Primary Data Collection Flashcards
Two types of data
- Qualitative data
- Quantitative data
- discrete
- continuous
Levels of Measurement of Data
- Nominal data
- Ordinal data
- Interval data
- Ratio data
Nominal Level Data
‘Lowest’ a.k.a most basic measure of data
No particular order to the labels
E.g. Classifying M&M’s by colour
Mutually Exclusive
Exhaustive - each item must appear in a category
Ordinal Level Data
Level higher than nominal
Mutually Exclusive & Exhaustive
Ranked/ ordered according to the particular trait they possess
E.g. superior, good, average, poor, inferior
Interval Level Data
Level higher than ordinal
Difference between values is a constant size
Zero is just a point on the scale. It does not represent the absence of the condition.
Scaled according to the amount of the characteristic they possess
E.g. Temperature
Ratio Level Data
Highest level of measurement
Zero point is meaningful - it reflects the absence of the characteristic
The ratio between two numbers is meaningful
Simple random sampling
A sample is selected so that each item in the population has an equal chance of being selected
e.g. names in a hat
table of random numbers
Systematic random samopling
Items or individuals of the population are arranged in some way, e.g. alphabetically. A random point is then selected, and then every kth member of the population is selected from the sample
Stratified random sampling
A population is divided into subgroups, called strata, and a sample is selected from each stratum. Either a proportional or non-proportional sample can be selected. In proportional sampling, each stratum is in the sample proportional as in the population
Cluster sampling
With cluster sampling, the researcher divides the population into separate groups, called clusters. Then, a simple random sample of clusters is selected from the population. Every member of the cluster is then sampled
Quota sampling
Sample whoever you meet until the quota is filled
Error in Surveys: Sampling Error
It is unlikely that the mean and standard deviation of the sample will be identical to the mean and standard deviation of the population.
The difference between the sample statistic and the population parameter is called sampling error.
Sampling Error
Sampling Error
Random Sampling Error
Error due to chance, may need to increase the sample size
Non Random Sampling Error
Target population not properly identified
Non-Sampling Error
Non Sampling Error Non-Response Error If non-respondents differ systematically to respondents then we cannot generalise or make assumptions about the population based on the sample Remedies include Call backs or follow ups Reducing the cost to the respondent Inducements Good survey methods
Types of questions
- Dichotomous
- Multiple choice
- Open ended
Dichotomous Questions
Allow for two answers only; for example true/false, yes/no, male/female etc.
Examples: Are you living in student accommodation Yes or No Are you Male or Female
It is important not to overuse dichotomous questions where there might be middle ground or a “don’t know” answer
Multiple Choice Questions
Respondents can choose from several indicated possibilities (included in this category are those questions which may ask the respondent to rank choices).
It is important that the possibilities are both ‘comprehensive’ and ‘mutually exclusive’ (that is every respondent should, unless otherwise stated be able to pick one and only one of the possibilities)
Open Ended Questions
The main advantage of this type of question is that it allows for an infinite number of divergent answers.
This, however, is also their greatest disadvantage, as the responses to such questions are the hardest to process and analyse.
Therefore one needs to be wary not to use too many on any one questionnaire.
Advantages of Questionnaires
Allows contact with otherwise inaccessible respondents
Incentives may be used
Low cost option
Larger geographical area does not increase costs(by email)
Minimal staff – not labour intensive
Anonymous
Respondents have time to think about answers
Disadvantages of Questionnaires
Low response rates No interview intervention Cannot be long or complex Accurate mailing list required Perhaps directions needed if questionnaire is on internet. Reponses can be skewed, ie biased.