Research Methods AS L5 - 8 (sampling, designs, effects and variables) Flashcards
3 steps of conducting research:
1) Deciding an aim
2) Decide IV and DV + operationalise
3) Decide a hypothesis + operationalise
Aim:
Precise statement about purpose of study + what it intends to find out
Hypothesis:
Precise, testable statement about expected outcome of investigation
Null hypothesis:
IV will have no effect on DV
Alternative/Experimental Hypothesis:
IV will have an effect on DV
Non-directional hypothesis (two-tailed):
Direction of predicted differences between conditions not shown eg. eating chocolate will significantly affect a person’s mood
Directional hypothesis (one-tailed):
Direction of the predicted difference between condition is shown eg. eating chocolate will increase a person’s mood
Target population:
The group who researchers are studying and want to generalise their results to
Why are sampling techniques used?
Obtain a sample of target population to avoid studying entire populations, as it is too long and expensive
What should a sample aim to be like?
- Representative of population
- Same characteristic as population
Random sampling:
- Every member of target population has an equal chance of being selected
- Use a bias-free method to select a sample
Strengths and weaknesses of random sampling:
+ If target population is large and sample is large, sample is likely to be representative
+ No researcher bias as sample has been chosen by chance
- Difficult to get full details of target population
- Not all members of target pop will be willing to participate/be available, making sample unreliable and reducing the sample size
Systematic sampling:
Participants are selected by taking every nth person from a list of the total pop
Strengths and weaknesses of systematic sampling:
+ Much simpler than random sampling as it only requires a list of target pop
- Process of selection can interact with a hidden periodic trait within pop due to coincidence
- Not all members of target pop will be willing to participate/be available, making sample unreliable and reducing the sample size
Stratified sampling:
- Classifying pop into categories and randomly choosing a sample that consists of participants from each category in the same proportion as they are in the pop
- Divided into strata in terms of characteristics
Strengths and weaknesses of stratified sampling:
+ All grps within pop are included, therefore it is representative
- Time-consuming
- Hard if you do not have details of everyone in stratified sample
Opportunity sampling (Convenience sampling):
- Selecting participants who are readily available and willing to participate
- Possibly easy access to a group of people they know well
Strengths and weaknesses of opportunity sampling:
+ Easiest and most practical method for large samples in comparison to stratified sampling
- High chance sample will not be representative
- Sometimes people feel obliged to participate in research (especially if researcher is known to them), which is possibly unethical
Volunteer sampling:
- Volunteering to participate in a study
- Researcher usually advertises for people to participate thru leaflets, posters, radio or TV broadcasts
Strengths and weaknesses of volunteer sampling:
+ Time and effort is saved as the researcher only needs to wait for pps
- Certain type of person tends to volunteer eg. Someone who is interested about the topic therefore it is possible the sample is unrepresentative
In a pilot study, what is it important to check when using self-report measures?
- Understanding of questions
- Closed questions still offered suitable options
- Whether open questions are needed to elicit unpredictable responses
- Appropriateness of reporting methods
In a pilot study, what is it important to check when using observations?
- Operationalised definitions of behavioural categories are agreeed by observers
- Inter-observer reliability (do researchers need to practise observing behaviours?)
- That behavioural categories include all important behaviours
- Behavioural categories don’t overlap
- Whether pps are affected by observers (should it be covert?)
Which two effects should be checked in a pilot study?
- Floor effect = None of the pps can complete the task as it is too hard
- Ceiling effect = Task is so easy pps can virtually achieve full marks
Experimental designs:
- How pps are organised w/in the experiment
3 types of experimental designs:
1) Independent grps
2) Repeated measures
3) Matched pairs
2 different conditions psychologist may want to compare (+ meanings):
1) Experimental condition –> grp of ppl are exposed to IV
2) Control condition –> grp received no treatment + are used as baseline level to compare results
3 types of controls:
- Randomisation
- Standardisation
- Random allocation
Randomisation:
- Use of chance to control effects of bias when deciding order of conditions
- Put in random order by computer/manually
- Especially important using repeated measures design and same pps taking part in diff conditions
Standardisation:
Using same formalised procedures and instructions for all pps in study, which improves reliability
Random allocation:
Attempt to control for pp variables (individual differences) in an independent groups design
Pilot study:
- Preliminary small scale investigation of procedures to be used in main study
- Involves selecting a few people and trying our study on them
Strengths and weaknesses of pilot study:
+ Time and money can be saved by identifying flaws in procedure early on
+ Can help spot ambiguities/confusion/practical issues/suitability
- Time-consuming
- May not be beneficial overall as not all EV can be eliminated regardless
Independent groups design:
- Different pps used in each of the conditions
- Each grp of pps are independent from one another
- Pps randomly allocated to balance out individual differences
Strengths and weaknesses of independent groups design:
+ Order effects will not occur as there are different pps in each condition (pps may get tired/bored)
+ Chance of demand characteristics reduced as pps take part in only one condition each
+ Pps not lost between conditions as pps take part in only one condition each compared to repeated measures design where they take part in 2
- More pps needed for this design
- Different results gained could be due to individual differences rather than manipulation of IV
Repeated measures design:
Each pp is tested in all conditions of experiment eg. tested today to see effect of chocolate on mood, tested next week as well
Strengths and weaknesses of repeated measures design:
+ No individual differences between conditions as same pps are used
+ Half pps needed compared to an independent groups design
- Order effect may affect results
- Demand characteristics more likely as pps are involved in entire study + both conditions
Matched pairs design:
- Diff pps used in all of conditions
- Pps in 2 grps measured on characteristics important for study eg. condition one: male aged 45, condition two: male aged 45
- Identical twins used often
Strengths and weaknesses of matched pairs design:
+ Less risk of order effect as pps only take part in one condition
+ Individual differences are less likely to occur as grps have been closely matched on specific characteristics
- Matching process incredibly difficult and time-consuming as twice as many pps required
- Even 2 closely matched individuals have diff levels of motivation and fatigue
Order effect:
Sequence in which pps take part in conditions influence their performance/behaviour
One way to avoid order effect:
Counterbalancing:
- Half pps do Condition A first then Condition B
- Other half do Condition B first then Condition A
Ecological validity:
Ability to generalise findings of piece of research to other settings
Validity:
Study is measuring what intends to measure when referring to aim of study
4 types of validity:
- Ecological
- Participant
- Temporal
- Internal
Participant validity:
Results from pps of study can be generalised to target pop
Temporal validity:
Results of study can be generalised to people in today’s contemporary society
Internal validity:
Results of study is a direct result of manipulation of IV upon DV and has been unaffected by EV
Reliability:
Consistency of a research study
What method is used to assess external reliability, what is this and what outcome shows external reliability?
- Test-retest method
- Research study is conducted once and then conducted again
- If results are same, the study has external reliability
Internal reliability:
Whether test and results gained are consistent within itself
External reliability:
Whether test and results gained are consistent over time
What method is used to assess internal reliability, what is this and what outcome shows internal reliability?
- Split-half technique
- Questionnaire split in half and pps complete both halves
- If scores are similar on both halves, it has internal reliability
What needs to be well controlled to ensure a study has internal reliability?
Extraneous variables
3 categories of extraneous variables:
- Participant variables
- Situational variables
- Experimenter variables
Participant variables:
Characteristics of pps that may affect DV eg. age, personality
Which designs can help avoid participant variables and how must they be altered to avoid other issues?
- Matched pairs + Repeated measures design
- Repeated measures w/ counterbalancing to avoid order effects + random allocation to avoid biased grps
Situational variables:
Factors in environment where experiment is conducted that could affect DV eg. temperature, time of day
How can the issue of situational and experimenter variables be resolved?
Standardisation
Experimenter variables:
Factors to do with experimenter which can affect DV eg. personality, appearance
Investigator effects:
- Investigator inadvertently influences research results due to certain physical characteristics eg. age, gender
- If investigator knows hypothesis, they may be biased when interpreting results
Types of investigator effects:
- Observer bias
- Interviewer effects
How can the issue of investigator effects be resolved?
Double blind technique = Neither pps nor investigator knows aim of study
Demand characteristics:
When features of research studies enable pps to guess aim of research and hence, they act unnaturally out of nervousness (try to annoy by giving incorrect answers) /social desirability (try to please by giving expected answers)
How can the issue of demand characteristics be resolved and what experimental design is this hard to accomplish in?
- Single blind technique = Only pps do not know the aim of study
- More difficult to do in repeated measures design
