HAN 251 Final
Proportional Stratified Sampling
A population with different strata or subgroups, each of which has different sizes. Sample becomes representative of larger population. Probability Sampling. Unequal groups → pull a sample from larger population to do study with.
Grounded Theory
Approach for developing a theory. Substantive theory grounded in the experiences and views of participants. More participants: 20-50. Uses coding. Qualitative.
standard deviation
Descriptive Statistics. Tells how far away the values are from the mean.
Descriptive research
Developmental (cross-sectional and longitudinal designs), observational, correlational, survey. Non-experimental designs.
mixed methods
Develops rationale for mixing methodologies → then employs the practices of both qualitative and quantitative based on the rationale, combines both QUAN and QUAL
Case control study
Exposure vs. non-exposure. Follow 2 groups, follow rates of disease in two groups. Look for incidence of disease. See if variable increases the incidence of disease. Starts with disease and looks for variable.
explanatory design
First Quan, then follow up with Qual to explain the Quan results. Quan then Qual
Case Study Design
In-depth look at a particular program, event, or individual. Studied for a prescribed period of time to understand, inform practice, or illustrate how something changes as the result of a new event. Data collection is extensive / includes multiple strategies & data sets. Done to inform practice & illustrate how things change. Qualitative research design.
Ethnography
In-depth look at entire group to describe a culture or an aspect of a culture. Goal is that over time the researcher will develop a detailed portrait of the group's shared culture, behaviors, beliefs, and language. Study all forms of communication (verbal, nonverbal, symbolic) to understand culture. Understand day-to-day behavior. Study culture from subject's point of view to gain a subjective understanding of behavior. When you want to look at a culture group and look at shared implicit and explicit communication and patterns. Qualitative research design.
cluster sampling
Pulling intact similar groups → if they are all the same, you randomly select a few to study. Use when population is spread across wide area and it may not be feasible to study everyone. All clusters are equal, pull random cluster to study. Probability Sampling.
convergent design
Quan and qual are equal and happen at the same time. Survey also includes open ended questions - given at the same time and analyzed at the same time. Look for similarities between the two.
random sampling
Simple random, stratified random, proportional random, cluster, systematic
exploratory design
Starts with Qual then use those results to develop a Quan instrument/tool (ex: survey). Explore Qual to make a better Quan. Qual then Quan
Phenomenological
The lived experiences. Study of a multiple perceptions and perspectives of the meaning of an event. Relies heavily on interviews with a carefully selected group of people (5 to 25 different people). Interviews are unstructured and all of the participants drive the process. Researchers must suspend preconceived notions or personal interpretations (called bracketing). Focuses on common and divergent experiences of a phenomena or experience. What is it like to be you after experiencing it? Qualitative research design.
placebo effect
a beneficial effect, produced by a placebo drug or treatment, that cannot be attributed to the properties of the placebo itself, and must therefore be due to the patient's belief in that treatment.
theoretical sampling
a grounded theory, qual sampling method. They choose data sources that are most likely to help them develop a theory of the process in question.
Stratified Random Sampling
a population with different strata, or subgroups, each of which are roughly equal in size. Identifies strata or subgroups (gender, education level, religion, etc.) then random selection. Strata are chosen to divide a population into important categories relevant to the research interest. Probability Sampling.
focus group
a qual researcher can interview several participants simultaneously. No more than 10-12 participants. Has a moderator. Group discusses a particular issue for 1 to 2 hours.
cohort-sequential study
a researcher beings with 2 or more age groups (cross-sectional piece) and follows each age group over a period of time (longitudinal piece).
factorial designs
a researcher examines the effects of 2 or more independent variables in a single study.
ex post facto designs
a researcher identifies events that have already occurred or conditions that are already present and then collects data to investigate a possible relationship between these facts and subsequent characteristics or behaviors. Investigates extent to which specific independent variables may possibly effect dependent variables.
statistical significance
a result that is deemed not due to chance and we can reject the null hypothesis. A finding is this when it allows the researcher to reject the null hypothesis at a pre-specified level of confidence or probability.
Longitudinal design
a single group is followed over time (months and/or years), many data collection points. Following over time → see how people change over time. Cannot determine causality → no cause and effect determination. Enroll more people. Can't quickly test new hypotheses. Long time before results are obtained. Follow-up is key on a specific time interval is essential in this research. "Loss to follow-up" is a problem due to death, move, etc. Non-experimental.
triangulation
a strategy used by qual researchers. They collect multiple forms of data related to the same research question, with the goal of finding consistencies or inconsistencies among data.
structured questions
a survey question that provides answers and a respondent must pick one
systematic review
a viable alternative for qual and mixed method studies, in which research reports rather than individual people or other individual entities, are the objects of the study. A researcher implements an explicit, rigorous method for selecting and analyzing the reports.
mean
adding all of the scores and dividing by total number in the group.
inferential statistics
allow inferences and correlations to be made from the sample to population. Parametric data (interval and ratio) - significant difference (t-test) and correlations (ANOVA, ANCOVA, MANOVA, Regression, Multiple regression, Pearson r). Nonparametric data (nominal and ordinal) - significant difference: pearson's chi-squared, etc.
descriptive statistics
allows researchers to describe, sample, organize, and summarize data. Frequencies, percentages, description of central tendency (mean, median, mode), description of relative position (range, standard deviation).
meta-analysis
an analysis of the analysis. The researcher combines the results of many experimental and/or ex post facto studies to determine whether they lead to consistent conclusions. once researchers have conducted many replications of their research, another researcher may conduct this.
Percentage of agreement
another coder(s) use coding structure and definitions and code a randomly selected portion of the data. Team of coders code the data independently and meet to discuss until they reach an agreement. Discuss why until mutual agreement occurs between the coders (what we did in class).
random sampling
any differences between the groups are small and only due to chance; everyone in the study has equal chance of treatment or control group.
survey
any measurement procedures that involve asking questions of respondents.
focus groups
are a way to listen to people and see how interactions between people impact their answers. Should not have more than 8-12 people. Props (video clips, products, etc.) can be used to generate discussions. Length varies between 90-120 minutes. A problem: one person can dominate.
low response rates
are among the most difficult of problems in survey research
Internal consistency reliability
are questions designed to evaluate the same concept/obtain equivalent responses? The extent to which all of the items within a single instrument yield similar results. Asking same questions/content - sees whether you answer the same.
construct validity
are the hypotheses concerning the relationships between the underlying concepts borne out by the responses? The extent to which an instrument measures a characteristic that cannot be directly observed but is assumed to be true. Measuring something assumed to be existent.
Probability sampling
attempts to specify a random sample from segments of a larger population that will be representative of the whole.
loaded
biased questions that use words that cue a certain response.
embedded design
both quan and qual are collected in same general time frame. However, one general approach dominates, with the other approach serving a secondary, supplemental role. Mixed method design.
Observational Design
can provide QUAN alternative to QUAL (grounded theory, ethnography) approaches. Pre-specified focus to observations, underlying theoretical framework. Go in & watch to collect data. Interrater reliability is very important → researchers/observers need to be trained to observe something. Operational define what is being observed in order to count or evaluate observation. Non-experimental.
non-probability sampling
convenience, quota, purposive, snowball
Emergent design
data collected early in the investigation influences the kinds of data the researcher subsequently gathers. Qual design.
Internal validity
degree to which extraneous variables have been controlled; experimental effects can be attributed solely to treatment/intervention.
practical significance
degree to which findings related to a research intervention are useful in real-world procedures and practices; often involves a cost-benefit analysis.
mean, median, mode
description of central tendency
interrater reliability
do 2 face to face interviewers with the same questionaires get the same response? the extent to which two or more individuals evaluating the same product or performance give identical judgments. Make sure that two researchers see the same exact thing.
content validity
do the questions express the underlying concept they were designed to reflect? The extent to which a measurement instrument is a representative sample of the content area being measured. When a test has this, the items on the test represent the entire range of possible items the test should cover. Are you hitting all areas that are important to measure?
face validity
do the questions make sense? The extent to which an instrument looks like it's measuring a particular characteristic; relies on subjective judgment. Does it look like it measures what it's supposed to measure? Your subjective judgment - Are questions asking what they are want answered.
criterion validity
do the responses to the questions agree with an objective criterion or gold standard for the underlying concepts? The extent to which the results of an assessment correlate with another, related measure. Correlates with another established, validated measure.
Test-retest reliability
does the same question have the same response over time? The extent to which a single instrument yields the same results for the same people on two different occasions.
survey research
encompasses any measurement procedures that involve asking questions of respondents. Can collect data about: knowledge, attitudes, feelings, perceptions, beliefs, behaviors. Can correlate variables but the main focus is on learning about a population. Any measurement involving asking questions. Can look at a variable of interest. Can pull out variables but is not correlational research. Help you to understand the population.
Simple Random Sampling
everyone in population has equal chance of being selected. People on master list could be selected or not selected. Probability Sampling.
Descriptive research
examines a situation as it is. Does not involve changing or modifying the situation under investigation. Does not determine cause and effect relationships.
Correlational design
examines the extent (strength and direction, degree of association) of relationship between characteristics or variables within a group or between two or more groups. Looks for strengths of relationships. Explains/predicts why something happens. Statistical test. Interpret results with caution. Non-experimental.
credibility/trustworthiness
extent to which other individuals perceive the study's findings to be convincing and worth taking seriously.
internal validity
extent to which the study enables defensible conclusions about cause-and-effect and other between-variable relationships.
external validity
extent to which the study's results can be generalized to a larger population or broader context.
Qualitative
generally begin with statements like how, what, why.
placebo
given to control group, might on the surface appear to be influential but in reality should not be.
unbiased survey question
gives you acceptable, not acceptable, and is NOT a moral issue. Get to answer yes, no, or neutral.
discriminant sampling
grounded theory researchers may employ this, returning to particular data sources that can help them substantiate their theory.
screening/filter question
helps you get population you want. May bring respondent to a different set of questions or thank them for their participation and time (their survey is complete).
sampling
how participants are selected. The act, process, or technique of selecting a representative part of a population for the purpose of determining parameters or characteristics of the whole population.
constant comparative method
in Qual research, the methodology that often involves an interative process in which the researcher moves back and forth between data collection and data analysis.
multi-phase interative design
includes 3 or more phases with early ones providing foundational data on which later phases can build. Is interative because researcher moves back and forth between quan and qual methods, with each new body of data informing the conceptualized and implementation of subsequent phases.
Parametric data
interval and ratio, significant difference (t-test) and correlations (ANOVA, ANCOVA, MANOVA, Regression, Multiple regression, Pearson r).
structured interview
interview in which a researcher predetermines all of the questions to be asked.
interview distortion
interviewer may ask a question in a way that distorts the meaning of a question, may make the respondent feel uncomfortable, may not listen carefully to respondents on topics for which they have strong opinions, and may make judgements that they already know what the respondent would say to a question based on their prior responses, even though that may not be true. Interviewer can affect it, can skew it and shouldn't bias an interview.
effect size
is the statistical measure of the strength of an observed difference between groups.
experimental research
manipulates variables. Measure the effects of this manipulation on other variables. Data analysis also calculates "correlations" between variables, specifically, those manipulated and those affected by the manipulation. Has the power to prove causality.
Convergent, Explanatory, Exploratory
mixed methods designs
Non-probability Sampling
no way to guarantee that subgroups of a population will be included in the sample. You can't prove something, NO random sampling.
Nonparametric data
nominal and ordinal, significant difference: pearson's chi-squared, etc.
Cohort studies
observational studies that aid in evaluating associations between diseases and exposures.
participant
observer is a member of the group.
nonreactive
observer is removed from participants.
Ordinal
one object is bigger or better or more of anything than another. Rank order. Ex: rank the following exercises in order of preference from best to worst.
Nominal
one object is different from another; the number next to each response has no meaning except as a placeholder for that response. Ex: what is your occupation? Choices: Physician, DO, nurse practitioner, physician assistant, occupational therapist.
interval
one object is so many units more than the other.
experimental
only study design where you can say that you've proved something. Random sampling methods ONLY.
matched pairs
pairs of people (one in each group of 2) being compared who are identical or very similar with respect to characteristics that might possibly have an effect on the dependent variable.
cross-sectional design
people from different age groups are sampled, data collected once and compared. To see if there is a relationship. No causal relationship, just see how things change. Data collected in one instance. Non-experimental.
reactivity/hawthorne effect
people sometimes show improved performance simply because they know they're participating in research.
response rates
percentages of people agreeing to participate.
Quota
predetermined categories. Selections based on predetermined categories until filled. Ex: put up flyers, first 11 people get in. 11 females ages 36-40; 11 females ages 41-45. Not random but sample is stratified well. Non-probability sampling.
Retrospective cohort study
previously collected data is reviewed to identify the population and the expusre status (exposed/not exposed groups). Determines at present the (development) status of disease. Look back at data you already have. → study was not preplanned. The investigator has to go back to pre-existing data.
Phenomenological, Ethnography, Case Study Design, Grounded Theory
qualitative research designs
Sensitive
questions asking respondents about difficult or uncomfortable subjects. It helps to develop some trust or rapport with the respondent first.
Double-barreled
questions that ask more than one thing. You can often spot these kinds of problems by looking for the conjunction of "and" in the question. Should be split into 2 separate questions. Ex: what are your feelings towards using condoms and promoting abstinence only for the reduction of HIV transmission?
Convenience Sampling
readily available participants. Ease in selection, like a class. People available to you, no master list. Non-probability sampling.
clinical significance
refers to the practical, applied value, or importance of the effect of the intervention. Whether the intervention makes a real difference in everyday life (genuine, palpable, practical, noticeable). Are treated individuals as a group indistinguishable from non-treated with respect to the primary complaints following treatment?
Local Groundedness
researchers will collect proximity data before the study is underway, look at the setting beforehand. Qualitative.
Guttman cumulative scale
respondent checks each item with which they agree. The items themselves are constructed so that they are cumulative, ranked ordered in difficulty from least to most extreme or most to least extreme. If you agree with a statement, you agree with the statements above it as well. Bogardus social distance scale is an example of this.
single-subject designs
reversal, alternating treatments and multiple base -line designs can be sued to study not only with groups, but also with single individuals.
Qualitative
seeks saturation
Quantiative
seeks statistical validity
Purposive
select participants due to known similarities to target population. Knowledge of group is required before selection. Risk of researcher selection bias. Qualitative. Non-probability sampling.
systematic sampling
selecting individuals, or clusters, according to a predetermined sequence, which originates by chance. Ex: every 10 people off the list chosen. Probability Sampling.
reflexivity
self-reflection to actively try and identify personal, social, political, or philosophical biases that are likely to affect the researchers' ability to collect and interpret data.
baseline data
sequence of observations made prior to the treatment.
Naturalistic settings
spend bulk of time in the field observing and interviewing subjects in their own environment. Both data collection process and the outcome of data analysis is vital to the study. Qualitative.
Prospective cohort study
study begins w/ the identification of a population and exposure status (exposed/not exposed groups). Population is followed over a period of time for the development of the disease. Follow over a period of time → follow forward. The study is planned and designed to answer questions in a specific area. Non-diseased subjects meeting eligibility criteria are enrolled. Look forward, look to see development of disease over time.
Developmental
study of characteristics / variables develop or change over time. Non-experimental.
hawthorne effect
the alteration of behavior by the subjects of a study due to their awareness of being observed.
Equivalent forms reliability
the extent to which two different versions (Form A, Form B) of the same instrument yield similar results.
median
the midpoint among all of the scores (might not be a whole number).
mode
the most commonly occurring score. There might be more than 1 (bimodal). Ex: in a sample of 8, 4 people score a 100 and 4 score a 90.
Triangulation
the purpose is to confirm information about a phenomenon and to obtain convergent validity → confidence the finding is valid since it was verified by 2 different methods.
structured interview
the researcher asks certain questions and nothing more.
semi-structured interview
the researcher may follow the standard questions with one or more individually tailored questions to get clarifications or probe a person's reasoning.
combined experimental and ex post facto design
to study the possible effects of an experimenter-manipulated variable, a previously existing condition, and the interaction between the two.
prospective and retrospective
two types of cohort studies
Quantitative
typically ask things like: what is the relationship...to what extent...what is the difference...how much...was a significant difference found...impact.
likert scale
used to quantitatively assess abstract concepts, attitudes, behaviors, etc. Traditional 1-to-5 rating. Ask an opinion question on a 1-5 bipolar scale with a neutral point and the 2 ends are at opposite positions of the opinion. Ex: please indicate your agreement with the following statement.
correlational research
usually does not influence any variables. Measure and look for relations between some set of variables. Data analysis looks at the strength and direction of the relationship. Cannot prove causality. Main focus is determining how the variables are related or how variables can predict outcomes.
Member checking
went back to participants and make sure codes / themes are correct or not. Ask them if they are correct.
dichotomous
when a question has 2 possible responses. yes/no, true/false, agree/disagree. Ex: Are you male or female?
unstructured questions
when a respondent is not given answer choices. Ex: anything else you want us to know?
Biased survey question
when a respondent is not given the option to say that they do not agree.
Snowball
when it is hard to identify/find participants. Ask participants to help find participants, people need to be comfortable w/ researcher in order to refer people. Non-probability sampling.
return rate
when questions are distributed by mail or e-mail, the majority of people who receive questionnaires don't return them.
false respondents
with mail and phone surveys, is the person you are speaking to actually who they say they are?