Consumer Behavior Research Methods
Lernen
« zurück zur Kartenübersicht « Liste
96/157 [96]

Definitions

Probability vs. Non-probability Sampling (12 Points)

  • For exploratory research, non-probability sampling could be favorable – Why? Because findings are tend to be treated as preliminary.

  • For conclusive research, probability sampling is favorable.

    • Why? Because statistical projection of the result to the population is necessary

    • Only random (probability) samples let you generalize to a larger population!

  • Be aware, non-probability sampling suffers from selection bias

  • Properly carried out probability sampling eliminates the selection bias

  • But, probability sampling is costly and time-consuming

  • A common problem with sampling is non-response. How it can be reduced?  Use prior notification, motivating the respondents, incentive, and good questionnaire design, administration and follow up

  • Another very important topic: If we take a sample and e.g. compute the mean for this sample, then this measurement will be prone to errors

  •  There are two types of errors when you use samples:

    1. Random sampling error occurs because our sample is an imperfect representation of the population we are interested in
    -->After all, we have chosen a sample that is only one of a very large number of samples which may have been chosen, each giving rise to different sample estimates

    2. Non-sampling error occurs due to sources other than sampling (e.g. a bad questionnaire, respondents do not respond, wrong target population), and they may be random or nonrandom

  • If we indeed have used a random sampling technique, then there are statistical ways to estimate the random sampling error!

  • The most common measurement of sampling error is to compute standard errors of your estimates and a confidence interval

Diskussion