Week 6 - Control Comparison Groups

non-randomly assigned or nonequivalent control groups definition

A group might be added to the experimental design even though this group could not be randomly assigned to various conditions as were subjects in the other groups. Even so, this group may be selected and used to make implausible specific influences that are not otherwise well controlled (e.g., testing, maturation).

Non-specific Rx, Attention-Placebo, Common Factors Control group definition

An "intervention" is provided to include common factors or nonspecific features of the intervention (e.g., credible rationale, some procedures, participation or attendance) that are not considered to be the critical components that are responsible for intervention effects achieved in the intervention group. This is equivalent conceptually to a psychological "placebo," ergo the alternative name of "attention-placebo" control group.

instrumentation (threat to internal validity)

Changes in the scoring criteria, the instructions for completing the measures, or a response shift in the participants internal frame of reference for judging/measuring/rating

wait list control group (definition)

Identical to the no-treatment control group except after the second (post) assessment, participants now receive the intervention. Sometimes an additional assessment is provided after the wait list group receives the intervention.

no-treatment control group (definition)

No intervention is provided to this group and assessments (pre/post) are obtained over the same interval that the intervention is provided to the intervention group.

No-Contact Control Group

No intervention is provided, but participants are not aware that they are participating in a study and have no contact with staff involved in the project.

yoked control group definition

Not necessarily a separate group from one of the prior groups in this table. Yoking refers to matching subjects in different groups on some variable (e.g., duration or number of sessions) that might emerge during the course of the study. The investigator wishes to rule out the impact of these probably ancillary differences between intervention and nonintervention groups. Subjects in intervention and nonintervention groups are paired as "partners" so to speak and the emergent variable (e.g., more sessions) that was provided to the intervention subject is assigned to the partner.

no-treatment control group (pro and con)

Pro: the effects of history, maturation, and repeated testing as well as other threats to internal validity are directly controlled. con: the longer the waiting period, the more likely they will drop out

wait list control group pros and cons

Pro:This group allows careful evaluation of treatment effects at different points in the design. Because treatment eventually is provided to the wait-list control subjects, its effects can be evaluated empirically. Con: does not completely ameliorate the ethical problems of withholding treatment but may help a little.

Parametric Intervention Strategy

Question asked: What changes can be made in the specific treatment and its delivery that will enhance change? Requirements: Two or more intervention groups that differ in one or more facets (e.g., duration, intensity).

Dismantling Intervention Strategy

Question asked: What components are necessary, sufficient, or facilitative of change? requirements: Two or more intervention groups. One receives the full intervention package; other groups receive that package minus one or more components.

Constructive Intervention Strategy

Question asked: What components or other interventions can be added to enhance change? Requirements: Two or more intervention groups. One receives the full intervention package; other groups receive that package plus other components or the intervention.

Intervention Mediator Strategy

Question asked: What processes or constructs mediate the relation between the intervention and change? Requirement: One or more interventions with assessment on presumed processes that may be responsible, lead to, statistically account for change.

testing (threat to internal validity)

The effect of a test (e.g. pretest) on subsequent performance. can control for this by using a counterbalanced design

What is a control group?

The group that does not receive the experimental treatment in an experiment.

tau definition

The standard or routine treatment that is provided in a given setting for the same clinical problem or intervention focus. This group receives whatever is usually done, i.e., as usual care

What is internal validity?

This type of validity is focused on determining whether a study's findings are accurate, or are more the result of the influence of extraneous variables.

What is the purpose of a control group?

Used to address threats to internal validity (ex. history, maturation, selection, testing)

interaction of selection with other threat (threat to internal validity)

When the threats such as history, maturation, attrition etc. happen more to one group than another.

Maturation (threat to internal validity)

change that is internal to the participant, due to experience or practice, e.g. growing older, wiser, more tired, bored.

History (threat to internal validity)

events in the lives of all participants that might account for change. E.g. 9-11, school violence

attrition (threat to internal validity)

loss of participants when data is collected over more than one session.

tau pros and cons

pro: Demands for service and ethical issues associated with many other control conditions are met. con: It is difficult to know what these treatments entail at a clinic, hospital, or school, no matter what the descriptions and brochures actually say.

Non-specific Rx, Attention-Placebo, Common Factors Control group pros and cons

pro: address threats to internal validity (e.g., history, maturation, repeated testing) and can also address threats to construct validity con: one usually does not know in advance that the properties of the nonspecific-treatment group are inert. AND client who receives a "fake treatment" might be turned away from therapy in the future when a veridical treatment might help with the stresses and strains of life.

yoked control group pros and cons

pro: the investigator controls those variables that potentially can confound the results. Con: The variables that serve as the basis of yoking often are based on considerations of construct validity.

no contact control group pros and cons

pro: understanding how changes come about without interventions would be informative and potentially useful to harness. con: its rarely a viable option for research b/c you need informed consent

non-randomly assigned or nonequivalent control groups pro and con

pro:to reduce the plausibility that other influences (internal validity or construct validity) could explain the results. con: Because the group is not comprised randomly, the data may not be as persuasive as parallel data resulting from a randomly comprised control or comparison group.

Comparative Intervention Strategy

question asked: How effective is this intervention relative to other interventions for this clinical problem or intervention focus? requirements: Two or more groups that receive different interventions.

Intervention Package Strategy

question asked: Does the intervention lead to change (e.g., improvements after treatment)? basic requirements: Intervention vs. no-intervention or waiting-list control group

Intervention Moderator Strategy

question asked:What patient, family, contextual, or other characteristics influence the direction or magnitude of change with this intervention? requirement: One or more interventions but divided by levels of a predicted moderator (e.g., severity symptoms).

selection bias (threat to internal validity)

systematic differences between groups

statistical regression (threat to internal validity)

the tendency for extreme scores to regress towards the mean