Issues in experimental design

Question 1

What are experiments so important?

Answer 1

• Correlation does not imply causation

* Researchers can manipulate variables (Experimental manipulation) to examine and determine possible cause and effect

Question 2

What are independent and dependent variables?

Answer 2

• Independent variable: The manipulation used by the researcher
• Dependent variable: The outcome measured by the researcher

Question 3

What are independent designs?

Answer 3

Different independent participants in each group e.g. sex differences (Men vs Women)

Question 4

What are repeated designs?

Answer 4

Same participants repeat the experiment in each condition e.g. Time of day (Morning vs Evening)

Question 5

What are the advantages and disadvantages of independent participants?

Answer 5

Advantages:
• Participation in one condition does not influence another.

Disadvantages:
• Individual differences between groups (random variance)
• Need more participants!

Question 6

What are the advantages and disadvantages of repeated participants?

Answer 6

Advantages:
• Reduced random variability
• Less participants needed!

Disadvantages:
• Cross condition effects:
- Practice effects
- Fatigue effects

Question 7

What is an ABA experimental design and what is it often used in?

Answer 7

First there’s a baseline condition measurement, then an experimental condition measurement and a repeat measurement with no condition.

Often used in treatments, the experiment might involve some sort of intervention or treatment to see if it improves or decreases scores. Measurements taken before treatment, during the treatment and after treatment. Then you see if the scores changed during the treatment and whether they stay the same after the treatment or not.

Question 8

What is longitudinal research and what are it’s advantages and disadvantages?

Answer 8

Repeated testing of participants over time.

Advantages: Changes across time more accurately measured with the same participants

Disadvantages: Resources heavy – time, money and participants

Question 9

What is cross sectional research and what are it’s advantages and disadvantages?

Answer 9

Testing of participants at different “ages”

Advantages: Relatively easy, quick and cheap to run

Disadvantages: Need far larger effects to identify significant results

Question 10

When is longitudinal and cross sectional research most relevant?

Answer 10

Slightly different terminology for independent or repeated. Most relevant in developmental or lifespan studies.

Question 11

What should you consider when thinking of using a control group?

Answer 11

• Use control groups where appropriate
E.g. No treatment, low dose, high dose

• Consider if control groups should have “treatment”?

• No treatment at all
• Placebo treatment

• In medical research, is it ethical to give a control group a placebo? What is an appropriate control?
E.g. treatments for anxiety – is it ethical to give them no treatment.

Question 12

How can you overcome the ethical problem of whether it’s ethical to give a control group a placebo in medical research?

Answer 12

In medical research, it might not be ethical to give a control group a placebo.
To overcome this, the control group tends to be given the standard treatment.
So there would be some kind of standard treatment that everybody gets within NHS guidelines and the research would be looking at whether there is a new treatment that is even better than that.

Question 13

What is an experimental design?

Answer 13

Participants can be randomly allocated to the different conditions. E.g. Allocate participants to drug group or placebo control.

Question 14

What is a quasi experimental design?

Answer 14

Participants cannot be randomly allocated; groups are pre-defined. E.g. Sex differences (male vs. female) in verbal fluency.

Question 15

In what designs do practice and fatigue effects occur?

Answer 15

In repeated measures designs only.

Question 16

What are practice and fatigue effects and what is the problem here?

Answer 16

Practice effects
• Scores increase due to practice

Fatigue effects
• Scores decrease due to fatigue

The big problem here is that we don’t know if the changes in scores are due to our manipulation or due to practice and fatigue effects.

Question 17

How can you tell if changes in scores are due to the manipulation or due to practice and fatigue effects

Answer 17

By using counterbalancing

Question 18

What is counterbalancing?

Answer 18

Counterbalancing:

Alternate forms of measures
• Multiple versions of the same test, equivalent in measurement, well validated with reliable scores.

E.g. If we had version A and version B of a test, we can give half of the participants version A first and half of the participants version B first and then give them the other test after. Randomly allocate participants to either test A then B or test B then A.

Question 19

What are the possible sources of experimenter bias?

Answer 19

• In design of experiment
• In selecting participants
• In treatment of participants during study

Question 20

What are the possible sources of participant bias?

Answer 20

• Attempt to “guess” what the experiment is about

* Attempt to “help” researcher get “right” results

Question 21

What are some ways of reducing or avoiding bias in research?

Answer 21

• Double-blind design: Neither the experimenter nor the participant know which “group” they are in, therefore they can’t adjust behaviour
• Research protocols and standardisation: Clear and repeatable procedure (might have a script of things to say to the participant to make sure they say everything in the correct order and don’t miss anything out. The script creates a standardisation across participants, furthermore if there are multiple researchers running the study, they’ll all be doing it in the same way.)
• External review: For example, experts review applications for research funding and all publications are rigorously reviewed before acceptance

Question 22

What are confounding variables?

Answer 22

• Measurable variable that may explain variability in data

* Also sometimes called: covariate, control variable…..

Question 23

What is experimental variability?

Answer 23

Experimental variability is the variability between conditions, that we cause, the “good variability”

Question 24

What is random variability?

Answer 24

Random variability is within the group, this is where we may have some confounding variables like individual differences.

Question 25

What is a way of dealing with confounds?

Answer 25

By matching participants

One way of dealing with confounds through the design of the experiment
• Identify and measure possible variables
• Then ensure they are the same in both conditions

E.g. before you start the study test how many units of alcohol they drink per week and then match similar participants and put them in either condition. This makes sure that the participants are balanced in both conditions.

This means that if there is a difference it’s due to your manipulation rather than individual differences.

Question 26

What is the difference between a one tailed and two tailed hypothesis?

Answer 26

In a one tailed hypothesis you specify the direction and in a two tailed hypothesis you do not specify direction.

Question 27

How do you choose to make a one or two tailed hypothesis?

Answer 27

Make sure your prediction follows on from the research you do in the introduction.

So, if you have an introduction that is saying for example, that memory is going to be better for early life events than for current events and it’s really consistent in research, then you should have a one tailed hypothesis and specify the direction of your prediction.

If the introduction is inconsistent and you find research that says opposite things to each other, you will then want to come up with a two-tailed hypothesis saying you think there will be a difference but not specify the direction.