Week 5 Flashcards
Process of Sampling
- Define process
- Determine the availability of a good sampling frame
- Decide on the sampling design.
Types of sampling designs
Probability sampling and non-probability sampling
Types of probability sampling
- Simple random sampling - every element has equal chance to get to sample
- Systematic sampling - choosing every i’th element for sample
- Stratified sampling - divide population into groups then simple random sampling within each group
- Cluster sampling - divide population into heterogenous groups then simple random sampling
Types of non-probability sampling
- Convenience sampling - Select subjects who are conveniently available
- Quota sampling - fix quota for each subgroup
- Judgement sampling - select subjects based on their knowledge/professional judgement
- Snowball sampling - getting referrals from first respondees
Operationalization/Measurement meaning
process of turning conceptual variables (happiness, etc) into measurable observations
Ratio vs Interval variable
ratio - has lowest point of 0 (absence of something), interval does not (temperature 0 unequal to no temperature)
Coverage error meaning
sample frame mismatching population
Coverage error types
Undercoverage - true population members are excluded
Miss-coverage - non-population members are included
How to deal with coverage error
If you notice and the coverage error is small, acknowledge it in research, but ignore. If coverage error is large, redefine population in terms of the sampling frame
Four requirements needed for causality claim:
- Existing correlation
- Measured cause needs to come first before the measured effect
- Control for confounding variables
- Existing explanatory and logical theory
Research strategies
Quantitative & Qualitative
Quantitative research strategy types
- Archival research (correlational) - research based on secondary data
- Survey research (correlational) - Respondents typically read the questions and record their answers themselves
- Experimental research (causal) - one or more independent variables are manipulated
Qualitative research strategy types
- Interviews
- Focus groups
- Case studies
Measurement levels
Metric (counted)
Categorical (Color, gender)
When is Pearson’s correlation coefficient used
used when both, dependent variable and independent variable are metric.
When is chi-square test used
When both dependent variable and independent variable are categorical
When is t-test used
When dependent variable - metric, independent variable - categorical AND number of levels of independent variable = 2.
When is one-way ANOVA used
When dependent variable - metric, independent variable - categorical AND number of levels of independent variable = 3+
[Is there significant difference in salary between 1. HR 2. Marketing 3.Finance managers]
When is logit analysis used
With multiple independent variables and categorical dependent variable
What to use in case of Multiple independent variables and metric dependent variable
Anova or linear regression
Pearson’s correlation coefficient what it does
measures the strength of the linear relationship between two metric (interval or ratio) variables. -1.0 to 1.0
Chi-square tests what it does
tests whether there is a relationship between two categorical variables (nominal or ordinal)
Independent vs Paired samples of t-Test
Independent - is there a difference between two tested groups? Samples are unrelated to each other
Paired - is there a difference between two points in time? Samples are the same
Covariate meaning
third variable that could be confounding your results
