Specialized Designs:
Single Group, Quasi Experimental & Developmental

(read: Cozby Chapter 11)
 


I. Single group Designs

The single group designs were discussed in the beginning of experimental design two topics ago under the heading of Conducting Research -IV Experimental Designs (to avoid).


While this may be the case for the strongest experimental interests (trying to establish causes), these designs may still be valuable for specific situations where the cost or other limitations lead one to make use of them.

This might be the case of experimental treatments of that have a high degree of risk, and only one or a few "patients" or subjects may be examined with the new (medical) treatment. These are called single-subject or single participant designs.

The most commonly used style of these are the "Baseline" studies where pre- and post- functioning are examined.

Another method to test the efficacy of such treatments is also to use the "Reversal Designs" where the treatment is given then taken away to see if the effect diminishes.  Here one makes use of the A-B-A style of where A is the baseline and B is the treatment.

As Cozby points out it might be unethical to remove the treatment (if there is an effect) then the ABAB design is used to reestablish the treatment after examining the effect of having removed it.


II Quasi Experimental Designs

These may also be used where it is impossible to control or manipulate the "independent variable.

This is the case where naturally occurring phenomena are examined, such as disasters or immigration or other uncontrollably events.

 

As discussed earlier they may be of one of these types:

One group Post-test only - where there is only a limited opportunity to follow-up, such as workshop or training sessions on study skills, reading or writing ability.

 

One group Pretest-post-test designs which may be very useful when testing  the effect of a new teaching method, first measure students' performance, introduce the method, and then measure performance again. Challenges here, as discussed earlier may include: history, maturation, testing, instrument decay.

Interrupted Time Series

Another Specialized type of design is the Interrupted time series design where archival records may be examined be and after an intervention or manipulation such as changing the penalties for drinking and driving or speeding. Here rates of car crashes may be used to examine the efficacy of the reformed laws.

Multiple Base-line designs may also be used where it is unethical or impractical to do reversal designs.

Here the treatment is given to participants at different times through the study. This should show the efficacy of the treatment under different circumstances, and it can be done across subjects or across behaviours.

Across subjects looks at the treatment (i.e., relaxation therapy) across various subjects getting it at different times as indicated in Figure 11.2 (p.204).

The across behaviours design will apply the treatment (behavioural technique) to the same participant(s) to different behaviours.


III. Program Evaluation

Program Evaluation is another important application of research design that may use experimental, but is more likely to use quasi-experimental designs (i.e., the efficacy of changing vending machine contents).

There are a number of aspects or stages of program evaluation that may be considered.

Needs Assessment - involves the examination of a community or individual's needs and whether or not they are being met.

Program Theory Assessment - involves the evaluation of a theory as applied to a specific program, testing its efficacy.

Process of Evaluation - examines the implementation of a program and whether or not it meets the specific needs that were assessed.

Outcome Evaluation - evaluates the impact of the program and seeks to establish impact assessment.

Efficiency Evaluation - examines the overall program and what may need to be changed or altered to meet the original needs and process and outcomes.


 

IV. Developmental Designs

 

Are used to evaluate changes in behaviour that relate to changes in chronological age or growth. 

Cross-sectional studies - Examine people of different ages at the same time. Here several groups of people of different ages will be examined and compared.

 

Advantages include the lower cost and easy availability of people  while the disadvantages include cohort effects. Cohorts are groups of people who are born and live through the same time periods, here people of different ages may be different not due to their ages, but rather to changes in culture.

 

This designs may be more useful when you are comparing ages that are closer (i.e. 2-10 vs. 20 -80)

  

Longitudinal designs - Examine the same people at different ages to see what has changed over a set period of time.

Seven-up is an example of one such study as is the 1991 "Nuns Study" where these women were followed up over their lifetime in serving the catholic church. Here archival medical and psychological records were used along with earlier written autobiographies.  Cozby also discuss Louis Terman's Life Cycle Study which began in 1921 with "gifted" children called "termites".

  

The advantages of longitudinal studies are that one can learn a great deal about intraindividual change, however attrition is problem that can lead to considerable difficulties.

The Sequential Model is a compromise between cross-sectional and Longitudinal studies where one starts off with a cross-section of ages and then follows up (at least the younger group) as they ages over a shorter period of time.