3. Data Collection

How will data be collected

  1. Observational research - gather info about smth that little is know about by looking, listening, judging - study people/phen. in natural setting (answers how/why questions)

  2. Existing info/ big data

  3. Surveys - study behav/opinions - can be on paper/ online/ phone

    • (+) multiple Qs can be asked about same topic —> different aspect of same theoretical concept can be measure

    • Likert scales

Scale score = combination/mean of all different answers (e.g. answers to dif question on IQ test) —> is a construct (IQ is a constructed variable bc it is the result of lots of dif Qs)


Sampling options

A. Nonrandom samples

  • chance of bias/ distortion

  • generalising is NOT possible

  • low ext. val.

  • still often used in qualitative studies


  1. Convenience sampling - researchers use participants who are easy to reach (e.g. UU psychologists recruiting UU students for small experiments)

  2. Purposive sampling - researchers look for specific individuals who can provide important info (e.g. study about coping mechanisms in refugees will look for refugees who had it tough)

  3. Quota sampling - convenience sample with requirement for no. within the groups

  4. Snowball sampling - targeted sampling where participants are asked to recommend other ppl for the study

B. Random sample

  • sample is representative of the entire pop.

  • generalising to pop. is possible

  • high ext. val.


  1. Simple random sample - requires list of all members of pop. and uses computer to select random participants (all participants have same chance of being selected)


  2. Stratified sample - combo of multiple random samples drawn from subgroup of pop. (to compentate for groups varying in size or having systematically dif experiences)

  1. Cluster sample - divide pop. into clusters and randomly select some of them

    • Multistage sample - from prev random cluster select random individuals


  2. Systematic sample - assuming pop. is ordered in some way, you pick every n-th individual


Potential errors:

Coverage error - not everyone is included in the list we draw from —> biased results

Non-response errors - some people refuse to participate or answer Qs —> biased results