Rsh 1 Exam 2 Flashcards
experiment studies
1 group is the experimental, one is the control, this includes random selection and random assignment into each group
quasi-experimental studies
1 group that fits into 1 category of the study and another that does not because we cannot random assign people who are not that specific category to be in that category
Non-experimental studies
independent variable is not being manipulated, based on observation and measuring instead of experimenting on randomly assigned subjects
self-report
what people tell you about themselves
behavioral
watching how people act or measuring how people act, measuring how people feel about something
reliability
answering questions the same way
internal consistency
how well items within the same scale relate to one another, 3 construct items should have answers that relate to each other based on the person answering them
test-retest
does our test produce similar results across different time points?
same test 2 different time points
alternate form
do different versions of our measure produce similar results?
different version of the same test either given to the same group or a different group
validity
are we measuring what we believe we are measuring?
content validity
does our study cover an appropriate breadth of information?
construct validity
are we measuring the intended construct?
are the pieces really a part of what we think we are studying?
is this the construct of what we want to measure really what we want to measure?
convergent validity
do scores on our measure correlate with similar variables?
number of times you use something should be related to how useful that thing is
discriminant validity
refers to the degree to which a measure or construct is distinct or different from other related measures
open-ended
participant is at liberty to answer how they would like, researcher gives no options
close-ended
researcher provides limited options to an item, usually yes or no
forced choice
researcher gives options and participant chooses which option is how they would like to respond
Likert scale
a point scale to measure people’s attitudes or opinions through a range of options on how strongly they agree or disagree with a statement
how to write proper survey questions and the problems that a well-designed survey can help avoid
- should be presented in statement form
- answerable on 5-point scale
- written so they are consistent with response choices
- clearly written, avoid ambiguous language, do not ask questions
- referring to the past few years
- do not ask 2 questions in one item
- grammatically correct
- no leading questions
know the basics of creating subscales in SPSS
creating subscales to test the reliability of the items and the validity of the subscale
reverse coding questions
questions are coded reversed, 1 becomes 5 and so on
test-retest
does our measure produce similar results across different points?
same test 2 different time points
alternate form
do different versions of our measure similar results?
Different version of the same test either given to the same group or a different group
mean
average
median
the middle
mode
most frequent
normal and skewed distributions
normal are when the answers the same
skewed are when there are outliers
variability
how much the responses to the items varies, measured by standard deviation and variance
central tendencies
mean, median, mode
probability sampling
everyone in the population has an equal chance of being sampled, unbiased sample
non-probability sampling
everyone does not have an equal chance of being sampled, biased sample
simple random sampling
probability sampling method which a subset of individuals is randomly selected from population members, need access to full population
stratified random sampling
includes dividing population into subpopulations that specific charactertistics in common and taking random samples form these subpopulations for data
cluster random sampling
dividing total population into groups or clusters and then using simple random sampling to select which clusters participate, all observation in a selected cluster is included in the sample
convenience sampling
ask people who are convenient, nonrandom selection of participants who are readily available to the researcher to serve as the sample
quota sampling
freely choosing any participant as long as they meet an already established quota
purposive sampling
choosing sample based on who the researcher thinks would be appropriate for the study, used when limited number of people have expertise in the areas under investigation
snowball sampling
existing study participants recruit future participants from among their acquaintances
nonresponse bias
potential systematic difference between those who refused to participate in a study and those who participated
correlation
particular type of statistical test we do to measure relation between 2 variables
how do you interpret an r statistic?
tells us degrees of freedom, r(N-2)
how do you report an r statistic?
ex: with an n of 120, r(118) = .68, p<.001
Understand the steps, why we are doing it, be able to interpret provided SPSS reliability output
we are doing it to make sure our reliability is a .7 or higher because if not then our study is not reliable and if it is then our study is reliable and we can proceed
Cronbach’s alpha interpretation, understanding the alpha if item deleted stats
- Cronbach’s alpha is testing how good a question is and if it is a reliable question that is benefitting the research
- an item should be deleted if it is decreasing the reliability of the research, if the research has a higher Cronbach’s alpha if that question is deleted then it should be deleted in order to make the research more reliable
what was the point of the M&M activity?
The point of the activity was to figure out how many of each color M&Ms came in a pack. We wanted to see if the numbers M&M provided were true. We started with small individual samples then started with group samples and finally did an entire class sample. The entire class sample was more accurate when compared ott eh numbers provided by M&M.
what was the take home message of the M&M activity?
The take home message was that the larger the sample, the more accurate the results. The better we can generalize our findings.
what were the takeaway messages from the lecture provided by the Dr. Carolyn Lau?
Why survey research
- Purpose
- Help strategize messages and actions for optimal impact
- Help us understand who we are and how our attitudes and behaviors are impacted by or can have an impact on the world around us
- Who does survey research?
- Marketers, policy makers, human resources, institutions, academics, pollsters
- Types of survey research
- Image measurement, product/concept testing, needs assessment, program evaluation, support barometers
- Weighting
- Not everyone sampled responds to survey - skewed sample
- Can be used to rebalance sample to match key population metrics
- Weight = population proportion/sample proportion
- If respondents differ systematically from non-respondents on variables of interest, leads to biased estimates
- Survey research practicum
- Conduct survey project for outside client, such as needs assessment or program evaluation
- Habitat for humanity in Monmouth county
- Boys and girl club of monmouth county - asbury park
- Students guided through entire research consulting process form research design in collab with client, through collecting/analyzing data, finally presenting results
qualitative research
non-numerical, based on opinions and experiences, words, pictures, observations, typically small samples, focus groups, interviews, case studies, data analysis is in depth on fewer observation, focus on individuals responses rather than averages
quantitative research
based on numbers, typically larger sample, experimental (cause and effect relationships), quasi-experimental, correlational, data analysis is descriptive statistics and inferential statistics
pros and cons of behavioral and self-report
advantages of self-report = cheap and easy, provides firsthand info about thoughts and feelings
disadvantages of self-report = accuracy of self-knowledge is questionable, social desirability bias, demand characteristics, retrospective bias
Behavioral is better because there is less bias and people will act normally if they do not know they are being evaluated