Term
|
Definition
• To control for the possible effects of extraneous variables on the dependent variable
• Observed effects of the independent variable on the dependent variable are not caused by extraneous factors
• More difficult in field research; ideal in controlled, clinical settings |
|
|
Term
|
Definition
• The ability to generalize the study results to other groups and settings beyond those in the current experiment
• More difficult in controlled environments (less like the real world); higher in field research |
|
|
Term
| Threats to internal validity |
|
Definition
| • Extraneous factors that allow for alternative explanations as to what caused a given effect on the dependent variable |
|
|
Term
|
Definition
• Threat to internal validity
• When events occur between the pretest and the posttest of a research study that could affect participants in such a way as to impact the dependent variable
• Ex: stress, illness, natural disasters
• How to avoid: use a control group that was not exposed to the event in question for comparison to see if the event had an effect on the experimental group |
|
|
Term
|
Definition
• Threat to internal validity
• Changes occurring in participants because of passage of time
• Ex: physical, mental, experiences
• Most likely to happen with longer studies using children
• How to avoid: use a control group to note any differences between the two groups |
|
|
Term
|
Definition
• Threat to internal validity
• Taking a pretest somehow affects the taking of the posttest (and thus threatens the results of the experiment)
• May become “test-wise” and learn how to answer better
• How to avoid: use a control group that does not receive any pretesting |
|
|
Term
|
Definition
• Threat to internal validity
• Instruments are not accurate/precise enough, or do not measure what they are supposed to measure
• When procedures are not standardized
• How to avoid: Avoid poorly designed questions; make sure your instruments are well-designed and thought-through |
|
|
Term
|
Definition
• Threat to internal validity
• Participants selected in a non-random manner differ in some way
• Participants may have differences prior to the study that account for the differences in the results, unrelated to the effects of the independent variable
• How to avoid: recruit volunteers and then randomly assign them to groups; pretest participants on measures of the dependent variable to make sure there are no pretreatment differences between groups |
|
|
Term
|
Definition
• Threat to internal validity
• Combines selection bias with a maturation threat
• Occurs when using intact groups that vary in some element of maturity
• How to avoid: pretest and/or prescreen groups on maturity levels |
|
|
Term
|
Definition
• Threat to internal validity
• Occurs when participants are selected on the basis of their extremely high or low scores
• If retested, those with the higher scores would most likely score lower, and vice versa
• How to avoid: choose participants from random sampling; study participants with a full range of scores and not divide groups based on scores |
|
|
Term
|
Definition
• Threat to internal validity
• Participants drop out of the study or cannot be located
• Smaller numbers or biased groups can create statistical errors
• How to avoid: oversampling and using large group sizes; use incentives to improve retention rates; obtain good demographic information about the study participants before the study to document and analyze in findings |
|
|
Term
|
Definition
• Threat to internal validity
• When participant’s attitudes toward being involved in a study affect the way they behave
• Results may be from attention given to the experimental group over the control group, and not from the effect of the independent variable
• How to avoid: provide the control group with some type of special treatment that is comparable to the experimental group but would not have direct impact on the dependent variable; keep participants from knowing that they are taking part in a study or being observed (these are both very difficult) |
|
|
Term
|
Definition
• Threat to internal validity
• Altered behavior because of expectations
• How to avoid: when testing new drugs, give a control group a placebo and do not tell which group is receiving the placebo (blind study); the researcher does not know which group is receiving the medication (double-blind study); provide the same information to each group, or minimal information to reduce expectations |
|
|
Term
|
Definition
• Threat to internal validity
• Treatment of experimental group contaminates or spills over to comparison or control groups
• How to avoid: stress the importance of confidentiality until the end of the study; use groups that do not have contact with one another |
|
|
Term
|
Definition
• Threat to internal validity
• The location of program or data collection affects participant responses
• How to avoid: make the locations the same for all participants; try to minimize location differences that could impact the dependent variable |
|
|
Term
|
Definition
• Threat to internal validity
• Differences in persons presenting a program affect the program
• Persons implementing the experiment may introduce inequality or bias into the study (possible with two different presenters with different styles)
• Occurs if an individual implementing an intervention favors one group over another
• How to avoid: choose people to implement the programs who are equally trained and competent, and who follow a standardized protocol; have all presenters present to all groups; the presenter should be neutral, or a neutral observer |
|
|
Term
| Threats to External Validity |
|
Definition
• Anything that makes the study results difficult to generalize to a larger population
• Any threat to internal validity can influence the results and make them not generalizable to any other population |
|
|
Term
| Selection Treatment Interaction |
|
Definition
• Threat to external validity
• The ability of a researcher to generalize the results of a study beyond the groups involved in the study
• When using intact groups, the researcher is unable to know if they are truly representative of a larger population
• One can only generalize about the population of which a study population is randomly selected from, not to a larger population
• How to avoid: randomly select study participants from a larger population, although this is more difficult |
|
|
Term
| Setting Treatment Interaction |
|
Definition
• Threat to external validity
• The extent to which the environmental conditions or setting under which an experimental study was conducted can be duplicated in other settings
• How to avoid: try to make the environments of both study groups comparable; pretest participants about demographics and their setting |
|
|
Term
| History Treatment Interaction |
|
Definition
• Threat to external validity
• Occurs when researchers try to generalize findings to past and future situations
• How to avoid: retest at different times of the day/month/year |
|
|
Term
| Increasing Internal and External Validity |
|
Definition
• Randomly select participants from a well-defined study population
• Randomly assign selected participants to groups
• Include non-treatment control groups in the research design
• Use the statistical procedure called analysis of covariance to equalize many initial differences that might exist between groups
• If the results of an experiment are not internally valid, there is nothing to generalize, so first ensure internal validity before external validity |
|
|
