Survey – theoretical outline Jiří Navrátil jiri.navratil@fss.muni.cz GLCb1008 Introduction to Methodology of Social Sciences Lecture outline •Context of survey design •Sampling •Data collection •Asking questions Explanation vs. understanding •Idiographic (deterministic): understanding, revealing motives, reasons, justifications… •Idiographic method – problem with zero •Nomothetic methods (probabilistic): explanation, identifying causal relationships Two logics of research process (de Vaus 1996: 21) Main types of research strategies •Experiment (small N) •Case study (small N) •Comparative design (small N) •Longitudinal design (small/large N) •Cross-sectional design (large N) • Social survey •Probabilistic/nomothetic logic •Deductive logic •Cross-sectional design (large N) Cross-sectional studies •More than 1 case (in fact large N) in one time frame •Collection of quantifiable data on the characteristics („variables“) of these cases •Logic of correlation between and among varibles •Typically – social survey/questionnaire (but other data collection methods available too) •Variables are not manipulated Time frame for cross-sectional studies Sampling •Population – all units (individuals, nations, organizations, cities …) from which we select sample and to which we refer with the results •Can we handle whole population? ( census …) •Sample – part of units selected for analysis •Probability sampling (logic of chance) vs. Non-probability (purpose) • Random sample •Standard type of sampling in large N •Representativity – known and unknown characteristics of population •Option of statistical inference from the sample to the population •Types of random sample: simple, systematic, stratified, multi-stage Simple random sample Systematic random sample Multi-stage cluster sample Non-probability sampling •convenience sampling – who do you succeed to get in the sample (street distribution, relatives, friends, social media) •purposive sampling – focusing at target group within population (study of student environmental activists – looking for the most visible ones at the faculty) •snowball sampling – members of population are hard to reach (subcultural or social movement studies) Non-probability sampling •Quota sampling – when we know the characteristics of target population (typically: sex, education, ethnicity, age) •Example: Czech adult population (attitudes towards the conflict in Ukraine) bydliště kvóta Hlavní město Praha 122 Středočeský kraj 129 Jihočeský kraj 61 Plzeňský kraj 55 Karlovarský kraj 27 Ústecký kraj 76 Liberecký kraj 42 Královéhradecký kraj 52 Pardubický kraj 49 Kraj Vysočina 48 Jihomoravský kraj 112 Olomoucký kraj 59 Zlínský kraj 55 Moravskoslezský kraj 113 pohlaví kvóta muž 490 žena 510 věk kvóta 18 - 24 80 25 - 34 150 35 - 44 190 45 - 54 180 55 - 64 190 65 + 210 vzdělání kvóta ZŠ 136 SŠ bez maturity 326 SŠ s maturitou 351 VŠ 187 Example of quota sampling Sample size •Absolute, not relative size of sample is important •Increasing sample size – decreasing possibility of sampling error (random sampling) • •https://www.surveymonkey.com/mp/sample-size-calculator/ • Types of data collection Data collection Online tools https://www.kobotoolbox.org/ https://www.surveymonkey.com/ Data matrix (column = variable; row = case) Modes of survey administration compared Types of questions •Open-ended vs. Closed questions (time, validity, reliability, comparability, clarification, easiness, spontaneity, exhaustiveness) •Personal factual, others factual, informant factual, attitudes, beliefs, normative values, knowledge How to design the questions •Follow research question! •Be specific! (What is your opinion about contemporary party politics?) •Avoid ambiguous questions! (How often do you eat fast food?) •Avoid long questions! (On a scale of 1-10, how likely are you to recommend political party you voted for in last elections to a friend or colleague who works in a similar industry as you do?) •Avoid double-barrelled questions! (Do you think that students should have more classes about history and culture?) •Avoid leading questions! (Do you think that our country is a true democracy?) •Think about filters and conditions! (What movie did you see in a cinema last month?) •Avoid using jargon and technical terms! (What is your opinion of the efficacy of the Twitter´s new search algorithm?) •Think about the respondent´s capacities, memory and knowledge! (When did you drink alcohol for the first time?) •Always do the piloting! •Use existing questions! References