Quiz 3 Flashcards
What are the factors that quantitative and qualitative researchers should consider when making decisions about study designs?
- Intervention: introducing an intervention is the difference between an experimental study and a nonexperimental study.
- Comparisons: will the pt be compared with themself at a different time or with others in a different group. both experimental and nonexperimental studies can compare different people or the same people at different times
- Controlling confounding variables: in quantitative studies efforts are made to control extraneous variables
- Blinding: minimize who has access to info about participants in order to prevent expectation bias
- How often will data be collected: data collected at a single point in time (cross-sectional) vs. data collection at multiple points (longitudinal)
- When will effects be measured, relative to potential causes: Retrospective study (look back at data and determine the cause) Prospective study (begin with a prospective cause and determine results)
- Where will the study take place? (Location): data for quantitative studies can be gathered in a variety of places. Consider the number of test sites when determining generalizability
How do quantitative researchers evaluate the quality of their sample?
-
a criterion for judging a sample is its representativeness (does it represent the population?)
- both sample size and the selection method are used to judge the quality of a sample.
- The number of participants is a major concern in quantitative studies - a larger group is typically better than a small one (less sampling error)(statistical conclusion validity; conclusion is considered to be valid because it is less prone to error). Large samples are not immune to inaccuracy.
- When differences between groups of participants are expected to be large then a small group will suffice. If the differences between groups are nuanced then a larger sample will be needed.
- Power analysis can be performed to estimate how large the sample should be
- Study should have:
- Defined population with clear eligibility criteria
- Mention of the type of sampling plan
- A Sampling plan that produces the most representative sample possible
- Is the sample size large enough to support statistical conclusion validity?
- Are key characteristics such as average age and percentage of male/female included?
- Generalizable results
Differentiate between retrospective and prospective research designs.
-
Retrospective: (case-control designs) involve collecting data about an outcome in the present and then looking back in time for possible causes.
- For example, in retrospective lung cancer research, researchers begin with some people who have lung cancer and others who do not and then look for differences in antecedent behaviors or conditions, such as smoking habits. Such a study uses a case-control design—that is, cases with a certain condition such as lung cancer are compared to controls without it. In designing a case-control study, researchers try to identify controls who are as similar as possible to cases with regard to confounding variables (e.g., age, gender). The difficulty, however, is that the two groups are almost never comparable with respect to all factors influencing the outcome.
-
Prospective: (cohort) designs, researchers begin with a possible cause and then subsequently collect data about outcomes.
- follow smokers around for years to see if they develop cancer
- follow a cohort of people around to observe
Differentiate between cross-sectional and longitudinal research designs.
- Cross-sectional:
data are collected at one point in time. For example, a researcher might study whether psychological symptoms in menopausal women are correlated contemporaneously with physiologic symptoms. Retrospective studies are usually cross-sectional: Data on the independent and outcome variables are collected concurrently (e.g., participants’ lung cancer status and smoking habits), but the independent variable usually concerns events or behaviors occurring in the past. Cross-sectional designs can be used to study time-related phenomena, but they are less persuasive than longitudinal designs. Cross-sectional designs are economical, but they pose problems for inferring changes over time. The amount of social and technological change that characterizes our society makes it questionable to assume that differences in the behaviors or characteristics of different age groups are the result of the passage through time rather than cohort differences
- Longitudinal: involve collecting data multiple times over an extended period. Such designs are useful for studying changes over time and for establishing the sequencing of phenomena, which is a criterion for inferring causality. In nursing research, longitudinal studies are often follow-up studies of a clinical population, undertaken to assess the subsequent status of people with a specified condition or who received an intervention
Differentiate between experimental and quasi-experimental designs
- Experimental Designs
- sometimes called a posttest-only design
- most basic experimental design involves randomizing people to different groups and then measuring outcomes
- true experiments only if people are randomly assigned to different orderings of treatment
- RCTs are the “gold standard” for intervention studies (Therapy questions) because they yield the most persuasive evidence about the effects of an intervention. Through randomization to groups, researchers come as close as possible to attaining an “ideal” counterfactual.
- First, many interesting variables simply are not amenable to intervention. A large number of human traits, such as disease or health habits, cannot be randomly conferred; cannot randomly assign to genetic dz
- many variables could technically—but not ethically—be experimentally varied.
- Sometimes, RCTs are not feasible because of practical issues. It may, for instance, be impossible to secure the administrative approval to randomize people to groups
- Quasi-experimental
- Trials without randomization (medical literature)
- Involve intervention
- lacks randomization (hallmark of true experiment)
- may lack a control group
- signature of quasi experimental design is implementation and testing of an intervention in the absence of randomization
- nonequivalent control group: pretest-posttest design; compare 2 or more groups of people before and after implementing an intervention
- design is weaker bc without randomization it cannot be proven that groups are equal from the beginning
- uses term comparison group instead of control group
- strong quasi studies will introduce some controls; ex. baseline measurements
- quasi experiments tend to be practical
- might not be ethical to randomize; can still add controls
- less conclusive results and less generalizable
- causal inferences cannot be readily made
- usually at least one possible rival explanation
Blinding
In controlled trials the term blinding, and in particular “double-blind,” usually refers to keeping study participants, those involved with their management, and those collecting and analyzing clinical data unaware of the assigned treatment, so that they should not be influenced by that knowledge.
Randomization
A study design that randomly assigns participants into an experimental group or a control group. As the study is conducted, the only expected difference between the control and experimental groups in a randomized controlled trial (RCT) is the outcome variable being studied.
Matching
Matching is a statistical technique which is used to evaluate the effect of a treatment by comparing the treated and the non-treated units in an observational study or quasi-experiment.
same characteristics between groups (match characteristics) allows for equity
Sampling
Sampling is the process of selecting units (e.g., people, organizations) from a population of interest so that by studying the sample we may fairly generalize our results back to the population from which they were chosen
Sampling Bias
Sampling bias occurs when some members of a population are systematically more likely to be selected in a sample than others. It is also called ascertainment bias in medical fields. … In other words, findings from biased samples can only be generalized to populations that share characteristics with the sample.
Describe sampling and sample considerations in qualitative research?
- Rely on small nonprobability samples
- Quantitative researchers measure attributes and identify relationships in a population; they want a representative sample so findings can be generalized
- Researchers will ask questions like who will give the most valuable info, who should I talk to, who should I observe - in order to maximize my understanding of a phenomenon
- first step in qualitative sampling is selecting settings with information richness
- with progression of the study researchers will look for participants who confirm, challenge, or enrich understandings
- researchers will have eligibility criteria like quantitative studies but they do not have explicit target population
- Avoids random sampling bc they are not the best method of selecting people who are knowledgable, articulate, reflective, and willing to talk with researchers
- they use various nonprobability sampling designs
Types of qualitative sampling
- Convenience - volunteer sampling; efficient but not preferred because it may not provide the most info-rich sources (goal is to get the most info with a smaller group.)
- Snowball - current participants refer potential participants to the researchers; sample may be restricted to a small network of acquaintances, quality of referrals may also be affected by the new participants trust and desire to be involved in research.
- Purposive - studies may begin with snowballing or convenience but mat eventually become purposeful; researchers pick out participants
-
Maximum variation - deliberately selecting cases with a wide range of variation on dimensions of interest
- Extreme (Deviant) - allows learning from participants who are unusual or extreme (outstanding success or failure).
- Typical case - selection of participants who are the typical or average example/case.
- Criterion - studying cases who meet a predetermined criterion of importance
What are the criteria for selecting the sample for qualitative studies?
-
Sample size is based on the information needs in a qualitative study
- The number of participants depends on data saturation (redundancy of info) How many individuals need to be included until this achieved.
- Individuals who are insightful, reflective, and articulate are desired in order to obtain quality data; data saturation can be achieved quickly with participants like this.
- A larger sample may be needed with maximum variation sampling and typical case sampling
- The main criterion of selecting sample size is data saturation (information redundancy).
- Appropriateness (of methods) - An appropriate sample results from the selection of participants who can best supply information that meets the study’s conceptual requirements.
- The sampling strategy must yield a full understanding of the phenomenon of interest
- Participants with diverse experiences are included - A sampling approach that excludes negative cases or that fails to include people with unusual experiences may not fully address the study’s information needs.
- Potential transferability of findings - similarity between the study sample and other people to whom the findings might be applied
What are the purposes of mixed methods? What are the strengths of this design?
- Purpose: Research that integrates qualitative and quantitative data and strategies in a single study or coordinated set of studies, some questions require mixed methods
-
Strengths:
- complementarity of quantitative and qualitative data and the practicality of using methods that best address a question
- practicality: given the complexity of phenomena, it is practical to use whatever methodological tools are best suited to addressing pressing research questions
- enhanced validity: hypothesis or model is supported by multiple and complementary types of data, researchers can be more confident about their inferences
Describe phases of clinical trials
- Studies that develop clinical interventions and test their efficacy and effectiveness
- Phase I: designed to establish safety, tolerance, and dose with a design; focus is on developing the best treatment
- Phase II: pilot test of treatment effectiveness. Researchers see if the intervention is feasible and acceptable or holds promise; this phase is designed as small-scale experiment or quasi-experiment
- Phase III: full experimental test of the intervention, an RCT. The objective is to develop evidence about the treatment’s efficacy – when term “clinical trial” is used it is referring to this stage
- Phase IV: involves studies of the effectiveness of an intervention in the general population. Emphasis is on external validity
Intervention =
experimental study/quasi
Quasi-experimental studies don’t have
randomization; it has a control/comparison group and an intervention
nonexperimental design
- observational or descriptional
- it is considered a quantitative study
- Any study that does not have an intervention (observational)
within-subjects design
same people compared at different times or under different conditions; pretest and baseline data; then collect posttest data to compare
Between-subjects design
Different people are compared (ex. men and women); give intervention to each group and collect data, then compare control to experimental group.
Analysis of covariance
A statistical test that can be performed to control for confounding (extraneous) variables
Masking
making intervention unknown to researchers or participants
Causality
most quantitative research focuses on looking to answers cause and effect questions.
Experimental designs are the only type of design which can examine cause and effect relationships
Criteria for Causality:
- cause must happen before effect
- must have a demonstratable empirical relationship btwn cause and effect
- cause cannot be explained by a third variable
Posttest only (after only) design
- collect data after the end (after intervention)
- hard to tell if intervention contributed to outcome without baseline
Pretest - posttest (/before-after) design
- collect baseline data and data after intervention
- usually more telling
