used to determine if IVs have indirect effects on an outcome?

by testing theory, explains how an IV and DV are related and tested using multiple regression.

What does mediation add to research? and how is it tested?

determine if IV and DV are related, if IV and mediator are related, and if mediator and DV are related while controlling for IV

What are the three steps to mediation?

reliability, measure of effect size, can test hypothesis, and parts of causal relationships

What are four things that correlational research provides?

linear relationship

'"r" values and "b" values tell us about?

Any systematic association other than a straight line

What is the definition for a curvilinear relationship?

0

What would "r" equal in a curvilinear relationship?

the likelihood of obtaining our r value if r \= 0.00

What does statistical significance tell us?

sample size, magnitude, and alpha (cut off)

What are three things that influence significance?

our cut off, and our willingness to make a Type I error

What does are alpha tell us?

easier and harder

Making a directional hypothesis makes it \---- and \---- to find a significant result?

r may be influenced artificially for various reasons

What is the idea of confounding?

restricted range, preventing restricted range, and inreliable measures

What are three things that will artificially influence r? Otherwise known as confounding?

outliners

What are extreme scores (usually 3 SD's away from the M)?

inflate r and decrease r

What do we do if the outlier is on the line and off the line?