An interrupted time-series design is a quasi-experimental approach and a variant of the pretest-posttest design. In this methodology, a researcher takes a time series—multiple measurements of a dependent variable at regular intervals—both before and after a treatment is introduced. The treatment effectively 'interrupts' the ongoing measurements. Unlike the simpler pretest-posttest design, which relies on a single before-and-after measurement, the inclusion of multiple measurements allows researchers to better distinguish the actual effect of the intervention from normal variation.

Interrupted Time-Series Design

To effectively rule out threats to internal validity—such as history, maturation, testing, instrumentation, regression to the mean, and spontaneous remission—researchers can add a control group to a pretest-posttest design. A control group does not receive the experimental treatment but is subjected to the same external events, passage of time, and repeated measurements as the treatment group. By comparing the posttest changes between the two groups, researchers can isolate the actual effect of the treatment from these alternative explanations, although doing so means the study is no longer a one-group design.

Control Group in Pretest-Posttest Designs

In research design, history serves as an alternative explanation for changes in a dependent variable, representing a significant threat to internal validity. This threat occurs when extraneous external events happen between the pretest and posttest measurements, unintentionally causing a shift in the participants' scores. For instance, if an anti-drug education program is being evaluated, an unexpected and highly publicized celebrity drug overdose occurring during the study period could independently influence participants' attitudes, confounding the results and making it impossible to determine if the education program was the true cause of the change.

History as a Threat to Internal Validity

Maturation is a threat to internal validity that occurs when participants undergo natural developmental or physiological changes—such as growing older, learning, or becoming fatigued—between the pretest and posttest measurements of a study. These natural, ongoing changes, rather than the experimental treatment itself, may be the actual cause of any observed differences in the dependent variable over time. For example, over the course of a year-long intervention, children naturally become better reasoners, which could be misconstrued as an effect of the intervention.

Maturation as a Threat to Internal Validity

The testing threat to internal validity occurs when the mere act of taking a pretest inadvertently influences a participant's responses on the subsequent posttest. Completing the initial measurement might inspire further thought, conversation, or practice regarding the topic being studied. This increased awareness or familiarity can independently cause a change in the posttest scores, regardless of whether the actual experimental treatment was effective.

Testing as a Threat to Internal Validity

Instrumentation threatens a study's internal validity when the fundamental characteristics of the measuring instrument or process change between the pretest and posttest. This issue frequently arises in observational studies where human observers are used to measure behavior, as they may gradually gain skill, become fatigued, or unconsciously alter their evaluation standards over time. Additionally, instrumentation effects can manifest when the participants themselves change their approach to the measurement tool, such as taking a survey very seriously during the novel pretest but becoming bored and less careful with their responses during the posttest.

Instrumentation as a Threat to Internal Validity

Regression to the mean acts as a severe threat to internal validity when participants are selected for a research study specifically because of their extreme initial scores. Statistically, individuals who achieve exceptionally high or low scores on a pretest are highly likely to score closer to the population average on a subsequent posttest simply due to natural, random variability. This makes it incredibly difficult to determine whether a shift in scores was genuinely caused by the experimental treatment or was merely the result of statistical regression.

Regression to the Mean as a Threat to Internal Validity

Spontaneous remission is the natural tendency for many medical conditions and psychological problems to gradually improve over time without any formal treatment or intervention. This phenomenon acts as a significant confounding variable in studies that lack a control group; researchers evaluating a treatment using a simple pretest-posttest design may falsely attribute a patient's recovery to their specific intervention, when in reality, the condition would have naturally improved on its own.

Spontaneous Remission

A practical example of a one-group pretest-posttest design involves evaluating an anti-drug education program for elementary school students. Researchers first measure the students' baseline attitudes toward illegal drugs (the pretest). Next, they implement the anti-drug curriculum over a specified period. Finally, they measure the students' attitudes again (the posttest). By comparing the pretest and posttest scores, researchers can see if attitudes changed, though the lack of a control group makes it difficult to definitively attribute any change solely to the educational program.

Example of a One-Group Pretest-Posttest Design

Why is it difficult to conclude with certainty that a treatment was effective when using a one-group pretest-posttest design?

A one-group pretest-posttest design is a quasi-experimental research method where a single group of participants is measured on a dependent variable both before and after a treatment is administered. Unlike a within-subjects experiment, this design does not counterbalance the order of conditions because the treatment must inherently occur between the initial baseline measurement (pretest) and the subsequent measurement (posttest). While an improvement in posttest scores compared to pretest scores might intuitively suggest that the treatment was effective, such conclusions are difficult to make with certainty because the absence of a control group leaves the study highly vulnerable to alternative explanations and threats to internal validity.

Google

A one-group posttest only design is a quasi-experimental approach where a researcher applies a treatment or manipulates an independent variable, and then measures the dependent variable just once after the intervention. It is considered the weakest quasi-experimental design because it lacks a control or comparison group. Without this comparison, there is no way to determine what the participants' outcomes would have been had they not completed the treatment.

One-Group Posttest Only Design

Quasi-experimental research is a methodological design that incorporates some, but not all, of the essential features of a true experiment. While researchers using this approach may introduce a treatment or comparison condition, they frequently fail to use random assignment to place participants into distinct groups or omit counterbalancing to control for potential order effects. Consequently, while it offers more control than purely correlational studies, quasi-experimental research generally possesses lower internal validity than true experimental research.

Quasi-Experimental Research

https://kpu.pressbooks.pub/psychmethods4e/

KPU Research Methods in Psychology - 4th American Edition

A practical example of a one-group posttest only design is measuring elementary school students' attitudes toward illegal drugs immediately after they complete an anti-drug education program, without any prior measurement or control group. Another common example involves media reports, such as an advertiser claiming that $$80\%$$ of women noticed brighter skin after using a specific cleanser for a month. In both cases, the lack of a comparison group makes it impossible to draw meaningful conclusions about the actual effectiveness of the treatment.

Examples of One-Group Posttest Only Designs

One-Group Pretest-Posttest Design

Why is the one-group posttest only design considered the weakest quasi-experimental approach?

An example of a quasi-experimental study is investigating the effect of a person's sex on their spatial memory. In this case, sex is the independent variable, and spatial memory is the dependent variable. Researchers can observe and compare differences between males and females, but they cannot manipulate or randomly assign a participant's sex, which prevents them from making a causal claim.

Example of a Quasi-Experimental Study: Sex and Spatial Memory

A nonequivalent groups design is a specific type of between-subjects design and a quasi-experimental approach in which researchers do not use random assignment to allocate participants to their respective conditions. Due to this absence of random assignment, the resulting comparison groups are nonequivalent, meaning they are likely to have inherent, pre-existing differences. These initial dissimilarities introduce potential confounding variables that can significantly threaten the internal validity of the research.

Nonequivalent Groups Design

An example of a quasi-experimental study lacking random assignment involves evaluating an anti-bullying program across two similar schools. A researcher implements the program in one school (the treatment group) and compares the bullying incidents to the other school (the control group). Although this design uses a comparison group, the inability to randomly assign individual students to the schools means there may be a selection effect; preexisting differences between the students in the two schools could alternatively explain any observed differences in bullying.

Example of a Quasi-Experimental Study: Anti-Bullying Program

A selection effect is a type of confounding variable that occurs when the participants in a treatment group differ from those in a control group in some systematic, pre-existing way. This issue frequently arises in quasi-experimental designs where researchers cannot use random assignment to place participants into conditions. When a selection effect is present, it provides an alternative explanation for the results, making it difficult to confidently determine whether the observed differences in the dependent variable were caused by the experimental treatment or by the inherent differences between the groups.

Selection Effect

Experimental, quasi-experimental, and non-experimental (correlational) research vary systematically along a spectrum of internal validity. True experimental research generally possesses the highest internal validity because it utilizes the manipulation of an independent variable and strict control over extraneous variables to confidently rule out alternative explanations. In contrast, non-experimental research ranks lowest in internal validity because it fundamentally lacks variable manipulation and control. Quasi-experimental research occupies the middle ground; it incorporates some structural features of an experiment but lacks others, such as the random assignment of participants.

Comparison of Internal Validity Across Research Designs

The Tuskegee syphilis study, conducted by the US Public Health Service from 1932 to 1972, is a tragic historical example of extreme injustice in scientific research. The study involved poor African American men in Alabama who were misled into believing they were receiving treatment for 'bad blood' while actually being observed for the untreated progression of syphilis. The researchers deliberately denied the men treatment even after penicillin became the standard cure, clearly violating the ethical mandate to treat participants fairly and highlighting the urgent need to protect marginalized groups from disproportionate research risks. The profound betrayal of these men was formally acknowledged in 1997 when President Bill Clinton issued a public apology on behalf of the US government, emphasizing how the participants had been exploited without their knowledge or consent.

Tuskegee Syphilis Study

In quasi-experimental research, researchers manipulate an independent variable prior to measuring the dependent variable. This chronological sequence effectively resolves the directionality problem—a common issue in purely non-experimental studies—because the established timeline clarifies which variable is the potential cause and which is the effect.

Elimination of Directionality Problem in Quasi-Experiments

Quasi-experiments are frequently employed to assess the effectiveness of specific treatments, such as educational interventions or psychotherapeutic approaches. They are particularly valuable and common in real-world field settings where practical, logistical, or ethical constraints make it challenging or impossible to randomly assign participants to different conditions.

Applications of Quasi-Experimental Research

While quasi-experimental research manipulates an independent variable before measuring a dependent variable, it still cannot definitively establish cause-and-effect relationships. The absence of random assignment or counterbalancing means researchers cannot confidently rule out confounding variables—pre-existing differences between conditions—as the true cause of any observed outcomes.

Learn Before

Related

Learn After