- In order to establish strong cause-and-effect relationships when conducing experimental research, internal validity should be strong.

- Internal validity exists if the observed effects of the independent variable on the dependent variable are real and not caused by extraneous variables.

- Internal validity is threatened when there are alternative explanations for the effect of the independent variable on the dependent variable.

- "Threatens internal validity when events occur between the pretest and posttest of a research study that could affect participants in such a way as to impact the dependent variable"

- Examples: stress, natural disasters, news events, illness

- Addressing the concern: Use a control group that is not exposed to the independent variable/treatment and an experimental group that is exposed to the independent variable/treatment. Compare the two groups.

- Threatens internal validity "when changes are seen in subjects because of the time that has elapsed since the study began and that may not be the result of any program."

Examples: Becoming older, wiser, stronger, or more experienced

Addressing the concern: Compare a control and experimental group that would maturate at the same rate.

- Threatens internal validity when any differences seen from baseline, measurment taken prior to a study, to posttest may be from a testing effect, learning how to do better on the test, and not the independent variable

- Addressing the concern: Use a control group that does not receive any pretesting.

- Threatens internal validity "when measurements are not accurate or procedures are not standardized."

- Addressing the concern: Physical instruments must be carefully calibrated and individuals using the instruments must be well trained. Written instruments need to accurately measure what they are supposed to measure and items must ask the correct questions.

- Threatens internal validity when individuals are selected in a nonrandom manner. Differences found between groups may be due to differences existing before the study instead of as a result of the study.

- Addressing the concern:Recruit volunteers, and randomly asssign them to gorups instead of having volunteers self-select their groups. Match participants on selected characteristics, and randomly assign them to groups. Pretest groups on measures of the dependent variable.

- Threatens internal validity by combining a selection bias with a maturation threat when using intact groups that vary in maturation level.

- Addressing the concern: Pretesting and/or prescreening groups on maturity levels

- Threatens internal validity when "participants are selected on the basis of their extremely high or low scores."

- Regression to the mean: Those with high scores may have nowhere to go but down, and those with low scores may have nowhere to go but up

- Addressing the concern: Follow a random sample of participants representing the full range of scores rather than placing students in high and low groups based on one test.

- Threatens internal validity when there is a loss of subjects due to death, illness, moving, or dropping out of the study.

- This may be difficult for group comparisons when the number of participants falls beneath established guidelines and may make it difficult to generalize to larger populations. Dropouts may introduce bias because those left in the study are the ones who want to remain.

- Addressing the concern: Oversample and use large group sizes. Use incentives to encourage participants to stay in the study. Document demographics of the participants before and after the study.

- Threatens internal validity when "participants' attitudes towards being involved in a study affect the way they behave."

- Differnces in the dependent variable between the experimental group that receives special attention and the control group that does not, may not be the effect of the independent variable but rather a result of the attention given.

- Addressing the concerns: The control group could have some special treatment comparable to the experimental group that would not directly impact the dependent variable. Do not let participants know they are taking part in a study.

- Threatens internal validity because the participants' expectations effect the results.

- Addressing the concerns: Give the control group a placebo, something that looks like the actual treatment. Conduct a bind study where the participants do not know whether or not they are receiving the placebo or a double bind study where both the participants and the researcher so not know. Before a study begins, give both groups similar amounts of information.

- Threatens internal validity when " the treatment being applied to one group spills over or contaminates another group." Members of the experimental group could tell members of the control group about the treatment. Members of the control group could implement the treatment on their own.

- Addressing the concerns: Make sure the control group and experimental groups are not closely located to each other. Stress the importance of keeping the treatment confidential.

- Threatens internal validity "when there are differences in the locations where interventions take place."

- Addressing the concerns: Make the locations the same for all participants or minimize any location differences that would effect the dependent variable.

- Threatens internal validity when individuals implementing the experimental treatment may inadvertently introduce inequality or bias into the study. People providing the treatment may not be equal in knowledge, understanding, personality traits, presentation skills.

- Addressing the concerns:  Make sure all persons implementing the treatment are equally trained and competent, make sure to follow a standardized protocol for implementation, have individuals present to all groups.

- An individual implementing the program may inadvertently favor one group over another.

- Addressing the concern: Do not have the program developer present the program or have a neutral observer observe the presenter and note any differences.

- In order to establish strong cause-and-effect relationships when conducing experimental research, external validity should be strong.

- External validity "is concerned with the ability to generalize the study results to other groups and settings beyond those in the current experiment."

- External validity is threatened when there are alternative explanations for the effect of the independent variable on the dependent variable and, therefore, cannot be generalized to any other population.

- Threatens external validity when the results of the study may not be generalized to a different population. Overgeneralization can occur when the results of an intace group are generalized to a larger population.

- Addressing the concerns: Choose a random sample making sure that generalizations apply to the population of the study and are not applied beyond to a larger population.

- Threatens external validity when considering the "extent to which the environmental conditions or settings under which an experimental study was conducted can be duplicated in other settings."

- Addressing the concerns: Note the differences in the participants and the settings which may influence the results of the study.

- Threatens external validity when "the researcher tries to generalize findings to past and future situations."

- Addressing the concerns: Note if the experiment is time-sensitive.

Internal Validity:

     - Randomly select participants from a well-defined

       study population.

     - Randomly assign selected participants to groups.

     - Include non-treatment control groups in the

       research design.

     - Analysis of covariance: equalizes many differences

       that might exist between groups.

External Validity:

     - Consider where generalizations are appropriate.

     - Duplicate setting and historical factors when

       when replicating a program with a different

       population.

EXTERNAL VALIDITY CANNOT EXIST WITHOUT INTERNAL VALIDITY.