Survey & Questionnaire Flashcards
Survey measures a particular…
Construct
Research methods are ways data is collected to
Answer q’s
Test hypotheses
Choice of research methodology is shaped by
Epistemological framework (how we can know something)
The choice of data collection is independent of
Epistemological framework
Observational research
Watching people do stuff
Types of observational research
Fully structured Naturalistic Participant Event sampling Time sampling
ainsworth strange situation is an example of
Fully structured observational research
What does strange situation measure
Attachment
Participant observation
You might be one of the participant in the situation. Different levels of immersion
Event sampling observation
Interested in observing ‘events’ we set up
Time sampling observation
Repeated observations over longer period
Fully structured observation pros
Systematic data collection
Confounding variables controlled
Constant setting
Replication
Fully structured observations cons
Lacks ecological validity (too artificial)
Behaviour spontaneous
Ethical issues
Naturalistic observation pros
Good ecological validity
Rich data
Behaviour less subject to demand characteristics (Doing things you think the research wants to see) of setting
Naturalistic observation cons
Difficult for observer to be unobtrusive
No control over confounding variables
Difficult to replicate
Ethical issues
Archival research methods
Use of already existing documents and records
Types of archival research methods
Census data Police fines Clinical records Organisational record Newspaper magazines Letters papers
Archival research pros
Data already collected
Cost effective
No ethical review required
Allows investigation of q’s that can be investigated in no other way
Archival research cons
Data may not be in format to answer research q
May not include info about variable of interest
May contain different measures used at different times
Case studies are
Detailed description of single unit of analysis (person or organisation)
Used to formulate broader research hypothesis/questions
Case study pros
Rich data Context for behaviour identified Promotes methodological trigangulation (integration of info from diverse sources) Can include temporal element (over time) Facilitate theory generation
Case study cons
Does not allow generalisation
Time consuming
Open to researcher “bias”
Ethical issues
Interviews are
A conversation aimed towards understanding the experience of another
Interviews can be:
Structured
Semi-structured
Unstructured
Interviews are used in ……….. Research
Qualitative
The more structure you have in the instrument the more ……….. It is
Reliable
Interview structured pros
Easy to administer
Easily replicated
Lower influence of interpersonal factors
Simple data analysis
Interview structured cons
Participant constrained
Reduced richness of data
Info distorted through poor q wording
Interview unstructured pros
Flexible
Rich data
Valid data
Relaxes participant
Interview unstructured cons
Unsystematic
Difficult to analyse data
Strongly influenced by interpersonal variables
Focus groups are
Discussion-based interview that generates data from group interactions. You are the moderator.
Focus groups aim to
Capture understandings, perspectives, stories, discourses and experiences not otherwise meaningfully expressed by numbers
Focus group pros
Dynamic Involves participants in research Easy to assemble Economical Rich data Good for ppl with lower literacy levels
Focus groups cons
Required moderator good facilitation skills
Data difficult to analyse
Group processes affect data obtained
Artificial setting
Experiments allow the researcher to:
Control or manipulate the environment to see effects on human behaviour
Experiments include:
Variables of interest
Control conditions
Random assignment
Experiments aim to
Establish that a change in one variable CAUSES a change in another
IV is the variable that the
Experimenter manipulates
DV is the variable the experimenter predicts will be affected by the
IV
Experiment pros
Controls for external influences and affect of confounding variables
Demonstrates causal relationships btwn varables
Accurate measurement of variables
Easy replication
Experiment cons
Artificial, low ecological validity
Results may not generalise outside of lab
Variables may be narrowly defined and lack construct validity
Quasi-experiments explore
Relationships btwn variables
In Quasi-experiments random assirng to conditions is
Not possible
Quasi-experiments must have…
Pre-existing groups ie male and female groups
Quasi-experiment pros
Good face validity
Good ecological validity
Economical
Quasi-experiments cons
No casual relationships established btwn variables
Experimental error more likely
Face validity is
Appear to measure what we’re trying to measure
Pre or no experimental methods are:
Non causal (correlation)
Causal relationship is:
Hot day = better ice cream sales
Hot = more robberies
BUT hot day, ice cream sales and robberies not causality related…
Questionnaires
Method of data collection
Questionnaires allow
Collection of data from large numbers of ppl
Questionnaires measure something that is
Not directly observable ie a theoretical construct
Questionnaires provide info on
Incidence of some behaviour of interest
Opinions
Beliefs
Attitudes of large no of ppl
Survey is
Study design where questionnaire is primary focus
Items are
The questions/statements in questionnaire
Scale is
Set of items that measures something psychological or not directly measurable (construct)
Inventory is
Publication of several scales
1st general principle for making good questionnaires
Checking questionnaire does not already exist
2nd questionnaire rule
Lenght
3rd questionnaire rule
Readability
4th questionnaire rule
Writing questions Open/closed Language Response sets Measurement techniques (ie likert scales, semantic differential)
5th questionnaire rule
Assessing psychometrics
Ie reliability, validity
The longer the questionnaire… The less
Reliable (too broad) not focused.
Jangles are
Different labels for what is the same as something else
Jingles
Giving constructs the same name
Readability got questionnaires grade level is
Btwn 7-8 grades
4 things to consider when writing questions…
Open/closed questions
Language
Response sets
Measurement techniques (likert scales, semantic differential)
Open questions are qualitative and responses need to be
Analysed in depth
Closed responses are
Quantative and can be easily analysed
5 language rules for questionnaires
Ask 1 q at a time Avoid ambiguity Avoid double negatives Keep neutral Avoid technical terms and jargon
Response sets are
How you went about answering the items. Trying to make yourself look as good as possible. Can pollute results.
Best way to get around social desirability bias in questionnaires is to
Incorporate social desirability scale into questionnaire
What is response acquiescence
Tendency to agree rather than disagree with statements
Combat response acquiescence
Change wording so not are all positive and negative
Make q’s unpredictable forcing thought
Likert scale is
- Or 7 point scale. Participants indicate how much they agree or disagree with a statement
Likert scales rules
Agreement
Symmetrical
Labels are roughly equivalent moving upwards
Semantic differential scale
Takes a more indirect approach to measuring attitude toward behaviour or event. Participants indicate thoughts/feelings
Psychometrics is
Area of psychology concerned with the quality of scales and items designed to measure psychological constructs
2 psychometric criteria
Reliability!
Validity!
