Week 14 - Designing Online Questionaires Flashcards
(12 cards)
Recap of cross-sectional research
- Can tell us about the strength + direction of a relationship between variables
- Used to collect quantitative data
- Whether we use an experimental design or cross-sectional design depends on the research question
- Measure variable/s once for each ppt
- Provide correlational data
- Can’t inform us whether one variable causes the other
-> can’t enable us to establish cause + effect - would need to carry out an experiment
Cross-sectional design vs experimental design
Cross-sectional:
- Assesses direction + strength of the relationship between variables
- Measure variable/s once for each ppt
- Can take place in more naturalistic settings
Experimental:
- Used to establish cause + effect
- Typically, take more than one measurement
- Need to be carried out in highly controlled conditions
-> to control for possible effects other variables may have on our results
What are three ways the internet can be used as a research tool?
- Internet allows large volumes of data to be collected
- Participation can be reactive -> ppts who are interested in the study would read the info and consent to take part - they are reacting to a request and giving consent to participate in the study
- Participation can also be non-reactive -> ppts didn’t know they were participating in the research - no request from researcher
-> analysing existing posts on discussion boards and forums may provide insight into different behaviours - Psychologists are also interested in how internet use affects people’s behaviour
What are the benefits of online data collection?
- Enables access to wider or specialist populations
-> removes geographical restrictions on our research
-> can access different populations, not just student populations, + individuals who might be difficult to contact face to face (e.g. patient populations) - Data collection is automated -> researcher doesn’t need to be physically present - ppt accesses study via electronic link
-> time saving + reduces the number of researchers that need to be involved
-> every ppt is receiving the exact same task instructions - Stores data automatically
-> researcher doesn’t need to keep a log
-> enables us to collect data from a large volume of people
What are the challenges of online data collection?
- No control who participates
-> individuals who might not meet the inclusion criteria may take part
-> ppts may try to participate in study more than once - problematic as might lead to biases in your results - High drop out rates
-> ppt may be more likely to get distracted (may have multiple tabs open on computer) - Difficult to control for environmental + procedural variables
-> researcher has little control over environment the study is being conducted in
-> procedural variables e.g. screen size (using a phone vs a computer)
What are the ethical considerations when doing data collection online?
- Unlike face-to-face research, researcher is unable to monitor, support + end the study if required
-> so additional precautions should be taken to protect ppts - Informed consent
-> In online studies, we want this to be the first screen
-> want to be sure that ppts have read the statements related to giving informed consent - in online research, ppts may rush past this page to proceed to the study
-> can get around this by using checkboxes associated with each statement - Withdraw
-> if ppt want to terminate their participation during the study they can close their internet browser - can withdraw data after study by emailing one of members of research team with a unique identifier
- Minimising harm
-> need to think if the topic wee are investigating is it appropriate to collect data online - some sensitive topics can’t collect data online as researcher not physically present - can’t check if in distress
-> need to pre-warn the pptts about the nature of the research before they agree to take part
-> need to also signpost ppts to appropriate support agencies - in ppt info sheet + debrief
Where can you find psychological tests?
PsychTESTS
-> a database of psychological measures, scales, surveys and other research instruments
Recap of reliability + validity
Reliability -> relates to the consistency of the test
Validity -> relates to whether the test measures what its supposed to be measuring
Describe external reliability
- Relates to the consistency of the test over time
Test-retest reliability:
- Assess the correlation between scores taken at two points in time from the same sample
-> give test to group of ppts then wait a bit of time and give the same ppts the same test
-> would conduct a correlation analysis + if a strong correlation was found between the two scores at the two different time points, suggests the test has good external reliability
Describe internal reliability
- Refers to consistency within the test itself
- If test has high internal reliability, this suggests ppts would answer in a similar manner across the different items
- If low in internal reliability, they may show a lack of correspondence across the different items
-> this lack of correspondence may indicate that these items aren’t measuring all the same construct
What is one way to check if a test has internal reliability?
Split-half method:
- Administering the test to a group of ppts and then halving the items so you have two scores and then looking at the correlation between those two scores
-> if two scores strongly correlate, this would suggest the test had good internal reliability
Cronbach alpha:
- Reflects all possible split half combinations
- Alpha values range from 0 to 1
- Alpha values ranging over 0.70 considered to be acceptable
What are 4 types of validity we need to consider when selecting a test?
Face validity:
- Do the items on the test appear to be measuring what they’re supposed to be measuring?
-> can be judged by reading through the items + forming a judgement
Predictive validity:
- Can test scores be used to predict future events?
Concurrent validity:
- Do scores correlate with other measures taken at the same time?
-> e.g. using a new personality test to measure extraversion - then checking whether scores on this newly developed test correlate with existing measures for extraversion
Construct validity:
- Does the test measure the construct it was designed to measure?
-> can test this by checking whether that test score correlates with other measures that have been designed to use the same construct
