Research Methods Flashcards

Question

Evaluate Natural Experiments

Answer 1

Good - Creates opportunities for studies that might not have been possible (earthquakes, volcano eruptions) - High ecological validity due to using natural problems and real world issues Bad - Rare to find a naturally occurring event - Limits generalisation - No control as IV exists already - Participants may not be randomly allocated

Answer 2

An Experimental Method where the variables already exist, so the IV has not and cannot be determined by anyone e.g. twins, adoption - cannot control who is a twin and who has been adopted

Answer 3

Good - controlled conditions - good replicability - can be certain of cause and effect Bad - Possible confounding variables as you cannot randomly allocate participants - Cannot claim the IV has caused any observed change as it has not deliberately been changed by the researcher

Answer 4

Demand Characteristics Confounding Variables Extraneous Variables Standardisation (lack of) Investigator Effects Randomisation

Answer 5

When participants try to work out what is going on in the experiment (the aim), and change their behaviour so it is no longer natural They can show the 'please you' or 'screw you' effect where they deliberately please the experimenter or sabotage the experimenter

Answer 6

Unconscious actions or Unstandardised procedures that may influence the research outcome

Answer 7

Participant Variables They are the differences between participants such as age, gender or personality

Answer 8

Situational Variables They are features of the experimental situation such as noise, temperature and weather

Answer 9

Independent Groups Repeated Measures Matched Pairs

Answer 10

An Experimental Design where two separate groups of participants experience two different conditions of the experiment - all participants experience only one level of the IV - one group would do the control condition and one would do the experimental condition - performance of the 2 groups would be compared

Answer 11

Good - Less chance of demand characteristics as participants only complete one so cannot figure out the aim - No order effects as they have only done one - No practice effects as they have only done one - Can use random allocation Bad - Participant variables might make it difficult to establish cause and effect - there may be confounding variables - Need a larger sample - Takes longer/costs more

Answer 12

An Experimental Design where all participants experience both conditions of the experiment - Each participant completes one condition (either experimental or control) - Participants then swap and complete the other condition afterwards The scores from both conditions would be compared to see the differences

Answer 13

Good - Counterbalancing means that participant effects are minimised as every participant is completing every condition - Fewer people are needed as they take part in both conditions - Allows for Random Allocation Bad - Practice effects - participants may be used to the task, or may work out the aim if they complete it more than once - Order effects - participants may be tired after the first condition, or may care less and not try as hard on the second condition

Answer 14

An experimental design where participants are matched based on possible participant variables that may affect the DV e.g. 2 people with glasses, 2 people with hearing aids, 2 people aged 22 etc. - one person from the pair completes one condition and the other person completes the other

Answer 15

Good - Reduces Participant Variables by having similar people in each condition - Avoids practice effects - demand characteristics are less likely - Avoids order effects Bad - Participants can never be matched exactly, even if they were identical twins - Matching might be time-consuming and expensive

Answer 16

The large group of individuals a researcher is interested in studying Also called the target population

Answer 17

The smaller group who take part in the research

Answer 18

Volunteer Sampling Random Sampling Opportunity Sampling Stratified Sampling Systematic Sampling

Answer 19

Every member of the target population has an equal chance of being selected e.g. names in a hat, random number generator

Answer 20

Good - Should represent the target population - Should eliminate a sampling bias Bad - Difficult to achieve (effort, money) - Time consuming - some people might say no

Answer 21

The researcher selects participants from whoever is available at the time

Answer 22

Good - Quick - Easy (cost effective) Bad - Might be biased - Might not provide a representative sample

Answer 23

Participants put themselves forward to be put in the sample They self-select

Answer 24

Good - Easy to find people willing to participate (less time consuming) Bad - Might not provide a sample representative of the whole population - Might not have enough participants - Might suffer Volunteer Bias if they are too keen to participate (please you effect), or they might all be the same type of person

Answer 25

The target population is broken down into smaller groups, and these are then sampled from The sample is a proportional representation of the target population

Answer 26

Good - Avoids researcher/sampling bias - Can generalise results as sample should be representative of the population Bad - Takes a lot of time - Difficult to do, and some people might say no, so you'd have to start again

Answer 27

Every nth member of the target population is selected e.g. every 8th house on the street every 5th person on the register

Answer 28

Good - Should provide a representative sample as it is a set selection across the whole target population - Reduces researcher bias Bad - Difficult to achieve (time, effort, money) - There might not be enough diversity to represent the whole population

Answer 29

The strength and direction of an association between co-variables (the relationship between them)

Answer 30

Using a Scattergram One Co-Variable is on the X axis, and one is on the Y axis Each point on the graph is the X and Y position of each co-variable

Answer 31

Positive correlation - as one variable increases so does the other Negative correlation - as one variable increases, the other rises No correlation - there is no pattern between the variables

Answer 32

Correlation does not equal causation It simply points out the relationship between the variables

Answer 33

Using a Correlation Coefficient Between 0 and +1 shows a Positive Correlation Between -1 and 0 shows a Negative Correlation For a Correlation to be strong, it must have a Coefficient of at least .8 (-0.8 or +0.8)

Answer 34

Measures of Central Tendency Measures of Dispersion

Answer 35

Measures that give us averages about the typical values in a set: Mean Median Mode

Answer 36

Measures that are based on the spread of scores and how they vary and differ from one another: Range Standard Deviation

Answer 37

The most frequent value There can be 2 modes (bimodal) There can be no mode

Answer 38

See which value is the most frequent

Answer 39

Good - Easy to find Bad - May be more than one answer - May not represent the whole results

Answer 40

The Middle Value when results are placed in order Used for Ordinal and Interval data

Answer 41

Place results in order Find the Middle Value

Answer 42

Good - Easy to calculate for small data sets - Not affected by anomalies (outliers) Bad - Not a true representative of all data - Can be difficult for larger data sets

Answer 43

The overall average of a data set

Answer 44

Add up all the results Divide by the number of results there are

Answer 45

Good - Uses all the data Bad - Can be distorted by outliers

Answer 46

The largest value minus the smallest value It tells us how spread out the data is

Answer 47

Subtract the Smallest Value from the Largest Value

Answer 48

Good - Easy to calculate Bad - Affected by extreme data

Answer 49

A measure of Dispersion that tells us how far the data is from the mean

Answer 50

Good - Accurate Bad - Difficult

Answer 51

Avoid Jargon Avoid Emotive Language Avoid Leading Questions Avoid Double-Barrelled Questions Avoid Double Negatives

Answer 52

The assessment of scientific work by others who are specialists in the same field. They should be objective and unknown by the Researcher.

Answer 53

1) Decide if a proposed research project should receive funding 2) Assess research for quality and accuracy by validating the quality of the hypotheses, methodology, statistical tests, and conclusions 3) To suggest amendments or improvements, or even suggest it is unsuitable for publication

Answer 54

Good - Spots mistakes and prevents invalid or inappropriate data from being published Bad - Some reviewers may use their anonymity to criticise rival researchers on purpose - Burying groundbreaking research - reviewers might reject research that goes against mainstream theory or the status quo - Publication bias - publishers may want research that will make good headlines to increase the readers, or they might want positive results - Takes a long time - Sometimes fraud can be missed if results are fabricated - Bias - peer reviewers might know the person and dislike them or their university - File Drawer Phenomenon - work might be left for too long

Answer 55

Consistency - you get the same results every time you complete the research

Answer 56

1) Test Re-Test (inter-rater reliability) - Carry out the research once - Carry the same research on the same participants at a later date - Calculate the correlation to check how similar the results are - If the coefficient is 0.8 or above, it is reliable 2) Inter-Observer Reliability - Carry out the research once - Have a separate second researcher carry out the research - Correlate the results - If the coefficient is 0.8 or above, it is reliable

Answer 57

If the Correlation Coefficient is less than .8

Answer 58

Questionnaires - remove some questions - re-write some questions - replace open questions with closed questions to reduce ambiguity Interviews - use the same interviewer each time or train all interviewers to standardise the procedure - avoid leading questions - use a structured interview for more control Experiments - use a lab experiment for more control - ensure it is standardised for replicability Observations - operationalise behavioural categories - train observers - standardise

Answer 59

Whether we are measuring what we set out to measure - there are no confounding or extraneous variables that could have affected what we are measuring

Answer 60

How generalisable the findings are beyond the research setting - e.g. to the real world (ecological validity), or across different eras (temporal validity)

Answer 61

It looks like the experiment measures what it is supposed to be measuring

Answer 62

If the results from this study are close to the results from another already established study

Answer 63

The extent to which findings can be generalised to other settings and situations

Answer 64

The extent to which findings can be generalised to other historical times and eras

Answer 65

The researcher or another expert will look at the test and see if it seems to measure what it is supposed to be measuring

Answer 66

Test your participants with your test Test them again with an already established test Collect Results and Correlate the scores High concurrent validity will have a correlation coefficient of .8 or more

Answer 67

Questionnaires - put in a lie scale to test consistency of results - make it anonymous to reduce social desirability Experiments - control extraneous variables by using a control group for comparisons - standardise procedures - use a double blind to reduce investigator effects Observations - use covert observations so behaviour is authentic - ensure behavioural categories are not too broad or ambiguous Aim to use quantitative methods If using qualitative methods, use quotes or triangulation (different sources of the same information)

Answer 68

Nominal Ordinal Interval

Answer 69

Categories Lowest and most basic data such as tally charts Discrete data (1 item can only appear in 1 category)

Answer 70

Ordered Data The rank/place/rating Lacks precision as it is subjective and there is not a set interval between each unit e.g. 1st, 2nd, 3rd

Answer 71

Based on Numerical Scales Includes units of equal, precisely defined size A standardised and operationalised unit of measurement

Answer 72

The Mode It is the most often in all categories

Answer 73

Median Range

Answer 74

Mean Standard Deviation It is the most precise measurement for the most precise level

Answer 75

1) Test of Difference or Test of Association? 2) What Experimental Design is used? 3) What is the Level of Measurement?

Answer 76

In 95% of cases the results would have happened anyway

Answer 77

In 5% or less of cases the results would have happened anyway This suggests the IV has caused the DV

Answer 78

Objectivity Empirical Method Replicability Falsifiability Theory Construction Hypothesis Testing Paradigms and Paradigm Shifts

Answer 79

Karl Popper

Answer 80

Not allowing personal biases to affect our data

Answer 81

Approaches based on gathering evidence through direct observation and experience

Answer 82

The extent to which procedures and findings can be repeated across different circumstances and contexts Important to find validity and reliability

Answer 83

Scientific theories should hold themselves up for hypothesis testing and the possibility of being disproven This is why we always have a null hypothesis Strong theories are those that have been repeatedly tested and not falsified

Answer 84

Gathering evidence to form a theory - a set of general laws or principles that can explain behaviours - could be a hunch that leads to experiments to develop a theory

Answer 85

Testing a theory through empirical methods Can support and strengthen a theory Can disprove a theory

Answer 86

Paradigm - a set of shared beliefs and assumptions - Kuhn suggested psychology has too many conflicting approaches to be a science Paradigm Shift - science progresses through scientific revolution - researchers may begin to question the accepted paradigm, and this gathers pace until a paradigm shift occurs - this is when contradictory evidence cannot be ignored, so there is a shift to the new belief/paradigm

Answer 87

Abstract Introduction Method Results Discussion References

Answer 88

A short summary It includes all major elements of a report - aims - hypotheses - method - procedure - results - conclusion It lets others know if they want to read it in full

Answer 89

A Literature Review of the past research and theories that relate to their study It progresses from general to specific and ends on the most relevant aims and hypotheses

Answer 90

A section including - Design - Sample/Participants - Apparatus/Materials - Procedure - Ethics Should be detailed enough to allow replication

Answer 91

Summarising key findings - Descriptive Statistics - Inferential Statistics - Qualitative results and findings Raw data goes in an appendix

Answer 92

Summarising the results in verbal form Discusses limitations Suggests how they can be modified for future studies Consider the wider implications

Answer 93

Adds any details of source materials

Answer 94

Surname, Initials (Date) Title of article, Journal title, edition. Page numbers

Answer 95

Author, Surnames and initials (Date) Title pf book, place of publication, publisher

Answer 96

Turns Qualitative Data into Quantitative Data

Answer 97

1) Coding - Categorise the data into meaningful units (codes) - These could be categories, themes, phrases, key words 2) Count - Count how many times these codes occur - Highlight the transcripts to count for occurrences There is your quantitative data

Answer 98

Leaves Qualitative Data as Qualitative Data but makes it easier to analyse

Answer 99

1) Identify Themes - identify key ideas that are recurrent Could put these into broader categories 2) Directly quote these to illustrate each theme in the final report

Answer 100

Divide by the total value Multiply by 100

Answer 101

Change/Original x 100

Answer 102

Calculate the Change Find the change as a % of the original value

Answer 103

Calculate the Change Find the change as a % of the original value

Answer 104

When the Null Hypothesis is rejected and the alternative is accepted when the Null Hypothesis is actually true e.g. saying a fat man is pregnant - He could not be pregnant - As people have accepted he is pregnant although he could not be, the alternative hypothesis has been accepted when the null hypothesis is actually true

Answer 105

When the Alternative Hypothesis is Rejected and the Null Hypothesis is Accepted although the Alternative Hypothesis is actually true e.g. saying a heavily pregnant woman is not pregnant - She is pregnant - As people have accepted she is not pregnant although she is, the null hypothesis has been accepted when the alternative hypothesis is actually true

Answer 106

To try to get a balance between Type 1 and Type 2 errors to reduce the risk of both of them