Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Experimental Design

Specific to Experimental Design
by

greg miller

on 3 October 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Experimental Design

experimental Design First, how do we select participants? Population: e.g. all people, all children, all high school students.

Would be best to test everyone. Would allow us to be very confident in our result. But that would take WAY too much time, money, etc. So we use a sample. (only some of the population) Representative Sample - very important

Random sampling - all members of population have same chance of being selected.

Random sampling a great way, but very rarely done. Also time consuming and expensive We usually use a sample of convenience. These are samples of people who are easy to find and collect.

(often college students) Population
(all members of the group) Random Sampling Sample of Convenience Best Option Next Best Option Not the Best, but Easy and Cheap Experiment

Changing only one variable to see if the other changes as a result. Usually testing for a difference between two groups, with only ONE difference between them. To find CAUSES.

Adv. + Dis. ? - Only way to determine causes
-Strong conclusions possible if designed carefully

- Sometimes impossible to do: practically or ethically (smoking)
- hard to control confounding variables Two things needed for it to be an experiment:

1. variables controlled or manipulated
2. participants randomly assigned to groups

Often you have an "experimental group" and a "control group." The experimental group gets the treatment Independent Variable (IV) - the one we control / change

Dependent Variable (DV) - the one that might change because of what we did. We measure this to see if there was a change.

Confounding Variables - Things that might influence the effect. A difference between the groups OTHER THAN the IV. You Try!

Find IV, DV, and any confounding variables.

E.g. Sleep and Memory Study. We will control how much sleep the participants get, and then measure their memory to see if getting less sleep will reduce memory ability. We put the men (who are older) in the 8 hours of sleep group, and the women (who are younger) in the 4 hours of sleep group. After their sleep, we test them in different rooms to see how well their memory works. Random Assignment to groups helps a lot. We can assume the two groups are equal if the people have been randomly assigned to the EXPERIMENTAL and CONTROL groups.

Doesn't necessarily mean they are REPRESENTATIVE of the POPULATION, just that they are equal to each other. e.g. Women who smoke during pregnancy and birth weight. Math used to describe, measure correlations, or test for causes. Validity - does your measurement tool measure what you want it to?

Reliability - does your measurement tool show the same results if used again and again? Blind Study - participants don't know what group they're in (experimental or control)
Reduces Response Bias / Placebo Effect

Double Blind Study - participants AND researchers don't know which group the participants are in.
Reduces Experimenter Bias

Counterbalancing - doing different tests in different orders for different participants. To reduce Order Effects

Order effects - maybe participants do better on test 1 because we always do it first, not because it's better. Therefore some will do test 1 first, some will do test 2 first. Placebo - sugar pill
Effect - thinking there will be an effect makes one happen

but not only for medical trials. Can effect many psychological processes.

e.g. "This music will make you work faster" Ethics Psychologists need to follow strict ethics rules.

They are stronger now than they have been in the past. IRB - Institutional Review Board
ERB - Ethics Review Board Some important guidelines:

- respect dignity and welfare of participants
- protect participants from harm, risk, or danger
- if these occur, fix / remove them (repair any harm)
- obtain informed consent / inform about right to withdraw


- ensure confidentiality
- only use deception if it's the only way, and if there is valueable info to be gained
- debrief to inform participants all about study and give them all info, and explain why any deception was used (including confederates) Some studies we'll talk about later would NOT pass ERBs today. E.g. Milgram Obediance Studies (1965) Ethics for Research on Animals:

- must also get ERB approval
- treat humanely and care for animals
- keep discomfort, illness, and pain to a minimum, and only allowed if there is valuable info to be gained, and no other way to get the info. Z score of 1 = the standard deviation

if data is normally distributed, z scores can show us how many data points will fall between each range.

e.g. if SD = 10, and average is 40 then 68% will have a score between 30 and 50. Remember, DV DEPENDS on the change we made. OPERATIONAL DEFINITIONS:

How you define your variables for the purpose of your study.

Need to be REALLY SPECIFIC and CLEAR so other scientists can also try the same thing you did (replicate)

e.g. violent acts = any contact between the participant's hands, feet, or an object held by the participant, and the doll.
Full transcript