Transcript Survey Methods: Lessons from public attitudes to price fixing

Survey Methods in Law Research:
Lessons from public attitudes to price fixing
Andreas Stephan
Norwich Law School
www.ccp.uea.ac.uk
What does a survey measure
• ‘A set of procedures for collecting information and making
quantitative inferences about populations’
• Attitudes, attributes, actions
• ‘Target population’ (usually too large)
• …so we take a ‘sample’
• Those who agree to participate are ‘respondents’
What does a survey measure
‘Samples’
Smoking
Y
Others in target
Population, not
taking part in
survey
X
Income
Using a survey in Legal Research
• Should start with a specific research question / hypothesis.
• Questionnaires inappropriate for exploratory research: key is
design of the questions.
• How law: is applied / perceived / affects people’s behaviour
• Where to begin:
– Literature review (is it necessary)
– Brainstorming sessions
– Qualitative research (get to know target population)
Public attitudes to price fixing survey
Hypothesis:
•
Public perceptions of price fixing do not reflect the severe
sanctions that exist for such behaviour in law.
Research Questions:
1.
Do members of the public think price fixing is wrong?
2.
What punishments do they feel are appropriate?
3.
Do they feel the use of immunity is justifiable?
4.
Do they perceive price fixing to be ‘dishonest’?
5.
Are attitudes different towards crisis cartels?
Designing unbiased questions





Dependant variables (e.g. why do people smoke?)
Independent variables (e.g. parents, friends, age, stress, income)
Most common explanatory variables will be demographic
Many pitfalls such as reverse causality (e.g. income and
education); overly narrow definitions; time.
Some common challenges:

Respondents dedicate limited time/effort to completing survey

Making sure respondents’ understand questions

Technical / specific language and the risk of ambiguity

What can actually be measured - is survey still worthwhile?

Avoiding bias (in language and context)

Problems of asking about the past
Designing unbiased questions
(Q1):
(Q2):
Do you feel that young people lack discipline?
Should national service be reintroduced?
(Q1):
Do you feel it is wrong to force people to take up arms
against their will?
(Q2):
Would you oppose the reintroduction of compulsory military
service?
Surveys should be piloted and developed over
many months…
SEPT 06
Is it wrong for the owners of 3 corner shops to meet
up once a month and agree on what prices to
charge?
 Yes  No
MAR 07
Imagine the owners of the 3 corner shops in your
area meet once a month to agree on what prices to
charge for groceries. Shopper A believes that this is
good for their customers because it ensures similar
prices and saves them the hassle of searching each
shop for the lowest price. Shopper B believes that
this is bad for their customers because it will result
in much higher prices. Which shopper do you agree
with? Please tick () only one.
 Strongly Agree  Agree  Disagree  Strongly Disagree
 Neither  Don’t Know
Choosing an appropriate survey method
Self completion (postal)
Online
Telephone
Face to face interview (Street or omnibus)
Probability and Quota surveys: the former is preferable, but not
practicable with large populations.
Sample size and response rates are important. The lower, the more
chance that a group has been excluded, resulting in perverse
survey results.
Important factors include: cost; incentive; timing effort
Choosing an appropriate survey method
‘Samples’
Y
Others in target
Population, not
taking part in
survey
X
Interpreting the results

In law, descriptive statistics are perfectly acceptable.

In Economics, statistical analysis is expected.

But, torture statistics long enough and they will confess…

Professional ethics / responsibility
Interpreting the results
Strong Positive
No Relationship
Y
Y
X
X
Weak Negative
Y
X
Time machine: What I would have done
differently



Including more questions
Ask if each question really of any use / or did it just seem
interesting at the time.
Expanding piloting beyond CCP?
Tips for effective survey design

Define the aims of the study

Review the existing literature / research on subject

Establish the research questions / hypotheses

What is it exactly you want to measure? How? - Think ahead!

Establish a realistic budget – is survey still worthwhile?

Design survey questions – repeatedly pilot to develop/refine

Choose the most appropriate survey method / sample size.

Carry out survey (allow plenty of time for delays)

Process, analyse, interpret and report.