The Student Room Group

Need your feedback on this...

Is this considered an acceptable design?

The type of experimental design used in this study is “within participants”, where all 135 of my participants will do the experimental condition (mental stress task).

"Stress" (independent variable) was controlled by exposing participants to a 50-minute stress mental task, which most of it contained "insoluble" anagrams, in addition to mathematical problems that cannot be solved without a calculator, all of this so the participants get the "below average" feedback.
"Headaches" (dependent variable) on the other hand are measured using a 5-point "headache rating scale" (0 = no headache; 5 = severe headache) that participants fill out at the beginning and at the end of the mental stress task.

This experimental design is trying to analyse if "stress and headaches are related" to each other.


Thanks.
You need a control condition, a more sensitive DV and generally a less obvious design to reduce the likelihood of demand characteristics.
Reply 2
Original post by GodspeedGehenna
You need a control condition, a more sensitive DV and generally a less obvious design to reduce the likelihood of demand characteristics.


Can you just explain this more further if possible? For example, what do you mean by control condition.

And, isn't headache an acceptable DV?

How can I make a less obvious design? And, why is that? What do you mean here by reducing the liklihood of demand characteristics?

Thanks.
(edited 13 years ago)
[QUOTE="SWEngineer;30578650"]Can you just explain this more further if possible? For example, what do you mean by control condition.

And, isn't headache[/I[ an acceptable DV?

How can I make a less obvious design? And, why is that? What do you mean here by reducing the liklihood of demand characteristics?

Thanks.


A control condition is where you have not manipulated the IV. This is the baseline level of "headache" to which you will compare the level obtained in your stressful group. You then need to be able to show that, statistically, this difference has not occurred by chance.

Your headache scale will inevitably get lots of zeros as, most of the time, healthy people do not have headaches and so will rate zero most of the time. Additionally, if you can induce a headache via stress, what are the chances that it will come on during the period that you test. Similarly, it may require chronic and extended periods of inescapable stress that causes a headache. Having an arbitrary one hour test probably won't do much.

Your design is too obvious. It's going to be obvious to participants that the problems you have given them aren't suitable and they will guess that something is up. Subsequently, they won't put that much effort in and won't get that stressed. Then when you give them a headache questionnaire, it will be very obvious what you are trying to do and the participants will either go with it or resist it. Neither is good. These are demand characteristics. Additionally, you need some sort of confirming measurement to actually observe whether or not you are succeeding in stressing them out. E.g. A galvanic skin response or something.
Reply 4
Original post by GodspeedGehenna
A control condition is where you have not manipulated the IV. This is the baseline level of "headache" to which you will compare the level obtained in your stressful group. You then need to be able to show that, statistically, this difference has not occurred by chance.

Your headache scale will inevitably get lots of zeros as, most of the time, healthy people do not have headaches and so will rate zero most of the time. Additionally, if you can induce a headache via stress, what are the chances that it will come on during the period that you test. Similarly, it may require chronic and extended periods of inescapable stress that causes a headache. Having an arbitrary one hour test probably won't do much.

Your design is too obvious. It's going to be obvious to participants that the problems you have given them aren't suitable and they will guess that something is up. Subsequently, they won't put that much effort in and won't get that stressed. Then when you give them a headache questionnaire, it will be very obvious what you are trying to do and the participants will either go with it or resist it. Neither is good. These are demand characteristics. Additionally, you need some sort of confirming measurement to actually observe whether or not you are succeeding in stressing them out. E.g. A galvanic skin response or something.


To be more specific, my design should include the following:

The type of design you used (e.g. between, within, mixed or matched subjects)
The independent variable(s) including the conditions you selected to represent different levels of the IV
The dependent variable including details of the units of measurement used (e.g. seconds, milliseconds, or number of correct responses)
The experimental hypothesis

Does my design above meet those requirements? Or, what is missing?
Original post by SWEngineer
To be more specific, my design should include the following:

The type of design you used (e.g. between, within, mixed or matched subjects)
The independent variable(s) including the conditions you selected to represent different levels of the IV
The dependent variable including details of the units of measurement used (e.g. seconds, milliseconds, or number of correct responses)
The experimental hypothesis

Does my design above meet those requirements? Or, what is missing?


Sure, but it doesn't mean it's a good experiment.

What is this for? Are you doing an assignment or something?
Reply 6
Original post by GodspeedGehenna
Sure, but it doesn't mean it's a good experiment.

What is this for? Are you doing an assignment or something?


It is just part of a coursework, so it is not a complete experiment, st designing one.
Original post by SWEngineer
It is just part of a coursework, so it is not a complete experiment, st designing one.


Have you not been given any research method resources or anything?
Reply 8
Original post by GodspeedGehenna
Have you not been given any research method resources or anything?


I just heard about PubMed.
Reply 9
Any more feedback on the design above?

Thanks.

Quick Reply

Latest

Trending

Trending