Hoofdstuk 8 treadwell (samenvatting Flow) - Sampling, who, what and how many? Flashcards

1
Q

Census: a

A

a study of every member in a population.

Tegenovergestelde van een sample

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Sample:

A

a selected segment of a population presumed to represent the population.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are Nonprobability samples:

A

sampling based on a judgment by the researcher. It contains five
other sorts of sampling.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Which five types of nonprobability samples are there?

A
  1. Convenience sampling
  2. purposive or judgmental sampling
  3. Quota sampling
  4. network or snowball sampling
  5. Volunteer sampling
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Convenience sampling

A

sampling based on convenience to the researcher. Can be
useful when pretesting a study or when the results of the research are not intended
for scholarly publication. Often used when there’s not many time or money.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Purposive or judgmental sampling:

A

sampling based on a specific (group of) person(s)
that match criteria the researcher may have. E.g.: when you think only the cook in a
restaurant knows everything about the hygiene and quality of the food and no one
else does, or when you reach out to the group of students that cannot eat at the
campus restaurant because of their religion.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Quota sampling:

A

sampling that attempts to replicate in a sample the features that the
researcher thinks are important in the population. For example: you are going to ask
students who live on and off campus. 80% of students live on campus, so you’ll ask 8
students from campus and 2 students who don’t live on campus. One important
feature is replicated now, but the students in the sample are not randomly selected,
it’s the researchers judgment.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Network or snowball sampling:

A

sampling using members of a network to introduce
you to other members of the network. Like finding one vegetarian for your sample who
introduces you to more vegetarians. Sample may consist of people with same opinions
and there is less diversity because friends are recommending friends.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Volunteer sampling

A

obtaining a sample by asking for volunteers. Not always good
because you are by definition recruiting one type of person; volunteers. Research
results are biased because you did not ask non-volunteers.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Which types of sampling can also be used for non-human subjects (like media content?)

A
  1. Convenience sampling
  2. Purposive or judgmental sampling
  3. Quota sampling
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Sampling frames:

A

the master lists from which a sample is selected. For example: the
membership list of a club or all registered members of a political party.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Sampling units:

A

units selected for the study. For example: for communication studies the units
will often be individuals, but it could also be couples, corporations, comic strips, phone
numbers or episodes from Star Wars.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Probability samples

A

sampling generated by random selection of the sample units. Reduces
the bias in sampling. There are four kinds of probability samples.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Random sampling:

A

sampling in which every member of a population has an equal
chance to be selected and in which selection is determined by ‘luck of the draw’ rather
than a decision by the researcher. Like throwing a dice or picking names out of a hat.
You could, for example, number all students and pick some numbers using a random
numbers generator to choose the sample. A random sample is not automatically
diverse too!

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Stratified random sampling:

A

sampling in which randomly selected units from small or
minority populations are forced into the sample to ensure that they are represented
in proportion to their presence in the population.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Systematic sampling:

A

sampling every nth person on a list. For example: taking every
10th person or every 100th person listed in a phone book. The interval you select is the
sampling interval. A problem with systematic sampling could be that there is a pattern
in the population that matches the sampling interval.

17
Q

Multistage cluster sampling:

A

sampling based on first sampling large units such as a
state or province, then sampling smaller units such as a town or a city block, and so on.
Advantage: the relative ease o identifying people or households. Disadvantage: every
state of sampling has the potential for bias in the final sample increases. For example:
same-sex marriage is in some placed in the US not aloud, so you’ll not find them in
some states but you’ll find more of them in some other states. Multistage cluster
sampling can completely eliminate these couples or overrepresent them.

18
Q

The perfect sample size depends on:

A
  1. Your level of resources. Every research project is a balancing act between striving for
    the ideal sample and the constraints of resource limitations like money and time.
  2. The nature of your research. If your survey is informal, like you just want to get an
    overview of what people are thinking, or you want to pilot a survey the sample size is
    less of an issue.
  3. The level of (statistical) confidence that you want or need in your research results. If
    you want a 100% confidence in your results, you will not be sampling at all but research
    the entire population; that is a conduct census.
19
Q

Higher confidence intervals demand

A

larger sample sizes.

20
Q

Homogeneity:

A

the degree of ‘sameness’ in the population. The greater the homogeneity, the
smaller the sample size required.

21
Q

Formula for required sample size:

A

n = pq / s2

n = sample size
p = probability of event, in decimal form
q = p-1
s = sampling error, in decimal form (the standard deviation
of the sample or the standard error)

22
Q

How are p and q related to represent the homogeneity of the population?

A

p and q are related and together represent the homogeneity of the population. They represent
the probability that a population is likely to split on an issue. If voters are split 50-50 on a
referendum for example, p and q each have a value of 0.5. If 90% of the vote something down,
then p will become 0.9 and q will become 0.1. Together they are always 1. The value of p and q is at it’s greatest if they are both 0.5, which means that the population has minimum
homogeneity.

23
Q

Random digit dialing:

A

a procedure that involves dialing computer-generated random
phonenumbers in the hopes of reaching unlisted numbers.