Flashcards in INRS 7311 Learning Unit 4 Sampling and Data collection Theme 2 Quantitative Data Collection techniques Deck (12):
Define survey research 5 marks
*Defined as a data collection tool consisting of a series of questions to gather information about a relatively large group of people .
*Popular research tool often used to get demographic and psychographic information . Also used to to get quantitative description of trends or opinions to be able to generalise to the greater population
Distinguish between different types of surveys . 8 marks
*Cross-sectional = only collect data from participants once .
*Before-and-after = data is collected at to preset points in time . Most useful to measure a change in a particular event/situation .
*Longitudinal survey = used to determine log term effects of an intervention . (3 main types : trend which looks at a change in the population over time according to a phenomenon , cohort where a different sample is used but drawn from at different times and panel where the same sample is used at different times)
Identify the different types of questionnaires and explain the application of each one 16 marks
*Mail = send questionnaires either via mail or email . +ve's=self-administered , cheap ,allows for sensitive topics . -ve's = can't clear up misunderstandings , low response rate , long time to collect .
*Telephone = similar to mail but over the phone . +ve's: cheap,quick,high response rate , easy to clarify misunderstandings . -ve's: can't observe participants , getting access to sample , can't read NV behaviour .
*Interview = Structured face to face interview in certain setting . Recorded answers to preset Q's . +ve's : can clarify queries , high response rate, respondents don't have to be literate , able to build rapport . -ve's : takes long time , expensive , can't be used on a large sample , impossible to be anonymous .
*Group administered surveys = Questionnaire filed in by a group during a single session . +ve's : high response rate , quick , can clarify questions . -ve's : expensive .
Discuss the nature and characteristics of questionnaires 6 marks
*They often use close ended questions with checklists and rating scales .
Keep the following in mind .
/Always include an intro
/Include a rough time of completion
/Clear instructions and language
/Logical layout and sequence
/Not too many Q's
/Pro tip = demographic Q's at the end
/Includes thanks for including the questionnaire .
/Pre-test if possible
Identify different types of questions that can be used in questionnaires 24 possible marks
/Direct = demand open response
/Indirect = More polite and less confidential
/General = Applicable to a greater situation
/Specific= Like direct but focus on one particular aspect
/Close-ended = Have a fixed number of options
/Open-ended = Allows respondents to answer in their own words and express their feelings and ideas
/Paired-comparison= Choice between two options
/Contingency = Answer the question that follows only if the answer to this Q is one specific thing (eg only answer Q2 if you answer yes to Q1)
/Ranking = Given a set of options and rate them on a given scale (eg rate your favourite foods from 1-5 with 1 being least favourite )
/Inventory = Given a list of options and must choose what personally applies to them
/Matrix = Scaled options that gauge feelings of respondent on one specific area .
/Multiple choice = Respondent chooses most applicable from list of choices .
Identify and discuss possible problems associated with the wording of the questions 22 marks max
/Double barrell = Two Q's in one sentence
/Questionable assumptions or presumptive Q's = Make assumptions about respondent
/Loaded language = Q phrased in a way that indicates the researchers positions
/Leading Q's = Subtle pressure on respondent to answer one way or another .
/Negative Q's = Q's that can be misinterpreted
/Incomplete Q's = Respodents must answer on incomplete info provided .
/Vague = Q is vague in its wording
/Lengthy = Confuse people and distracts respondent from the focus of the Q
/Ambigious language = Too open to interpretation and/or have multiple meanings
/Abbreviations or acronyms = Assumes respondents know what they mean or stand for
/Complex Q = Laden with industry specific jargon that's hard for a lay person to understand .
Explain different measurement scales that can be used 4 marks
*Can be part of questionnaires but can be used on their own
*What scale is chosen is vital for how the data is analysed later on
Ratio scale is highest level of measurement , nomial scale is the lowest .
Identify the different types of surveys and explain the application of each kind . 18 marks maximum ,
*Nominal scale = Use of numbers to categorise to label variables . These numbers have no statistical relevance .These categories must be : exhaustive
(all elements of the population must fall into one category) , mutually exclusive(be able to fit in ONLY one category ) , no numerical significance (category number labels musn't mean/indicate anything) .
*Ordinal scale = Ranking of objects with regards to a common variable . Each item has more or less of the variable measure against than the other items . This measurement doesn't include the difference in the variable presence of each item . )
*Interval scale = Similar to ordinal but differences are equidistant from each other . Important to remember with this scale is that 0 is a possible value .
*Ratio scale = Like an interval place but absolute zero is a possible figure(meaning that variable is absent) . eg How much money have you spent at Restaurant X ?
*Likert scale = asks respondents to indicate a level of agreement or disagreement to statements regarding a attitude or object . Usually a 1-5 scale from 1 being lowest to 5 highest .
*Semantic differential scale = Also uses a scale of five or seven but includes polar opposites on the scale . (eg rating a temperature from hot to cold)
Distinguish between quasi-experimental and true experimental designs 9 marks
*In quasi experimental designs the participants are not randomly assigned to groups . Instead they are assigned base on variable that could be predetermined . 3 Types =Propensity score matching (matches two groups according to a variable that could be similar , like income bracket) . Regression discontinuity (comparison of instances above or below a certain threshold ) . Difference in difference (arguably the most common . The difference in growth between two groups is measured aka the difference in difference)
True experimental The notes list 4 types , namely :Pretest/Posttest , Solomon four group design , Factorial design , Natural Experiments
Explain the threats to internal validity in experimental design 24 marks maximum
*History = event takes place in betwen pre and post test eg getting robbed between a crime survey .
*Maturation = Indivudals change over time eg a survey on cooldrink subliminal ads is irrelevant if people are buying drinks because they feel thirsty and not because of the ads
*Testing = Pre test affects post test .eg repeating the same maths test twice either side of a maths workshop could just mean the students remember the test better . *Instrumentation = The test doesnt actually measure what its meant to measure .
*Statistical regression = choosing participants that are extremely far from the mean . Regression to the mean is far more likely , regardless of any innovation .
*Selection bias = Members of control group and experimental group must be compatible
*Selection maturation interaction = similar to selection bias , but in this case the incompatibility happens because time passes .
*Experimental mortality = Some respondents are lose inbetween the pre and post test period .
Explain quantitative content analsyis 4 marks
*An analysis of the content in mass media
*Denotative = literal , dictionary meaning
*Free from any form of interpretation