Non-Experimental Design or Experimental in a PhD Study: Which to Choose?

Page content

Choosing a PhD study design for your dissertation can be difficult if you don’t know the advantages and disadvantages of the different types of research design. When it comes to study designs, there are two major types.

Which study design you select for a PhD dissertation depends on your goals for the research and other study parameters such as sample size, availability, and location. Each type of research design carries with it pertinent advantages and disadvantages. Choose the wrong design and you may not be able to complete your study as planned. Read on to learn about the advantages and disadvantages of non-experimental vs. experimental PhD study designs.

A non-experimental research design is often called a correlational (or quasi-experimental) design for good reason. Non-experimental designs allow the researcher to determine if a variable tends to occur in significant proportion with another variable. Using inferential statistical procedures such as the ANOVA or linear regression, data are tested to see if the variability in a dependent variable can be explained or partially explained by the variability in an independent variable.

Advantages of a non-experimental study design

Non-experimental designs are often easy to implement because the research does not have to manipulate any of the variables or conditions of the study. Consequently, non-experimental study designs are a good choice for field and exploratory research where the relationships among variables are unknown or need to occur freely without manipulation.

Disadvantages of a non-experimental study design

What gives non-experimental designs its advantages actually contributes to its disadvantages. Since they are easy to conduct, many doctoral researchers do not put enough effort into exploring extant literature for what has already been done on the subject and often conduct non-experimental research on a whim. In addition, non-experimental designs often suffer from common method bias, specifically single-source bias, because collection of both dependent and independent variables from the same source is so easy and convenient.

Experimental PhD Study Designs

Experimental designs represent true experiments where the researcher artificially manipulates some variable, condition, or treatment in the design to isolate and better observe dependent variables. Conducting experimental studies often takes more preparation and review of what extant literature suggests about the variables under study.

Advantages of an experimental study design

One of the most important advantages of using an experimental study design has to do with establishing causation. Conducted correctly, the researcher can establish or suggest causation rather than just correlation between and among variables in the study. Many researchers conducting non-experimental research often mistake the correlations they find with causation.

Disadvantages of an experimental study design

By nature of the manipulation of a variable, condition, or treatment imbedded in experimental designs, researchers lose external validity. External validity is concerned with the degree to which experimental conditions mirror the real world. Too much artificial manipulation of a variable in experimental research may make replication of the relationships among variables possible only in a lab. With no bearing on the real world, some experimental designs are not useful for explaining phenomena outside of a lab.


The advantages and disadvantages of non-experimental and experimental PhD study designs illustrate that the doctoral student must strike a balance between an easy to implement design and a design that better isolates and captures the variables under study. Not enough planning with a non-experimental design and the PhD student risks repeating what extant literature has already established about the variables under examination. Too much manipulation in an experimental design means that the study is only valid in a laboratory.