quantitative research

descriptive

- descriptive exploratory, descriptive comparative •Provide a picture of a situation as it is naturally happening without manipulation of any of the variables •Used to generate hypotheses •Examine different types of phenomena

pre experimental

-One-group post-test only X O1 - Nonequivalent groups posttest only design •Provides weak evidence •Symbolic representation: X O1 X O2

6 types of experimental design

1.Two groups, pretest-posttest (in both groups) 2. Two groups, posttest only (in both groups) 3. Solomon four-group 4. Multiple experimental groups 5. Factorial 6. Crossover designs

research design

It is the specific plan that is used to meet the stated purpose of the research study

random assignment

Subjects have an equal chance of being in either the treatment or the control group (randomly chosen from available group)

criteria for casuality

Three key criteria for making causal inferences: •The cause must precede the effect in time. •There must be a demonstrated empirical relationship between the cause and the effect. •The relationship between the presumed cause and effect cannot be explained by a third variable; another factor related to both the presumed cause and effect cannot be the "real" cause.

experimental

Two groups pretest-posttest design (or "before-after"), a classic design •Data collected both at baseline and after the intervention •Symbolic representation: R O X O ROO Two groups posttest-only design (or "after-only") •Outcome data collected only after the intervention •Symbolic representation: R X O R O Rationale: pre-test would change subjects' perceptions or pre-test would be nonsensical

correlational

descriptive correlational, predictive correlational •Change in one variable is associated with change in other variable •Note: Correlation does not prove causation •Can be model-testing study

between group design

different people are compared (e.g., men and women, college juniors and seniors; patients in two different hospitals; two groups of participants are compared)

prospective

stronger than retrospective designs in supporting causal inferences Experimental designs are the strongest but sometimes we cannot conduct them (because of ethical reasons)

manipulation

the ability of the researcher to manipulate/control the independent variable

control

the ability to control, regulate, manipulate or statistically adjust for factors that can affect the dependent variable/outcome

bias

when extraneous variables influence the relationship between the IV and DV

randomization

•: The selection, assignment, or arrangement of participants by chance •An effective way to control for extraneous variables

non experimental research

•A research design that lacks the manipulation of an IV and random assignment •No treatment/intervention •Are used to describe a phenomena in detail, to explain relationships or differences among variables, or to predict relationships among variables •Can be descriptive or correlational in nature •Situations in which the IV cannot ever be manipulated •Gender, blood type, personality, health beliefs, medical diagnosis •Situations where it would be unethical to manipulate the IV •Child abuse, cigarette smoking •Situations in which it is simply impractical to conduct a true experiment •Constraints include money, time, administrative approval, inconvenience to subjects or staff •Situations where an experimental design is not appropriate •Descriptive studies, qualitative studies

solomen four group

•Alternative to the posttest-only control group design that is used to control for the testing threat by measuring it; superior to two-group posttest only as selection bias is minimized • •Includes four groups (two intervention groups, two control groups), all randomly assigned: •Pretest - treatment - posttest •Pretest - no treatment - posttest •Treatment - posttest only Posttest only

quantitative studies

•Causality and Probability •Control and Manipulation •Extraneous variables and Bias •Randomization (sampling and assignment) •Designs (how many groups ; when data is collected) •

factorial design

•Classic experimental design plus at least one additional independent variable that is also randomly assigned •Tests for multiple causality using two or more independent variables •More than one independent variable is experimentally manipulated •Example: a 2 × 2 design = 2 levels of one IV, 2 levels of a second IV

crossover design

•Classic experimental design with at least one additional period of data collection, in which experimental and control conditions are reversed •All subjects receive the intervention once and the control condition once •Subjects are exposed to 2+ conditions in random order. •Subjects serve as their own control. •Symbolic representation: R O XA O XB O R O XB O XA O Advantage: smaller sample size is needed

classes (quantitative data)

•Experimental: •3 main conditions must be present •Intervention or treatment •Control •Randomization •Quasi-experimental •Intervention is present but control or randomization is lacking (or both) •Non-experimental Collect data withoutmaking changes or introducing treatments (such studies as descriptive, correlational, or observational studies)

quasi experimental

•Involves manipulation of the IV • •Missing a component of the true experiment: either randomization or a control group, or both. • •Always has IV and DV or it would not be an "experiment" • •Weaker in ability to make causal inferences (cause-and-effect relationships) •Nonequivalent control group pretest-posttest •Time series •Pre-experimental designs

time series design

•No control group nor randomization •Collection of information over an extended period of time AND the introduction of an experimental treatment during the course of the data collection period • •Multiple data collection points strengthens the design - but not as strong as an experiment • •OOOOOO X OOOOOO

relationships and casuality

•Not all relationships are cause and effect relationships •There are functional and causal relationships Examples: - A relationship between height and shoe size is functional - A relationship between maternal sleep one week before labor and preterm labor is causal

disadvantages of experiments

•Not always possible to meet criterion for randomization •Some variables not amenable to manipulation (e.g., age, gender, health history, disease) •Some variables may technically be manipulated but ethical considerations prohibit manipulation (e.g., cigarette smoking, child abuse) •Experiment as reductionistic and artificial (focus on a few variables) •Hawthorne effect (we will talk about it later)

nonequivalent control group

•Participants in the groups might differ •Most frequently used quasi-experimental design •Involves an experimental treatment & 2 or more groups of subjects •Control group used (which is usually called a comparison group) •Lack of randomization, much weaker design as it can no longer be assumed that the experimental and comparison groups are equal at the start of the study •Symbolic representation: O1 X O2 O1 O2

advantages of quasi experiments

•Practical, feasible, generalizable •Fits in with real world of nursing - when it is difficult to deliver innovative Rx to half a group; not possible to randomize or to secure a control group •Introduces some control when true experiment is not possible •Nursing frequently interested in research problems that take place in the natural setting

threats to external validity

•Reactivity: the influence of participating in the study on the subjects •Hawthorne effect: subjects' behaviors are affected by desires to please the experimenter and personal values/benefits) • •Construct validity (related to definitions of variables, concepts and measurement tools; aka measurement effects) • •Effects of selection (do subjects represent the target population?)

validity threats

•Selection bias: characteristics of subjects affect the DV (e.g., unplanned effect of them) •For example: People who choose to be in a study may produce better outcomes because they get paid, or more educated, etc) •History: DV is influenced by an event that occurred during the study •Maturation: subjects change by growing 'older' and more experienced in life, participating in research, etc. •Testing: e.g., a pretest can influence the results of a post-test because people can remember the questions •Instrumentation: There are inconsistencies in data collection (e.g., tools are not appropriate for the population studied - say, pediatrics tools used on toddlers) •Mortality: There is loss of subjects before the study is completed (it does not mean they died!) •Statistical conclusion: The degree that the results of the statistical analysis reflect the true relationship between the DV and IV •Type II Error: researchers inaccurately conclude that there is no relationship between IV and DV

Threats to statistical conclusion validity

•Statistical conclusion validity: correctness of the decisions the researcher makes regarding statistical tests •Violated assumptions of statistical tests •Low statistical power: study sample is not large enough to result in statistically significant findings •Fishing and the error rate problem: researchers performing multiple statistical tests, fishing for statistically significant results

random sampling

•Technique for selecting elements when each person in the population of interest has the same chance of being selected (sampling frame, random number generator is used)

internal validity

•The degree to which one can conclude that the independent variable (IV) produced changes in the dependent variable (DV) •The extent to which it is possible to make an inference that the IV is truly influencing the DV and nothing else •The IV did what it said it did • •Threats to internal validity: selection bias, history, testing, instrumentation, mortality, statistical conclusion

external validity

•The degree to which the results of a study can be generalized to other subjects, settings, and times •Research on cancer in adult and pediatric population •When considering external validity, you must reflect on issues related to the generalizability of the study, specifically: the people involved (the sample), where the study took place (the setting - say, urban or rural), and the time in which the study took place (was data collected 5 years ago? 10 years ago?).

validity

•We have to evaluate the soundness of the evidence - whether findings are convincing, well grounded •We have to evaluate whether the results are logical, reasonable, and justifiable based on the evidence presented

within group design

•Within-group design: the same people are compared at different times or under different conditions; one group of participants only (e.g., pain level before and after intervention)

probability

•a likelihood or chance that an event will occur in a situation • • Because it is very difficult to prove anything with 100% certainty, we ascertain with probability •Denoted as small letter 'p' in publications

cross sectional

•collecting data from a group of subjects at one point in time; non-experimental design Example: Anxiety in parents of preterm/term infants measured at 6 weeks post discharge - measured only once

longitudinal

•gathering data at more than one point in time; either experimental or non-experimental Example: Anxiety in parents of preterm infants measured at 6, 12, 18, and 24 weeks post discharge - measured several times

retrospective

•researchers look back in time to determine possible causative factors; never experimental. Presumed cause and anticipated effect have already occurred, and we're investigating what had happened •Example: Depressed women asked if they experienced miscarriage to see if miscarriage was the reason for depression

prospective

•studies over time that follow subjects to determine if the hypothesized effects actually occur Example: Depression is measured in women who had experienced miscarriage, after intervention program (to see the effects of the program); or asking about dietary patterns and asking occurrence of the heart disease later on.

casuality

•the relationship between cause and effect; there can be multiple causes or factors that contribute to an outcome •Many (if not most) quantitative research questions are about causes and effects. •Research questions that seek to illuminate causal relationships need to be addressed with appropriate designs (usually experimental) •Three conditions (criteria) should be present