Lecture 11: Quasi-experimental designs and applied research

applied research

has direct and immediate relevance to a solution of real-world problems

dual functions of applied research

useful and help people + introduce knowledge

ethical dilemmas of applied research

informed consent, privacy, coercion

tradeoff between internal and external validity in applied research

most applied research is in the field and not lab, ecological validity

problems unique to between subjects design in applied research

random assignment impossible, nonequivalent groups = confounds, matching

problems unique to within subjects design in appliedd research

cannot always counterbalance = order effects possible, attrition if longitudinal (some subjects dropout)

quasi-experimental design

occurs when casual conclusions about the effect of an IV cannot be drawn because subjects cannot be randomly assigned to the groups being given different levels of IV (no random assignment so we cannot talk about cause and effect, no manipulation)

what types of experiments can quasi be used on

single factor ex post facto designs with two or more levels, ex post factorial designs (more then one IV), person x environment (PxE) factorial designs, all correlational research

nonequivalent control group design

quasi-experimental design in which participants cannot be randomly assigned to control and experimental groups

interrupted time series designs

quasi-experimental design in which a program or treatment is evaluated by measuring performance several times before and after treatment is in place

program evaluation

a form of applied research that includes a number of research activities designed to evaluate programs from planning to completion

P.E - step one

need analysis: is there a need for the program? how large is the target population? is there already similar programs? data from several sources (census, key informants in community, surveys of available research)

P.E - step two

monitoring program (formative evaluation): early on: is the program being implemented as planned? do we need to tweak anything while there is still time?

P.E- step three

evaluation outcome (summative evaluation): final assessment about whether or not program was effecting (threatening: people may lose there jobs), can be tricky to make good decisions/interpet results- often cannot use true experiments; confounds, if you get a result that is not statistically significant it doesn't necessarily mean the program does not work

P.E - step 4

cost effectiveness analysis: even if program has a statistically significant benefit, the cost may be too high to continue

qualitative research

a strategy for systematic collection, organization, and interpretation of textual info... its about meaning, doesn't manipulate, its non-numerical, its interpretive and generates theory (inductive), it can work with quantitative data, it may generate hypothesis

defining features of qualitative research

concerned with richness of description, capturing individuals perspective, relatively unstructured research strategies

basic methods of qualitative research

interviews, focus groups, participant observation

interviews

a directed conversation; interviewer seeks description of subjects world, not just asking and responding to questions; it should be a knowledge-producing conversation, interactive process; what is being said and HOW it is being said, interview guides/protocols: flexible, open ended, uses probes

focus groups

collective discussion that guides understanding of a topic, group of people with certain characteristic, interaction and group dynamic are essential (widens range of responses, activates forgotten details, releases inhibition)

coding

review transcript repeatedly, develop codes and subsides, codes: words or short phrases that represent the essence or key attribute of narative/verbal info, describes data themes, interactive process

qualitative and quantitative measures

qualitative methods are not in opposition to quantitative methods, but are another way to answer research questions that cannot be answered using numerical data

first step in ALL research

what method fits with my research question?