STUDY UNIT 3 Data collection 1 Sampling 2

STUDY UNIT 3: Data collection: (1) Sampling; (2) Measuring, (3) Questioning and (4) Observing Learning Unit Objectives: ► Be able to draw a sample apply measurements ► formulate items for a questionnaire and an interview ► conduct observations ►

Sampling Target population Accessible population Sample Source: google images After describing the population parameters, a representative sample needs to be drawn, where after the results can accurately be applied to the whole target population.

Calculating ACCURACY of sample The accuracy of any sample must be no less than 95%, otherwise the reliability and validity of the results of the research might be called into question. A formula to calculate standard error (s) = the square root of population parameters (p) x (1 – p) divided by sample (n) (Du Plooy 2009: 110). After calculating the standard error, the researcher can note the degree of confidence of a representative sample.

Types of sampling (1) Probability sampling – units of analysis have an equal chance to be selected. Units of analysis are selected at random (1. 1) Simple random sampling - e. g. selecting random numbers from a table / group (1. 2) Stratified random sampling – selecting other sub-groups, where the population parameters homogenous (2) Non-probability sampling – units of analysis do not have an equal chance of being selected. (2. 1) Convenience sample – units of analysis are conveniently available (2. 2) Purposive sample – the researcher selects his / her own units of analysis (2. 3) Volunteer sample – people volunteering to be part of the study (2. 4) Snowball sample – e. g. placing an advertisement to invite people to partake in the study (3) Quasi-probability sampling – sampling somewhere between probabilityand non-probability sampling. Units of analysis are selected at random. (3. 1) Systematic random sampling – e. g. the units of analysis are not selected from a random table / group (3. 2) Cluster random sampling – although there are sub-groups, they are grouped in clusters (e. g. geographic / ethnic) (3. 3) Multistage random sampling – different types of sampling used

Source: Google images

Measurement “Assigning numerals to variables being studied” (Du. Plooy 2009: 126) – e. g. in a questionnaire 20 units of analysis replied “YES” to the question: “Do you vote for COPE? ” Levels of measurement 1. Nominal level: level using numbers to identify categories – e. g. man = 1; woman = 2 2. Ordinal level: level numbers used to rank variables – e. g. best speech = 1; average speech = 2; worst speech = 3 3. Interval level: level rank-ordered, where the difference between the values are equal – e. g. 5 meters = 1; 10 meters = 2; 15 meters = 3 4. Ratio level: level the same as an interval level, however it includes the 0 (absolute zero) numeral to indicate when a variable is absent

Reliability and Validity Reliability – “The likelihood of being accurate and providing the correct result” –Oxford Dictionary. Reliability consist of three (3) components, nl internal consistency, stability and equivalency. Validity – “Bringing about the results intended” – Oxford Dictionary. The different types of validity include: face/content validity, expert-jury validity, criterion-based validity, and construct validity.

Measurement scales – i. e. ways to measure information 1) Likert scales Respondents can either ‘strongly agree’ / ‘agree’, be ‘neutral’ or ‘disagree’ ‘strongly disagree’ with a statement. 2) Semantic differential scales Investigating the relationship between words and their meanings – e. g. Mr. Marchant is: Weird : …: …: . x. : …: Normal Funny : …. : . x. : …: …: Boring

Problems in wording questions or statements Have you and a boyfriend / girlfriend ever had an argument about whosaid-what? The same principle applies to asking research questions or making research statements: you have to make sure the WORDS you use are precise. Ask yourself: 1. Is this a double-barrelled question (statement) 2. Is the assumption clear 3. Is the language I’m using loaded (double meaning) In a self-administered (a researcher provide a questionnaire which needs to be answered by respondents), there are the following types of questions: 1. Direct- vs. Indirect questions 2. Specific- vs. General questions 3. Open-ended vs. Closed-ended Questions (including, pairedcomparison-, filter-, contingency-, ranking-, inventory-, matrixand multiple choice questions)

How the wrong wording can lead to an invalid result Source: Google/images

Collecting data through OBSERVATION We’ve spoken about collecting data through questionnaires, but often behaviour, opinion and other phenomena can provide data which is collected through field observation. When conducting quantitative research, the researcher will typically conduct his / her observation systematically. When conducting qualitative research, the researcher will typically conduct his / her observation enthnographically.