Causal-comparative/Ex Post Facto Design
Causal-comparative/Ex Post Facto Design
Causal-Comparative Research Design Introduction and Focus - While causal research is experimental research designed to compare groups in a more natural way, causal comparative research design attempts to identify causes or consequences of differences in a non-experimental setting. These differences already exist, and their impact on the outcome is identified by comparing groups. Causal-comparative designs can have different foci: (i) exploration of effects, (ii) exploration of causes, and (iii) exploration of consequence
When do we use the design? – This design is used as an alternative to experimental design because sometimes the latter is expensive, non-feasible and difficult to conduct and while in experimental design the independent variables are manipulated, in causal-comparative design the predictors are not. Causal research uses different terms: ex post facto studies gather data retrospectively (e.g. given the obvious effects of smoking, the researcher will look in the past to find the potential cause), causal comparison where data are gathered from pre-formed groups and the independent variable is not manipulated in the experiment. For this, the researcher will have either to find a population on which the data are available, or to find an already existing appropriate group.
Review of literature suggests that there are instances when causal comparative design” and “ex post facto design” are not clearly defined and the expressions are used interchangeably. While both designs are non-experimental. ex post facto design refers to studies that use extant or secondary data (i.e. data that has already been collected while in studies using causal comparative design data are obtained from pre-formed groups and the independent variable is not manipulated as it is done in experimental studies (Laura M. O'dwyer, 2013)
Additionally, some students make a wrong assumption that the word ‘causal’ means that the design allows causal inferences when it does not, but this case is very exceptional and not compare to the true (Laura M. O'dwyer, 2013; Leedy & Ormrod YEAR)
Type of problem appropriate for this design – The type of problem that this design addresses. should relate to the impact or effect of X on Y. This type of design has some similarities with the correlational design. Both designs are suitable when conducting an experiment is either impossible or unethical. Both try to establish relationships among variables, but the main difference is that causal comparative will compare two or more groups after one of the groups has been exposed to some treatment and/or condition (e.g. new training or intervention) ) and the design will be used to compare the grades and/or GPA in two or more groups. in causal comparison, we compare groups, while in correlational design we attempt to explain one variable by the other. (Brewer & Kuhn, 2010)
Theoretical framework/discipline background: Causal-comparative design is used in a number of context
Specific Characteristics –
Sample Size - Causal-comparative studies use different types of data analysis (e.g. if 2 groups are compared then a t-test or a one-way ANOVA is appropriate, or is more than 2 groups are compared and ANOVA is appropriate) and the sample size should be calculated accordingly. The sample size can be calculated using G*Power.
Sampling Method – In causal research samples will be selected because they will have certain characteristics and as stated above will be non-equivalent. The researcher can construct groups or use performed groups. For instance, if the researcher is interested in studying the effects of meat eating on health, he/she can construct three groups of meat eaters, fish eaters and vegetarians so that comparisons between them are conducted related to the effects of meat eating. An example of using a performed group could be when in an inclusion program the researcher is interested in studying the use of classroom time by disability students. In this scenario the researcher will use non-disability students as a control group, but will not have freedom for assigning students to control or experimental groups if the researcher will be studying both categories of learners in the same class. Another way of sample selection might be to find a population that already has the data (e.g. effect of smoking) or to find an existing appropriate group to study the effects of the cause.
Data Collection – Causal research will use questionnaires, tests, interviews as data gathering procedures.
Data Analysis – In causal-comparative design t-tests, ANOVA and its variations and a chi-square could be used. The most common tests used are paired-samples t-test, independent samples t-test, one-sample t-test (if we are interested in comparing results of one group to a norm), and the different variations of ANOVA. Chi-square is used when the level of measurement of the variables is nominal (Salkind, 2010).
Causal comparative designs have the same limitations as any other design undernon-experimental research that is the independent variable cannot be manipulated and the researcher has no control over other variables that can be impacting the dependent variable. Additionally, is impossible to choose the experimental groups since the events have already occurred. (Salkind, 2010).
Write up Results – The results are being reported in the following way:
Reporting an independent samples t-test:
In the same school, students participating in afterschool program tested on ELA state assessment higher (M = 121, SD = 14.2) than did those not participating in afterschool programs (M = 117, SD = 10.3), t(44) = 1.23, p = .09.
Guidelines, Tips, FAQs
Link to the Design Discussion Threads
Brewer, E. W., & Kuhn, J. (2010). Encyclopedia of Research Design. Thousand Oaks Thousand Oaks, California: SAGE Publications, Inc. Retrieved from http://sk.sagepub.com/reference/researchdesign. doi:10.4135/9781412961288
Krathwohl, D. R. (1993). Methods of educational and social science research: An integrated approach. Longman/Addison Wesley Longman.
Laura M. O'dwyer, J. A. B. (2013). Quantitative Research for the Qualitative Researcher Retrieved from https://bookshelf.vitalsource.com/books/9781483320663
Leedy, P. D., & Ormrod, J. E. Practical Research: Planning and Design Retrieved from SAGE Research Methods database Retrieved from https://bookshelf.vitalsource.com/books/9781323103036
Salkind, N. (2010). Encyclopedia of Research Design. doi:10.4135/9781412961288