experimental research methods Flashcards
1
Q
What are lab experiments conducted under?
A
highly controlled conditions , where accurate measurements are possible, a standardised procedure is used
2
Q
What are some strengths of lab experiments
A
-A high degree of control over the variables minimises any extraneous variables - higher internal validity and establishes a cause-and-effect relationship
- more replicable as the variables are highly controlled
3
Q
what are some limitations for lab experiments
A
-low ecological validity - low mundane realism
- high risk of demand characteristics
4
Q
What are field experiments?
A
- field experiments are done in the everyday in the everyday environment of the participants
- the experimenter still manipulates the independent variable
-participants are often unaware they are participating
5
Q
what are the strengths of field experiments?
A
- high ecological validity - research is conducted in the real world
- low risk of demand characteristics - unaware they are taking part in a study
6
Q
what are the limitations are field experiments?
A
- less control over extraneous variables - reduces the internal validity
- potentially more time-consuming and expensive
- difficult to replicate precisely
7
Q
what are natural experiments?
A
research is conducted in a natural setting
the experimenter does not manipulate the independent variable directly, it varies naturally
8
Q
what are the strengths of natural experiments?
A
-high ecological validity
- low risk of demand characteristics
- can be used in situation in which it would be ethically unacceptable unacceptable to manipulate the independent variable
9
Q
What are natural experiments?
A
-research is conducted in a natural setting
- the experimenter does not manipulate the independent variable directly, it varies naturally
10
Q
what are the strengths of natural experiments?
A
- high ecological validity
- low risk of demand characteristics
- can be used in situations in which it would be ethically unacceptable to manipulate the independent variable
11
Q
what are the limitations of the natural experiments?
A
- there is no control over extraneous variables
- its difficult for another research to replicate the study in the exact same way
12
Q
What are quasi experiments?
A
the IV is naturally occuring but the DV may be measured in a laboratory setting
The IV is a difference between that existse.g gender or personality differences
13
Q
What are the strengths of quasi experiments?
A
allows for comparisons between different types of people
quasi experiments are often carried out under controlled conditions and therefore share the strengths of a lab experiment
14
Q
What are limitations of the quasi experiments?
A
- researcher cannot randomly allocate participants to conditions and therefore there may be confounding variables
-share limitations of a lab experiment if in controlled conditions
15
Q
What is the dependent variable?
A
the variable that is being measured.
16
Q
what is the independent variable?
A
the variable that is being manipulated.
17
Q
What is the operationalisation of variables?
A
is crucial i.e. putting the variables into a form that can be easily tested or measured by defining them as precisely as possible.
18
Q
What is the aim of an experiment?
A
is crucial i.e. putting the variables into a form that can be easily tested or measured by defining them as precisely as possible.
19
Q
What is the hypothesis of a research study?
A
This is a precise and testable statement predicting the relationship between two variables. This is usually derived from a theoretical explanation.
20
Q
What is a directional hypothesis?
A
predicts the direction in which any differences (/correlations/associations – Non-Experimental Methods) in the results of an investigation are expected to occur.A directional hypothesis should only be generated if there is previous research upon which to base that prediction – it cannot be based on the researcher’s hunch!
21
Q
What is a non-directional hypothesis?
A
does not predict the direction in which any difference (/correlation/association – Non-Experimental Methods) in the results of an investigation are expected to occur.
This type of hypothesis would be generated if there is no previous research.
22
Q
What is a null hypothesis?
A
predicts that there will be no significant difference (/correlation/association) found e.g. it predicts that the IV will not affect the DV.
23
Q
What is the independent groups design?
A
This is where participants only take part in one of the conditions, so each group does one level of the IV.
Therefore, two entirely different groups of participants are compared against each other in terms of their performance.
24
Q
How could you allocate participants fairly to conditions to reduce individual differences?
A
Random allocation: Use a random technique, such as picking names out of a hat, or using a random name generator.
Participants will then have an equal chance of being in either condition and participant variables should be evenly distributed
25
No order effects strength for IGD
No order effects: As different participants do each condition there are no order effects whereby the order in which the conditions are done may have an effect on the outcome. Practice and boredom effects are types of order effects and they also do not occur when using this design.
Practice effects – the participant becomes practiced at the task because they are doing it more than once, which leads to improved performance in the second condition.
Boredom effects – the participant becomes bored with the task, which leads to a deterioration in performance in the second condition.
26
IGD: reduced risk of demand characteristics:
Participants are only exposed to one of the experimental conditions, therefore they are less likely to guess the purpose of the study and change their behaviour accordingly.
27
IGD: No control of participant variables/individual differences
for example, participants in group 1 may have superior memory ability, therefore results might reflect group differences as opposed to the manipulation of the IV.
28
IGD: less economical:
Twice as many participants are needed.
29
What is the repeated measures design?
This is where participants take part in all conditions within the experiment i.e. all participants experience all levels of the IV.
Participants are therefore compared against themselves.
30
RMD: participant variables/ individual differences:
(the different abilities or characteristics of each participant) are controlled as the same participants are used in each condition i.e. participants are compared against themselves.
31
RMD: more economical
Fewer participants are needed, compared to an independent groups design.
32
RMD: order effects
The order of the conditions might affect performance (order effect), which can act a confounding variable:
Participants may perform better in the second condition because of a practice effect (they become practiced at the task), a positive order effect;
Participants may perform worse in the second condition because they have grown bored with the task (boredom effect), a negative order effect.
33
RMD: High risk of demand characteristics:
Participants are exposed to both conditions, so are more likely to guess the purpose of the study and change their behaviour to please the experimenter.
34
How could you deal with the issue of order effects when using a repeated measures design?
Counterbalancing is a technique that can be used to distribute order and practice effects. It involves changing the order of the conditions from one participant to the next, so half the participants experience the conditions in one order and the other half in the opposite order.
It ensures that each condition is tested first or second in equal amounts.
35
Other ways of dealing with the issues created from using a repeated measures design
Create a time interval between conditions (e.g. ask ppts to return two weeks later to complete the second condition).
Vary the stimulus material between conditions (e.g. use different word lists, but, they must be of equal difficulty to learn – words of a similar length/familiarity).
Use a different experimental design (e.g. independent groups or matched pairs, if appropriate).
36
What is the matched pairs design?
In a matched pairs design, participants are matched across the conditions according to certain relevant key characteristics (that might have an impact on the DV), such as IQ, age, memory ability.
37
MPD:Reduced participant variables:
By trying to ensure that the two groups are similar on some level e.g. age, IQ, age or memory ability, the results should be less affected by participant variables/individual differences and instead be due to the manipulation of the IV.
38
MPD:No practice or order effects and less likelihood of demand characteristics,
because participants only take part in one condition, therefore they are less likely to guess the aim of the study and alter their behaviour accordingly.
39
MPD: Difficult to match participants exactly:
It is impossible to control all participant variables because you can only match on variables known to be relevant, but an unmatched variable might be vitally important. Even two closely-matched individuals will have different levels of motivation or fatigue at any given moment in time.
40
MPD:Time consuming to match participants
: It can take some time for the researcher to identify pairs of participants that are well matched.
41
What is target population?
The group of people that the researcher is interested in, from which a sample is drawn - who they will generalise (apply) their findings to.
42
What is a sample?
The group of participants who take part in the study.
43
What is opportunity sampling?
The researcher selects participants who are available and willing to take part, asking whoever is around at the time of their study to take part
e.g. asking people in the street who are passing.
44
Opportunity sampling : strength: convenient
The most straightforward sampling technique - it is quick and easy for the researcher to obtain participants, because they simply select the first suitable participants readily available to them.
45
Opportunity sampling limitation: researcher bias
The researcher may avoid people they do not like the look of and approach those who have more desirable characteristics.
46
Opportunity sampling limitation: Unrepresentative sample
The sample is likely to be biased as it will undoubtedly exclude certain types of people, because it is drawn from a very specific area, such as one street in one town, so findings can’t be generalised to the target population.
e.g. if you select your sample from people walking around a town centre on a Monday morning, then it would be unlikely to include professional people or people from rural areas.
47
What is volunteer sampling?
Participants select themselves to be part of the sample, often by replying to adverts.
48
Volunteer sampling strength: easy
Creating the sample requires little effort from the researcher, other than creating an advert.
49
Volunteer sampling strength:Less chance of ‘screw you’ phenomenon:
Participants are willing and likely to take it seriously; less chance of them deliberately trying to sabotage the study.
50
Volunteer sampling : limitation: volunteer bias
Volunteers may be a certain ‘type’ of person e.g. more motivated, helpful and curious, and thus unrepresentative, making it difficult to generalise findings to the target population.
51
Volunteer sampling : limitation: demand characteristics:
Volunteers are eager to please and may behave in a way they think is required.
52
What is systematic sampling?
Uses a predetermined system to select participants - every nth (where n is any number) member of the target population is selected e.g. every 5th pupil on a school register.
A sampling frame is produced, which is a list of people in the target population organised into alphabetical order, for example.
A sampling system is nominated by calculating the size of the population and then assessing what size the sample needs to be to work out what the sampling interval is.
53
systematic sampling strength: unbiased selection:
Participants are selected using an objective system, avoiding researcher bias. This increases the chance of a representative sample, making it possible to generalise the findings to the population.
54
systematic sampling : limitation: periodic traits
the process of selection can interact with a hidden periodic trait within the population. If the sampling technique coincides with the frequency of the trait, it is neither random nor representative of the target population e.g. every 5th person might be of a certain age/gender/ethnicity.
55
What is random sampling?
Every member of the target population has an equal chance of being selected.
56
random sampling : strength: potentially unbiased:
The probability of this method producing a biased sample is minimal, because all of the members of the target population have an equal chance of being chosen – it is free from researcher bias.
57
random sampling : limitations: potentially biased
Could still end up with a biased sample (e.g. all females could be selected), particularly if the sample size is too small, making the sample unrepresentative and the results not generalisable.
58
random sampling: limitation: time consuming and impractical:
it may be difficult to get full details of a target population and not all members may be available or wish to participate.
59
What is stratified sampling?
Divides a population into sub-groups (or strata) - the sub-groups are identified as characteristics important for the research.
Participants are then selected randomly from the sub-groups.
60
stratified sampling: strength: representative:
There is a proportional and randomly selected representation of sub-groups (avoiding researcher bias), making generalisation of findings possible
61
limitation of stratified sampling:
Time consuming to divide a population into sub-groups and randomly select participants from each.
Requires detailed knowledge of population characteristics, which may be unavailable.
Stratification is not perfect: The identified strata cannot reflect all the ways that people are different, so complete representation of the target population is not possible.
62
What is an extraneous variable?
An extraneous variable is any variable other than the independent variable (IV) that might have an effect on the dependent variable (DV) if it is not controlled.
They are nuisance variables that do not vary systematically with the IV.
63
Examples of extraneous variables include:
Age and intelligence of participants
Lighting, noise or temperature in the laboratory
Gender of the researcher
64
There are three main types of extraneous variables:
Participant variables – concern factors such as participants’ age and intelligence.
Situational variables – concern the experimental setting and surrounding environment, for example, temperature and noise levels.
Experimenter variables – concern changes in the personality, appearance and conduct of the researcher. For example, female researchers may gain different results from male ones.
65
What are confounding variables?
A variable under study that is not the IV, but which varies systematically with the IV (essentially, it’s an unwanted IV).
Changes in the DV may be due to the confounding variable rather than the IV, and therefore the outcome is meaningless.
66
What is participant reactivity/effects?
Participant reactivity/effects is a significant extraneous variable within experimental research and one that is very difficult for researchers to control.
Demand characteristics create participant effects
67
What are demand characteristics?
Demand characteristics are any cue from the researcher or from the research situation that makes participants unconsciously aware of the aims of a study or helps participants work out what the researcher expects to find.
68
what is the 'please-u effect'?
Participants may act in a way they think is expected and over-perform to please the experimenter.
69
what is the 'screw-u effect'?
Participants may deliberately under-perform to sabotage the results of the study.
70
What are investigator effects?
Investigator effects can also act as an extraneous variable in experimental research.
They are any effect of the investigator’s behaviour (conscious or unconscious) on the research outcome (the DV).
This may include everything from the design of the study (e.g. materials, instructions) to the selection of, and interaction with, participants during the research process.
71
What is a single blind design?
The participant is not aware of the research aims and/or which condition of the experiment they are receiving.
This prevents the participant from seeking cues about the aims and reacting to them, thus controlling for demand characteristics.
72
What is the double blind design?
Both the participant and the person conducting the experiment (who has not designed the study) are ‘blind’ to the aims and/or hypotheses of the research, as well as which condition participants are in.
This prevents investigators from unconsciously giving participants clues as to which condition they are in and therefore reduces demand characteristics.
For example, in drug trials, the drug and placebo would be allocated in such a way that neither the participants nor the researcher would know who was receiving which.
73
What is randomisation used for?
One of the ways researchers can minimise the effect of extraneous/confounding variables is to use randomisation, which refers to the use of chance wherever possible to reduce the researcher’s influence on the design of the investigation. Aside from using a double blind design, it is an attempt to control investigator effects.
74
What is meant by ‘standardisation’?
Standardisation means keeping everything the same for all participants so that the investigation is fair.
75
Why is it important for the procedures of a study to be ‘standardised’?
To ensure that all participants have the same experience, researchers should ensure that they are all tested:
In the same place, with the same equipment and materials placed in the same way.
Under the same conditions, so the level of lighting, noise and heat remains the same for all participants.
At roughly the same time of day, as people may behave differently if tested at nine o'clock in the morning rather than five o'clock at night.
Given identical instructions in exactly the same way – to minimise investigator effects.
76
What are standardised instructions?
The experimenter should write out a set of instructions ahead of the study, clearly outlining in detail what the participants are expected to do.
Researchers should present participants with the instructions to read, or read them out loud to each participant before the study commences.
77
Why is it important for researchers to produce a set of ‘standardised instructions’ as part of their materials?
Standardised instructions are given to the participants to ensure that all of the participants in the study receive precisely the same instructions.
It ensures that each participant is treated the same and should help to reduce investigator effects.
78
What is internal validity?
The extent to which the study is testing what it intends to test e.g. in the context of an experiment, it is the degree to which the observed change in the DV was the result of manipulation of the IV.
If extraneous/confounding variables are bringing about a change in the DV, then the internal validity of the study is lowered as cause and effect can no longer be established.
79
What is social reliability Bias?
Participants wish to present themselves in the best possible way and therefore may not behave as they would do normally, but in a more socially acceptable way.
80
What are investigator effects?
Investigator effects can also act as an extraneous variable in experimental research.
They are any effect of the investigator’s behaviour (conscious or unconscious) on the research outcome (the DV).
This may include everything from the design of the study (e.g. materials, instructions) to the selection of, and interaction with, participants during the research process.
81
What are order effects?
Order effects: where the order in which the conditions are done may have an effect on the outcome. e.g. practice & boredom effects
82
What is face validity?
The extent to which something looks like or appears to measure what it intends to measure. This could be established by using a rating scale to assess the suitability of the measure. Those asked to rate the suitability could be the participants themselves or others in a position to offer judgement. A high suitability rating might suggest a valid test.
83
What is concurrent validity?
Comparing the results of a new measure against the results of another measure (of the same behaviour) that has already been established as being valid. The scores are then correlated. A high positive correlation (exceeds +0.8) between the two measures would indicate concurrent validity.
84
What is external validity?
The extent to which the findings can be generalised beyond the study to…
Different places or settings (ecological validity)
Different people or populations (population validity)
Different points in time (temporal validity)
85
What is ecological validity?
Part of ecological validity concerns whether the experiment mirrors the real world. This is called mundane realism or ‘representativeness’.
Another part of ecological validity concerns ‘generalisability’ – the extent to which findings from one study (conducted in a unique setting) can be generalised to other settings (including the ‘real world’).
86
What is internal reliability?
Internal reliability is a measure of the extent to which something is consistent within itself.
87
What is external reliability?
External reliability is a measure of the extent to which one measure of something e.g. IQ, varies from another measure of the same thing.
88
What is the split- half method?
Participants complete a test e.g. IQ test. The researcher then splits the test into two in terms of the data collected.
If the two halves of the test provide similar results, then this would suggest that the test has high internal reliability.
A Spearman’s Rho test would be used to assess the level of consistency (or relationship) between the scores from both halves of the test.
A strong positive correlation (+0.8) would suggest high internal reliability.
89
What is the test-retest method?
This involves assessing the same participant twice over a period of time using the same measure e.g. administering an IQ test 6 months apart.
Similar scores would suggest that the test has high external reliability.
A Spearman’s Rho test would be used to assess the level of consistency between the scores collected on both occasions – there would be pairs of scores
A strong positive correlation (+0.8) would suggest high external reliability.
90
What is replication?
The likelihood of the same differences occurring twice (or more) by chance alone are much smaller than when they occur the first time.
Effects that occur in a study are more likely to be reliable if they occur in a repeat of the study.
Replication therefore increases (external) reliability.
91
What is a pilot study?
Pilot studies are small-scale (involving a handful of participants) trial-run investigation, carried out prior to the main study to identify potential problems with the design, materials, procedure or analysis, so they can be fixed.
92
What is descriptive statistics?
Provide a summary of a set of data, drawn from a sample, that applies to a whole target population.
Measures of Central Tendency
Provide information about central (or middle/mid-point) scores for a set of data.
There are three measures of central tendency:
Mean
Median
Mode
Measures of Dispersion
Provides information about how spread out the data items are.
Range
Standard Deviation
93
What is the mean?
The mean is the mid-point of the combined values of a data set.
It is calculated by adding all the scores and dividing by the total number of scores.
94
evaluation of the mean?
It is the most accurate and sensitive measure of central tendency, as it takes account of all values within a set of data.
It can only be used with interval or ratio data.
It can be easily distorted by one (or a few) extreme values/outliers (e.g. large or small scores).
The mean score may not be one of the actual scores in the data set.
95
What is the median?
The central score in a list of rank-ordered scores.
With an odd number of scores, the median is the middle number.
With an even number of scores, the median is the mid-point between the two middle scores.
96
What is the evaluation of the median?
It can be used with ratio, interval and ordinal data.
It is not affected by extreme scores.
It is easier to calculate than the mean.
It is not as sensitive as the mean, because not all scores are used in the calculation.
97
What is the mode?
The mode is the most common or “popular” number in a set of scores.
98
What is the evaluation of the mode?
It is useful when the data are in categories (nominal data).
It is less prone to distortion by extreme values.
It sometimes makes more sense than other measures e.g. the average number of children is better described as 2 (mode) than 2.4 (mean).
There can be more than one mode in a set of data.
99
What is the range?
The range is the distance between the top and bottom values in a set of data.
It is calculated by subtracting the lowest value from the highest value in a set of data.
100
Evaluation of the range?
It is easy and quick to calculate.
It takes full account of extreme values.
It can be distorted by extreme values.
It does not show whether data are clustered or spread evenly around the mean, therefore failing to take account of the distribution of the numbers.
101
What is standard deviation?
It is a measure of the spread of a set of scores from the mean.
It measures the average distance between each data item above and below the mean.
The larger the standard deviation, the larger the spread of scores (from the mean).
102
What is the evaluation of standard deviation?
It is a more precise and sensitive dispersion measure than the range, since all scores are used in its calculation.
It allows for the interpretation of individual scores.
It may hide some of the characteristics of the data set e.g. extreme values.
It is more complicated to calculate.
It is less meaningful if data are not normally distributed.
103
What are distribution curves?
Distribution curves can be formed by plotting the data from a histogram and then overlaying a smooth continuous line
104
What is normal distribution?
The normal distribution is a symmetric distribution with no skew. The tails are exactly the same, in that one half of the distribution is a mirror image of the other half.
Most of the data values in a normal distribution tend to cluster around the mean i.e. approximately 95% of the data lies within 2 standard deviations of the mean.
There are many things, such as intelligence, height, and blood pressure, that naturally follow a normal distribution.
For example, if you took the height of one hundred 18-year-old women and created a histogram by plotting height on the x-axis, and the frequency at which each of the heights occurred on the y-axis, you would get a normal distribution.
The mean, mode and median are all equal.
105
What is skewed distribution?
A distribution is skewed if one tail is longer than another. These distributions are sometimes called asymmetric distributions.
A left/negatively-skewed distribution has a long left tail in the negative direction on the number line - the scores fall toward the higher side of the scale and there are very few low scores.
106
A distribution that is negatively skewed has the following characteristics:
the mean is typically less than the median (which is less than the mode);
the tail of the distribution is longer on the left hand side than on the right hand side.
107
A distribution that is positively skewed has the following characteristics:
the mean is typically greater than the median (which is greater than the mode);
the tail of the distribution is longer on the right hand side than on the left hand side
108
Why do we use statistical tests in Psychology?
Researchers use statistical tests to test the hypotheses (‘hypothesis testing’) of an investigation by establishing whether the result is significant or not i.e. to determine the likelihood that the effect/difference/relationship they have found has occurred due to chance.
109
In order to apply an inferential statistical test to a set of data, the following must be in place:
- A null hypothesis (H0) – this predicts that there will be no significant difference / correlation / association found.
- And an alternative/research hypothesis (H1) – predicts that a significant difference / correlation / association will be found. This is either:
a directional hypothesis (one-tailed test) – if there is previous research
or a non-directional hypothesis (two-tailed test)
110
What are the steps of hypothesis testing?
Generate hypotheses
design a research study to test hypotheses
collect data
analyse data using a statistical test and inter outcome according to table of critical values, using an appropriate level of significance
accept or reject hypothesis
theory is strengthened if the hypothesis is supported or revised if refuted
111
112
112
113
What is the calculated value?
An inferential statistical test will produce a test statistic.
The test statistic is known as the calculated value - it is based on the observations made. It is the result generated from applying the statistical test to the raw data.
113
What is the sign test and when would you use it ?
The Sign Test is a test that looks for consistent differences between two sets of data.
It is used for related (repeated measures) data. This means the participants are the same in each condition. It can also be used for matched pairs design when each participant is paired with another.
It can be used when the level of measurement is nominal or above.
113
what is critical value?
To decide if the calculated value is significant or not, it is compared against the critical value.
This is the value that must be reached for the null hypothesis to be rejected.
114
What is a nominal level of measurement?
Data is organised in categories, according to frequencies e.g. number of smokers/non-smokers; number of males/females
115
what are ordinal levels of measurement?
Data can be ordered in some way e.g. lining classmates in height order – the difference between each item is not equal (e.g. psychological scales, where there is an element of subjectivity).
116
What are interval levels of measurement?
Data uses units of equal intervals e.g. temperature, using both +/- numbers (public scales of measurement e.g. cm, time in seconds)
117
How do you explain the outcome of statistical test?
“The result is/is not significant, because the calculated/observed value of S (___) is equal to/greater/less than the critical value of S (___), when p=0.05, N is ___, using a one-/two-tailed test.”
Significant result = accept alternative hypothesis; reject null hypothesis
Not significant = accept null hypothesis; reject alternative hypothesis.
