Quasi Experiments & Other Designs

Definition (#f7aeae)

Important (#edcae9)

Extra (#fffe9d)

Simple-Case Design:

  • The subject’s behaviour is measured over time during a baseline control period

  • A treatment period follows in which the experimental manipulation is introduced and the subject’s behaviour is observed.

  • Single-case designs are used in Applied Behavioural Analysis (ABA) or Behaviour Modification techniques.

Applied Behavior Analysis:

  • Discrete Trial Training (DTT): Skills are taught in small, repeated steps with clear prompts and rewards.

  • Task Analysis: Breaking complex tasks into smaller steps and teaching each step one by one.

  • Positive Reinforcement: Rewarding desired behaviours to increase their likelihood.

  • Prompting and Fading: Giving guidance or cues to help perform a behaviour, then gradually reducing the support.

Problem: Could be many explanations for the change other than the experimental treatment.

Solution: A simple reversal design takes the following form, known as the ABA design, or a withdrawal design.

ABA Method:

A (baseline period) → B (treatment period) → A (baseline period)

  • The ABA design can be greatly improved by extending to an ABAB design, the experimental treatment is done a second time.

  • Justification: A single reversal may not be sufficient evidence to explain the effectiveness of a treatment.

  • With repeated demonstrations of treatment effectiveness, random or coincidental events become less likely to be responsible for the consecutive.

  • Justification: Unethical to end the design with the withdrawal of a treatment if it’s been very beneficial for the participant.

Quasi-experimental Design:

To study the effect of an IV in settings where the control features of true experimental designs cannot be achieved.

  • The design allows for the effect of IV on DV to be examined, causal inference is much more difficult.

  • Likely to occur in natural/applied settings.

  • Quasi-experimental designs lack important features of true experiments such as random assignment to conditions.

One-Group Posttest Only Design:

  • A single group receives a treatment and then their outcome is measured only once after the treatment, with no pretest or control group for comparison.

  • Difficult to establish causality due to threats like maturation or history effects.

One-Group Pretest-Posttest Design:

Measurement is conducted before & after treatment period, and found an increase in knowledge.

There are still several threats to the internal validity of this design:

  1. History: Any event that occurs between the first and second measurements but is not part of the manipulation (confounding variables).

  2. Maturations: Any changes that occur systematically over time. Any time-related factors result in a change from the pretest to the posttest.

  3. Testing: Simply taking the pretest changes the participant’s behavior, causing a reduction. The reduction found on posttest, is a result of taking the pretest in the first place.

  4. Instrument decay: Over time, an observer may gain skills, become fatigued, or change the standards of the observations.

Non-equivalent Control Group Design:

  • Use of a control group, but participants in the two conditions: experimental & control group; are not equivalent.

  • The 2 groups are not the result of random assignment.

  • Selection bias occurs when participants who form the two groups in the experiment are chosen from existing natural groups. A problem arises because participants may differ in important ways.

  • Advantage: Knowing the pretest scores. Being able to evaluate to what extent the groups were the same on the pretest.

Developmental Research:

Study of ways individuals change as a function of age.

2 methods for studying individuals ages:

  1. Cross-Sectional Method: Sharing similarities with independent measures design.

  2. Longitudinal Method: Sharing similarities with repeated-measures design.

Comparison of longitudinal & cross-sectional:

Longitudinal:

The same group of people is observed at different points in time as they grow older.

Participants in a longitudinal study may move, die, or lose interest in the study.

Researchers have to convince participants to remain in study, continue collecting data, and gather enough participants and requires resources.

Cross-sectional:

Persons of different ages are studied at only 1 point in time.

Cross-sectional methods are more common, less expensive and immediately yields results.

  • The researcher must infer that differences among age groups are mainly due to age.

  • Cohort: group of people born around a similar period and influenced by the same demographic trends.

  • Differences among groups of different ages may reflect developmental age changes; Differences may result from cohort effects.