2. Learning Objectives
Define 4 sampling techniques
Practice 3 of these using Skittles
Evaluate the different sampling
techniques
3. Big Picture
• Takes too long to study
everyone
Population • Use a sample of the
population to represent
everyone
Sample
• Key word:
representative
4. Expert Input
Different sampling techniques
Stratified
Random
Volunteer
Opportunistic
5. Student Led
How to do it?
Follow the instructions on the worksheet
You will need a calculator!
Colour Number
Red Ask if you need help
Green
Yellow
Orange
Purple
7. Sampling procedures: Simple random
sampling
• Simple random sampling: E.g. pulling names out of a
hat. Though psychologists and sociologists tend to
number all the names in the sampling frame, then get a
computer to randomly select the numbers. It is better to
say this in exam than say names out of a hat.
8. Strength and weakness
Low bias because everone has equal chance of being chosen so you cant
knowingly or unknowingly introduce a bias by eg putting the better
looking and perhaps more confident people in 1 condition/group if
confidence was a factor in your d.v
Sample can be checked mathematically for any bias
- for true random sampling you need a perfect list for your sample frame on
a practical level this is almost impossible to achieve even for people with a
huge amount of funding ie even the electoral register is not a perfect list
of everyone in the uk.This can cause bias as those missing may be a
specific type of people.(eg homeless or those with mental health
problems or non comforming types.)
- Even if you have the perfect list once you have made your random choice
of participants it is very unlikely that you will get access to everyone you
originally choose .(again bias potential as it is particular type of person
that refuse to be involved)
9. Volunteer self selecting
Asking people to volunteer. If questionnaires
are sent out by post those returning them
would e considered volunteers. Milgram used
volunteers and he paid them for their time,
10. Strengths and weaknesses
Ethically good because people have freely put
themselves forward to take part
You are more likely to get full cooperation
- Only certain types of people may volunteer and this
is likely to cause a serious bias in the results and
limits the generalisability of the findings in certain
types of experiments especially those of a social
nature
- It may take a long time to get enough volunteers in
comparison to an opportunity sample
11. Opportunity sample
This is when the researcher takes who ever is
available. If the researcher asks someone
directly to take part and they agree this is an
opportunity sample and not a volunteer
sample. If a general advert is put out and
people put themselves forward then that
would be a volunteer sample
12. Strengths and weaknesses
It is ethical as long as pp given full information of the
study they are asked to participate in also the
experimenter can judge if the person is likely to be
upset an only target suitable participants
A good sized sample can be obtained far more
quickly than a volunteer or random sample
- Using only who is available may mean you only have
a certain group of people eg mums at home or
students who visit the library
- Open to the experimenter choosing a specific group
ie people they are confident to approach
13. The sampling frame is divided into ‘strata’. A
random sample is then taken from each
stratum .You generate categories that may
be relevant to the nature of the study such
: Stratified sampling
as age gender ethnicity and then ensure you
have a certain number from each. I t may be
400
a number proportional to the sample frame students
50% male 50% female
200 males 200 females
75% white 25% ethnic minority
75% white
25% ethnic minority
50 ethnic minority 150 white 50 ethnic minority 150 white
males males females females
5 ethnic 15 white 15 white
minority 5 ethnic
males females
males minority
14. Strengths and weaknesses
All the relevant groups or strata have at least some
representation
Limits the number of participants needed compared with
random sampling to be sure of having some representaion
from each group/strata
- The researcher is really unlikely to be aware of every group
and sub group so you are attempting the impossible and in
doing so may create a bias by missing out key groups and even
over representing certain groups that have similarities which
is less likely to happen in a random sample
- It is difficult to get a truly representative number for each
group and could end up with a bias in terms of proportions