Experimental vs. non-experimental Flashcards
Experimental Research
A researcher is able to manipulate variables and subjects to identify a causal relationship. This typically requires the research to be conducted in a lab, with one group in an experimental group (being manipulated) while the others are placed in a control group (non-manipulation group). A lab-based experiment gives a high level of control and reliability.
Non-experimental Research
A researcher cannot control, manipulate or alter the variables or subjects. Instead, it relies on observations, interpretation or interactions to come to a conclusion. Typically, this means the non-exp researcher must rely on correlations, surveys or case studies and cannot demonstrate a true causal relationship. tends to have a high level of external validity - meaning it can be generalized to the larger population.
Experimental Research vs Non-experimental Research - what are the differences?
Differences between exp and non-exp include:
Methodology - exp researchers are capable of performing experiments on subjects and manipulating variables. Non-exp researchers observe and interpret what they are looking at.
With manipulation and control, exp can label causal relationships while non-exp cannot say this with certainty due to risk of confounding variables influencing results/m
Trade off between internal and external validity
These concepts affect the way you design a study because you are often faced with a trade-ff situation between internal and external validity. In order to obtain high internal V, you need a lot of experimental control e.g. lab setting. However, having more control compromises the external V, meaning less certainty that results are valid in the real world. Therefore, it is important to consider what the goal of the research is:
1. If it is appropriate to use strict exp conditions,
or
2. A sample that will be more relevant to real-world situations.
Scientific Research
The purpose is to collect information about a phenomenon in an area of interest in order to build a reliable base of knowledge around it. Involved developing theories in order to better explain, describe and organise the information that is collected.
Pseudoscience
Pseudoscience includes beliefs, theories, or practices that have been or are considered scientific, but have no basis in scientific fact. This could mean they were disproved scientifically, can’t be tested or lack evidence to support them.
Scientific Research vs. Pseudoscience
Scientific articles undergo a process called peer review, where the results are scrutinized before it can be published - this does not occur in pseudoscience.
Scientific experiments are done precisely in order for them to be replicated at a later stage; follow a method. Pseudoscience lacks precise and detailed descriptions of their methods and often don’t report their failures.
Randomised Control Trial
RCT randomised who receives a program (treatment group) and who does not (control group). It then compares the outcomes between those two groups - this comparison gives us the impacts of the program. Using randomisation approach means that a target population is first identified, then the program access is randomised within that population. The sample is first randomised into the two groups - random number generator or excel. To ensure randomisation has succeeded, it is important to check that groups are equivalent in terms of baseline indicators and contextual variables that might be important. Groups should be statistically identical.
Randomisation of group assignment
An experimental technique for assigning participants to different groups using randomisation. This ensures that each participant has an equal chance of being placed in any group. Also, helps ensure the difference between and within groups recorded at the end of the experiment can be more confidently attributed to the experimental procedure or treatment. Random assignment, blinding and controlling are key aspects of the design of experiments, as they help ensure that the results are not deceptive via confounding. RCT are vital in clinical research.
Matched Pair Design
Special case of randomised block design. It can be used when an experiment only has two treatment conditions. Subjects can be grouped into pairs, based on some blocking variables. Then within each pair, subjects are randomly assigned to different treatments. The matched pair design might be an improvement over complete randomisation design - similar to RCT control for confounding varibales, yet explicity controls for potential varibales that may affect the results - e.g. age and gender. Reduced third varibales problem as each pair is statistically similar, accounting for the possible third varible in determining the cause of an outcome.
