Descriptive Research Flashcards
What is most important for a sample?
Must be representative!
- > Characteristics of sample should fit to the population
- > Must be possible to draw conclusions from the sample to the population
Possibilities of sample procedures
Probability sampling
Non-Probability sampling
3 types of probability sampling
Stratified sampling
Simply random sampling
Cluster sampling
3 Types of non probability sampling
Judgemental sampling
Quote sampling
Other types of convenience sampling
2 Advantages of simple random sampling
Easy to implement
Works well for homogeneuous sample frames
How does stratified sampling work?
Population is divided into mutually exclusive and exhaustive subsets.
Simple random samples are taken from each subset
2 Advantages of stratified sampling
More precise -> fewer means
Every subset is represented
How does cluster sampling work?
Subdivide the population into subsamples, which are within each other homogenes and between each other heterogeneous.
Sample from a subset
1 Advantage and 1 Disadvantage of Cluster sampling
Economically efficient, but statistically not efficient (compared to other sampling methods)
When can judgemental sampling be useful?
In exploratory design
Disadvantage of quota sampling
Other important characteristics could not be represented
Other types of convenience sampling
Snowball sampling
Mall intercepts
Volunteers
Possibilities to classify data collection methods of primary data
Longitudinal
Cross-sectional
2 longitudinal data collection methods of Primary data
True panel
Omnibus panel
1 Cross sectional data collection method of primary Data
Sample survey
Definition of True panel data collection
Repeated measurement of the same variables over time
-> Enables true longitudinal or time series analysis
Definition of omnibus panel
A sample is maintained, but the collected information varies over time
-> Easy access to panellists
What do Cross sectional panel lists do?
Provide a snapshot of variables of interest at a single point in time
Allows of cross classification of the variables
-> Enables an analysis of relationships of variables
3 Forms of response
Open end questions
Dichotomous questions
Multicholtomous questions
Scales of multichotomous questions
Rating scales
Likert Scales
Semantic differential
Constant sum scale
Scales of measurement
Nominal (Non-metric)
Ordinal (Non-metric)
Interval (Metric)
Ratio (Metric)
Measurement: Woraus setzt sich der Xo zusammen?
Sum out of Xt, Xr, Xs
Xt = true score Xr = random error Xs = systematic error
What does a systematic error do?
Affects the measurement in a predictable way
What does a random error do?
Is not systematic
What does „Validity“ say?
Refers to whether we are measuring what we want to measure
What does „Reliability“ say?
Is the degree to which what we measure is free of random errors
4 Types of nonsampling errors
Noncoverage error
Nonresponse error
Data collection error
Office processing error
Univariate forms of analyzing descriptive data
Indicative -> Bar Charts, Pie Charts, Frequency table
Statistics -> Mean, Median, Variance, Chi-Square test
Bivariate forms of analyzing descriptive data
Indicative -> Scatter plots, Cross tables
Statistics -> Correlations
Definition „Population“
Group of units about which to make judgements
Definition „Sample“
Subset of selected cases from the population
Definition „Sample Frame“
List of elements from which the sample is drawn
Explanation of simple random sampling
Random selection of the number of cases required
Each population element has an equal chance of being selected
Explanation of Stratified Sampling
Population is divided into mutually exclusive and exhaustive elements
A simple random sample of elements is chosen independently from each subset (strata)
Every population element is assigned to one and only one element
