Week 8 - Sources of error in experimental research Flashcards
What is an unsystematic error?
Random error (also called unsystematic error, system noise or random variation) has no pattern. One minute your readings might be too small. The next they might be too large. You can’t predict random error and these errors are usually unavoidable.
What is a systematic error?
Systematic error occurs when an observed or calculated value deviates from the true value in a consistent way.
What are the two main sources of bias we need to control for?
1) Extraneous variables
2) Confounding variables
What is an extraneous variable?
These are extra variables that also could influence the dependent variable that you do not want to influence the dependent variable
What is a confounding variable?
A confounding variable is a third variable that influences both the independent and dependent variables.
What is the difference between extraneous variables and confounding variables?
An extraneous variable is any variable that you’re not investigating that can potentially affect the dependent variable of your research study. A confounding variable is a type of extraneous variable that not only affects the dependent variable, but is also related to the independent variable.
Are extraneous variables a systematic error or unsystematic error?
Unsystematic error
How is random error important in significance testing?
Probability tells us if it is significant or not. Controlling for extraneous variables gives you a higher signal to noise ratio and helps you see the effect you are interested in
How can you combat unsystematic errors?
1) Create a more powerful manipulation (e.g., increase the dosage of nicotine administered)
2) Test a larger number of people
3) Using a matched pairs design
4) Situational variables - to control for possible participant variables that could add noise to the data. You can specify inclusion and exclusion criteria
5) Power - Random error is always there; however, you can try to overcome it by making sure that your experiment has enough power to find differences between the conditions. Power is the probability of not making a type 2 error
What is Power?
The probability of correctly rejecting the null hypothesis
What does it mean if a study has high power?
The higher the power the less likely it is that we will miss a false negative
What are the ways of increasing the power of your experiment?
1) More participants - Testing more participants will increase the power of the experiment. The random error will also ‘even out’ with more participants. Ensuring you have the appropriate number of participants will mean you are less likely to make a type 2 error.
2) Sensitive measurement - Try to choose a measurement that accurately represents the whole range of possible scores. If task too easy then there will be no difference between the groups. If task too hard than there will be no difference between groups
3) Experimental design - Some designs have more power than others. This is mainly because they control for extraneous and confounding variables differently. In independent measures designs individual differences can have a large impact on the data because they vary between our experimental conditions. In repeated measures the participants essentially control for this because they do both conditions. In matched pairs these differences are partially controlled but not completely
4) Statistical test - Parametric statistics have more power than non-parametric statistics.
If a memory task is too easy that most participants get nearly all of the correct answers correct than we won’t find any differences. What is this called?
Ceiling effect
If a test is too difficult that nearly everyone gets no correct answers or only a few answers correct. What is this called?
Floor effect
Are extraneous variables a systematic error or unsystematic error?
Systematic