Sampling Methods: The Two Main Sampling Methods Are Probability

Sampling Methods: The two main sampling methods are probability and non-probability based. Probability samples are determined by sampling the population at random – every person in the population has a known chance of being included in the sample. Non-probability samples are convenience based, meaning more accessible population members are more likely to be sampled.

Benefits of probability samples include higher quality, more precise estimates (can’t estimate precision at all with non-probability samples). Can be more confident in interpreting and using resulting data. On the downside, it’s difficult to do (requires defining “population” and figure out way of reaching out to them such that we can determine the probability they are included in the sample).

Benefits of conveniences samples are just that – they are convenient to get. On the downside, there is a good deal of uncertainty about whether results will generalize beyond that sample, and no statistical tests available to test generalizability.

- Probability Sample 1. Simple random sample: all population members have 1/n chance of being sampled. 2. Stratified sampling: divide frame into “strata” according to a characteristic. Pro: 1. Can ensure representation of smaller sub-groups 2. Cost efficient increases precision

- Non- Probability Sample 1. Convenience 2. Quota

-SE = SD of sample mean/ square root of the number of observations in the sample

-Sampling Error: Error caused by observing a sample instead of the whole population. Estimates are always subject to natural sample to sample variation (random error).

-Sample Biases: Expected value differs from true value because of unwitting under sampling or oversampling of certain types of sampling elements.

-BMR: Imagine the end of the process (writing the report or presenting findings in a meeting). What are the various decision alternatives you might recommend? Identify analysis that can be used to support choice among these alternatives. Determine the information need to conduct the analysis. Get the data needed for the analysis. Analyze data & make recommendation.

-Cons of Focus Group: 1. small sample 2. Non-representative sample 3. no model for turning “liked the concept” into actual trial and use 4. social pressure demands that can influence participants in focus groups (perhaps not an issue here) 5. Participants influence each other 6. Psychological biases

-True experiment: Random assignment and the presence of a control group are the central features of a true experiment. Random assignment refers to the fact that each experimental subject has an equally likely chance of being assigned to each condition. Control group is the group who is like the experimental group in all ways except they did not receive the key manipulation.

Random assignment and including a control condition reduce threats to internal validity, including: 1. History: should affect both control and experimental groups equally 2. Selection bias: random assignment means equal chance of being in each condition, so shouldn’t see this 3. Maturation: this too should not differ between control and experimental groups 4. Testing itself can affect responses

Descriptive Research: Collection of quantitative data

Market Intelligence: 3 key skills 1. Defining the problem and determining what information is needed 2. Getting data and judging its quality 3. Analysis frameworks and using data to make decisions

-Secondary: Exists already, was collected for some other purpose- advantage: Cheap/quick/sufficient/baseline for comparison; disadvantage: too many data/number conflicts/categories may not fit

...