Week Three Flashcards
population
a group of people about whom one would like to draw some meaningful conclusions.
sample
a subset of that population that is actually included in your research study
○ A full set of elements or people from which the sample was selected.
- To make generalisations of the sample it is important that your sample is representative of the population.
sampling frame
a list of members/elements of a population from which one might obtain a sample
○ Electoral role
○ Telephone directory
○ Etc.
sample statistic
a numeric characteristic of a sample (measured)
population parameter
a numeric characteristics of the population (often not known).
response rate
what proportion of people responded?
sampling error
the difference in value between the sample statistic and the population parameter (depends on sample size).
- The smaller the sample, the larger the sampling error. - Thus, if you are trying to relate a result back to the population, the less you have of that population, the more incorrect your data will be.
causation
Criteria for identifying a causal relationship
- Cause (IV) must be related to the effect (DV) (relationship condition). - Changes in IV must precede the changes in the DV (temporal order). - There must be no other plausible explanation for the effect.
probability sampling
- A way to ensure your sample is representative of the population.
- Basic principles:
○ A sample will be representative if all members have an equal chance of being selected in the sample.
○ Allows researcher to calculate the relationship between the sample statistics and the population parameter.
- Basic principles:
simple random sample
probability sample
Each member has an equal and independent chance of being selected.
systematic
probability sample
§ Every kth person
□ Randomly select the first person then divide the size of the population by the size of the desired sample, use this to determine the interval at which the sample is selected.
® E.g. for a sample of 1000 people form a list of 10,000 select the first then select every 10th person.
Need to ensure the list is not arranged in any way.
stratified probability sample
§ If you want to make sure the profile of the sample matches the profile of the population on some important characteristic.
§ Divide population into strata and randomly sample from the strata.
§ Used to reduce sampling error by ensuring ratios reflect the actual population
§ To ensure that small subpopulations are included in the sample.
multistage cluster probability sampling
§ Begin with a sample of groupings and then sample individuals.
Large sample obtained first in order to identify members of a sub-sample
advantages to probabillity sampling
○ Helps overcome sampling bias
○ Representativeness
disadvantages to probability sampling
○ Can be a big non-response rate (can be very expensive).
non-probability sampling
- Is sometimes more desirable because it is easier and quicker.
- Not every member of the population has an equal chance of being part of the sample.
- It is used when there are no lists etc. for the population under study
convenience non-prob sample
○ Convenience: a sample of available participants.
○ Advantages;
§ Easy
§ Inexpensive
○ Disadvantages
§ No control over representative, thus, any results would need to be interpreted with caution.
snowball sampling
non-prob
○ Used mainly for hard to study populations
§ Involves collecting data with members of the population that can be located and then asks those members to provide information for other members of the population.
quota
non-prob
○ Non-probability sampling equivalent of a stratified random sample
Want to reflect relative proportions of a population but you aren’t able to sample randomly from each strata as you do in stratified random samples
judgement/purposive sampling
non-prob
○ Clear purpose to the sampling strategy: select key informants, atypical cases etc.
○ Often used to
§ Select cases that are especially informative
§ Select cases in a difficult to reach population
§ Select cases for in depth investigation.
disadvantages to non-prob sampling
In non-probability you lose an element of representativeness.
increasing sample size
- More complex the analysis the larger the sample you require.
- Increases in sample size bring with them increases in accuracy/precision, reduces sampling error.
- Heterogeneity of the population, greater variation in the population, the larger the sample should be.
might need a large sample when…
○ Heterogeneous
○ You want to break the sample into subcategories
○ When you expect a small effect or weak relationship
○ When you use less efficient methods of sampling
○ For some statistical techniques
○ If you expect a low response rate.
validity
are we measuring what we think we are measuring?
reliability
The consistency or repeatability of your measurement
test retest reliability
§ Does the test measure the same thing every time you use it?
§ Addresses the stability of the measurement.
§ However cannot always be used as often there are carry over effects.
□ People can get bored
□ Can get better or worse at something
split half reliability
§ Administer a single measurement at one point in time
§ You split the measure into 2 and correlate the scores
§ Strength: eliminates memory and practice effects.
Limitation: are the two halves equivalent?
cronbach’s alpha
§ Measure of reliability
§ Coefficient ranges from 0 to 1
Closer to one it is, the more reliable.
inter-rater reliability
§ Do different raters measure the same thing?
§ Checking the match between two or more raters or judges
§ Calculation:
□ Nominal or ordinal scale (more common) - the percentage of times different raters agree
□ Interval or ration scale- correlation coefficient
face validity
§ Most subjective type of validity
§ Does it look like it measures what its designed to measure?
§ On the face value.
§ Weak method but good first step
content validity
§ The extent to which the measure represents a balanced adequate sampling of relevant dimensions.
Does it cover all aspects of the construct that it is aiming to measure?
criterion validity
- Involves checking the performance of your measure against some external criterion
§ Concurrent: does it relate to a known criterion that already exists?
□ Establish the validity of your measure by comparing it to a ‘gold standard’ (existing measure of the same construct).
§ Predictive: does the measure predict something later on down the track?
□ Does the measure predict something that its theoretically supposed to predict?
construct validity
- Establishes validity by showing that your measure relates to other constructs in a way that you would expect.
§ Convergent:
□ Measures of constructs that theoretically should be related to each other, are, in fact, observed to relate to each other
® There is correspondence or convergence between similar constructs.
§ Discriminant/divergent
□ Demonstrates that your scale doesn’t relate to something it should not relate to.
□ Trying to demonstrate that it is separate to something else.
□ Measures of constructs that theoretically should not be related to each other, are, in fact, observed not to relate to each other (i.e., you should be able to discriminate between dissimilar constructs)