Module 1 o Ordinal – specific order.

Ex: • Pretest-posttest – outcome data at

grades A-F, Income level baseline and after intervention
Types of Quantitative Research: 2. Continuous – infinite values. Ex: • Cross over design – subjects
time/weight exposed to 2+ conditions in
• Experimental (RCT) – Treatment, control, o Interval – numerical value, random, and serve as their own
intervention. Random assignment measured along equidistant control
• Quasi-experimental – may/may not have continuum. Ex: temp, weight, • Possible carryover effects –
control, no randomization credit score exposed to 2 diff treatments. May
• Non-experimental – no manipulation or o Ratio – has a zero value, be influenced in second condition
intervention similar to interval. Ex: dose, by first experience
flow rate, pulse
Types of Qualitative Main Categories of Quasi Experimental
Module 3
• Grounded theory – explain phenomenon • Nonequivalent control group
(interview) Abstact – Introduction – Methods – Results – designs – those getting
• Phenomenology – lived experience Discussion – Limitation – Conclusion intervention are compared with
(interview) nonrandomized comparison group
• Ethnography – understand culture P<0.05 is the cut off for statistical significance for most • Within subjects design – one group
(observational fieldwork) studies. Low p-value, good reliability is studied before and after
intervention
Terms • P=0.05 means 5/100 would be chance
• P=0.01 means 1/100 would be chance Criteria for experimental: intervention, control group,
• IV - Independent variable, random assignment
intervention/manipulation, controlled, Eliminating Bias
input, predictor Define new terms in the chapter
• DV - dependent variable, outcome, 1. Quantitative – have control,
operational, experimental, predicted, randomization, blinding • Attrition – how many people lost in study
response, measured, output 2. Qualitative – reflexivity or reflecting on • Case-control design – intervention grp,
• Operational definition - measurement own perspective/bias case grp, control grp
tool • Causality – 3 characteristics (variables
Terms:
• Conceptual definition – wordy definition must be related, cause → effect,
• PICO - population, intervention, relationship not due to confounding
• Directional hypothesis – when one extraneous variable)
comparison, outcome
goes up, other goes down/up.
• Cohort design – longitudinal study that
Shows expected direction of
Module 2 performs cross-section on a cohort
variable relationship
through time.
• Nondirectional hypothesis – does
Terms • Control– does not receive intervention
not specify direction of relationship
• Cross sectional – not looking back or
• Anonymity – even researcher does not between variables
forward
know which group, double blind. • Null hypothesis - no correlation, no
• Experimental group – receives
• Beneficence – there should be benefit for difference, no relationship
intervention.
participants/ future generation. Action • Reliability – accuracy, consistency
• Correlation – looking at relationship
should promote good, freedom from of information
between concepts
harm, discomfort, exploitation. • Validity – “soundness of evidence”
• Descriptive research
• Conceptual framework – limited to are you measuring what you set
• Intervention – researcher does something
research problem, synthesis of relevant out to measure?
to some subjects, manipulation
concepts. Qualitative/quantitative • Credibility – results are truthful and
• Longitudinal design – over a period of
• Confidentiality – researcher knows which accurate
time
group but does not tell anyone. Right to • Research hypothesis
• Nonequivalent control group design –
keep knowledge or information • Simple hypothesis – predicted
two groups where you compare each
concealed relationship between one IV and
group. Group 1: before and after
• Consent form – physical or digital form one DV
intervention; group 2: no intervention,
showing agreement • Complex hypothesis – predicted
control.
• Informed consent – receive information relationship between 2+ IV and 2+
• Nonexperimental design – does not meet
before agreeing DV
criteria for experimental (intervention,
• Institutional Review Board (IRB) – need to • Research problem – “obesity is
random assignment, control group)
get approval from IRB to be considered as becoming more common”
• Prospective design – looking forward
peer reviewed, to make sure scientifically • Research question – “does eating
• Quasi-experiment – Has intervention but
accurate/high standard, and safety of more sugar increase risk for
no randomization or control group
participants obesity?”
• Randomization –assign at random
• Nonmaleficence – do no harm to patients • Statement of purpose – can be
• Randomized controlled trial (RCT) – true
• Primary source - author himself seen in introduction
experimental
conducted research • Retrospective design – going back
• Secondary source - did not collect data, Module 4
• Hawthorne effect – altered behavior of
but analyzed it subjects due to awareness of being
• Theoretical framework – something that Experimental vs Non-experimental
observed
can guide study. Show relationships •
• Has intervention – experimental Within subjects’ design – when no
between concepts. Mostly quantitative comparison for control is available, look
• No intervention – nonexperimental
• Theory – belief, policy, procedure at interventional group before and after
followed as basis of action intervention
Control Group Conditions
• Vulnerable population – pregnant
women, elderly, institutionalized people,
• Attention control – extra attention Abstract - Intro (problem, significance of prob,
children, limited capability to sign on purpose) – method (research design, sampling
but not active ingredient of
consent form method, instrumentation/operational definition. Stat
intervention
• Delayed treatment – intervention analysis plan, intervention content, ethical
Variables consideration) – result (data)- discussion (based on
given at later date
result, compare w other study similarity and
1. Categorical – categories. Ex: cheerios, differences)
True Experimental Designs
fruit loops, captain crunch are in category
cereal
• Posttest only (after only) –
o Nominal – no rank/order. Ex:
outcome data collected only after
zip code, blood type, gender
intervention
Module 5 Module 6 through process of interest. Recruited to
develop well rounded theory.
Threats to Internal Validity: • Mixed methods research – combines
quantitative and qualitative method.
1. Temporal ambiguity: RCT. In correlational Used when construct is poorly
studies, it may be unclear whether the IV understood. Neither quanti or quali alone
preceded the DV, or vice versa. will answer question.
Correlation =/= causation Terms
2. Selection: biases arise from preexisting Module 8
• Face validity – whether instrument looks
differences in group. When not randomly
appropriate for contruct. Based on
assigned. When participants have Probability Sampling: simple random, stratified
judgment, no objective criteria
motivation to volunteer. random, systematic, cluster
3. History – cooccurring with causal factor • Construct validity – what is instrument
that can affect outcome measuring? Does it adequately measure? Nonprobability – purposive/judgment, quota,
4. Maturation – process that result simple • Content validity – degree to which convenience, snowball
from time. instrument has adequate sample of items
5. Mortality/Attrition – differential loss of for construct being measured. • Convenience sampling – group of people
participants • Criterion validity – degree to which easy to contact
instrument is related to external criterion • Eligibility criteria – like demographics
Threats to External Validity: • Internal consistency – extent to which • Nonprobability sampling – can not
items on instrument are measuring same calculate probability of getting sample.
1. Restricted characteristics of participants attribute. >0.80 are highly desirable. Not for inference from sample to general
2. Too homogenous participants Assessed via CRONBACH’S ALPHA population
3. Enhancing internal validity can have • Interrater reliability – comparing • Power analysis – estimate number of
adverse effects on external validity agreement between observations or sample size required to detect effect
ratings of two or more observers • Probability sampling – choose sample
Terms: • Likert scale - commonly used to measure from larger population using probability.
attitudes, knowledge, perceptions, Must be random
• Attrition – how many people lost in study values, and behavioral change. Similar to • Purposive sampling – chosen on basis of
• Blinding – keep data collectors and others 16 personality quiz researcher’s knowledge and judgment
unaware of group allocation, controls • Measurement – process of observing and • Quota sampling – sister of strat. divide
bias, ensures that others involved in the recording the observations that are into subgroups → “pick whoever is close
study will not influence results collected as part of a research study to you per group.”
• Confounding variable- extraneous • Reliability - the instrument measures the • Sample
variable, influences both IV and DV construct consistently in all populations • Sample size
causing fake association over time regardless of the instrument
• Simple random sampling – randomly
• Constructs – proposed attribute of person administrator
choose participants (names in paper in
that often can not be measured directly • Test-retest reliability - consistency over box)
but can be assessed using a number of time; the assumption is that for traits that
• Stratified random sampling – divide into
indicators or manifest variables. have not changed, any differences in
strata. From proportion determine how
• Construct Validity – is intervention a good people’s score on the two testings are the
many participants per strata.
representation of construct? Does DV result of measurement error
• Systematic sampling – “Systematic
really measure intended constructs? • Validity – the instrument actually
sampling is defined as a probability
Involves inferences from particulars of measures the construct it was designed
sampling method where the researcher
the study to the higher order constructs to measure
chooses elements from a target
they are intended to represent. • Validity coefficient – analyze relationship population by selecting a random starting
• External validity – generalizability (can between scores and criterion >0.70 is point and selects sample members after a
this sample/research be representative of desirable fixed ‘sampling interval’.”
the overall population/situation?) • Visual analogue scale – scale bar where
• History – other events co-occuring with you drag pointer. “no itch to worst
causal factor that can affect outcome imaginable itch”
• Homogeneity – restricting sample. Only
people who are similar with respect to Module 7
confounding variables are included. Very
specific inclusion criteria. Limited Terms:
generalizability.
• Internal validity – inferences can be made • Ethnography – interprets culture/cultural
that the independent variable is truly behavior. Relies on fieldwor
causing the outcome • Phenomenology - Focuses on the
• Intervention fidelity – monitoring of description and interpretation of people’s
whether intervention is faithfully lived experience. Interviews. What is
delivered in accordance with plan and essence of phenomenon experienced by
treatment was received people?
• Manipulation – introduction of • Grounded theory – studies experience
intervention in experimental or quasi with a process and creates a theory of
experimental study to assess DV how that process works. Used to develop
• Matching – consciously forming and refine theoretically relevant
comparable groups. Must know in categories
advance what confounders are. • Data Saturation - the point in the
• Maturation – process that results from research process when no new
time passage information is discovered in data analysis.
• Selection redundancy
• Statistical conclusion validity • Observation – naturalistic settings.
• Statistical power – ranges from 0-1. Descriptive: describes event, activity, ppl
Capacity to detect true relationship. Use Reflective: personal reflections that relate
large sample. Sensitivity. How likely study to insight, broad themes emerge
is to distinguish an actual effect from one • Snowball sampling - nonprobability
of chance. Ex: “study with 80% power = sampling technique where existing study
80% significance.” subjects recruit future subjects from
among their acquaintances.
• Unstructured interview – conversational,
flexible, grand tour questions
• Theoretical sampling – purposive.
Preferred. Recruits participants going