research methods advanced- week 3= survey design

survey design

survey designs

questionnaires, observations and interviews

survey research

involves asking questions and or observing individuals directly, conducted on sample of target population, aims to describe individuals as well as to offer predictions about them based on socio-demographic characteristics, can be used to study associations between phenomena but not causal

surveys and variables

no manipulation of variables, and focuses on relationships across variables. dependent on samplimg can then be generalised any relationships found to wider populaiton

stages of survey design

work ut aims, select appropriate question styles, design questions, pilot questionnaire, revise questionnaire, administer the questionnaire, analyse and report

standardised measures

rigorous design process, start with hundreads of candidate questions, each item is evaluated independently, undertake rigorous statistical analyses, validity and reliability repeatedly tested, tests have descriptive statisitcs to act as reference pount for new data

types of validity

predictive- able to predict validity in future

concurrent- is score related to another criterion of this construct tested at same time

convergent- is score related to other measures of the same construct

discriminant-is score different from measures that theoretically measure something else

types of reliability

test-retest =test across different periods and get same result

inter-rater= will results i get be the same from a different experimenter

inter-method= will i get the same results using a different method

internal consistency (cronbach’s alpha)= did you respond to similar questions in similar ways

item response thoery

provides an estimation of the discrimination parametre how well an item functions as a measure of a latent construct. allows the assessment of difficulty/ severity thresholds in the latent construct continuum

measurement invariance

captures the degree to which your measure is testing the same thing across conditions, supports results as increases certaintiy of finding true differences accross conditions

what to consider when putting survey together

demographic questions, standardised measures and unique questions whilst considering respondent motivation, length, density presentation and question order