research methods

• = -> equal

  < -> less than

  << -> much less than

  >> -> much greater than

  > -> greater than

  ~ -> approximately
• IV -> what you change

  How to operationalise an IV -> by clearly defining 2 (or more) conditions of the experiment
• DV -> what you measure

  How to operationalise the DV -> by measuring the concept numerically
• -> a target population of a survey is the population you wish to study

  -> a sample population is the population which you are able to observe in a sample
• OPPORTUNITY SAMPLE -> anyone whose available & agrees to take part can become a partiticpent

  STRENGTH -> quick & convenient way of obtaining large sample of participants

  -> useful for studying processes - E.G : memory ,, ETC!

  WEAKNESSES -> likely to be biased as culturally & ethically limited

  -> may not be able to generalise findings to all members of target population
• VOLUNTEER SAMPLING -> here participants choose to take part in study

  STRENGTHS -> quick & convenient way of obtaining participants

  -> people have volunteered ,, less likely to drop out

  WEAKNESSES -> volunteers may be more outspoken ,, have more time ,, may be more interested in the topic being studies

  -> unlikely to be representative & can't generalise findings to members of the target population
• RANDOM SAMPLING -> every member of the target population have an equal chance of being chosen

  STRENGTHS -> gives most unbiased ,, representative sample

  -> this method will mean results will be most generalisable to the target population

  WEAKNESSES -> once selected ,, participants may not want to participant

  -> difficult to gain access to names of all members of the target population
• SNOWBALL SAMPLING -> couple of participants who fit into the target population & then asks them to find similar people who also fit the target population

  STRENGTHS -> useful way of gathering participants that are hard to find

  WEAKNESSES -> likely to be biased as the participants have similar characteristics

  -> samples less representative ,, decreasing generalisability of results
• REPRESENTATIVENESS -> a sample that closely matches the target population as a whole

  GENERALISABILITY -> the extent to which the findings of a study can be applicable to other settings
• ● OPEN QUESTIONS -> '' why do you like chocolate?''

  STRENGTHS -> delivers detailed information

  -> more validity

  WEAKNESSES -> difficult to analyse & comparisons can't be made between different people's responses can't be made

  ● CLOSED QUESTIONS -> '' do you like chocolate? yes or no?''

  STRENGTHS -> easier to analyse

  -> allows different responses be compared

  WEAKNESSES -> can't understand reasons for behaviour - lacks detail

  -> less validity
• ● RATING SCALE QUESTIONS -> '' on a scale of 1-10 ,, how much do you like chocolate?''

  ● LIKERT SCALE -> '' mint chocolate is the best chocolate ''

  ● SEMANTIC-DIFFERENTIAL SCALE -> '' rate your attitude towards chocolate ''

  STRENGTHS -> allows comparisons to be made as quantitative data is collected

  -> have more options ,, more valid

  WEAKNESSES -> central tendency bias - may think answers too extreme so they choose around middle

  -> different people may interpret values if 1-10 / how far they agree than other participants
• LAB EXPERIMENT -> conducted in a special environment where variables can be carefully controlled

  STRENGTHS -> high control allows cause & effect relationships to be established

  -> high control means experiment can be replicated

  WEAKNESSES -> lacks ecological validity as too highly controls

  -> total control over extraneous variables is impossible

  -> ethical concern - deception
• FIELD EXPERIMENT -> an experiment conducted in a more natural environment

  STRENGTHS -> greater ecological validity than lab experiment

  -> less bias from demand characteristics if participant are unaware of being studied

  WEAKNESSES -> more difficult to control ecological validity so more likely to affect results

  -> more difficult to replicate exactly

  -> ethical problem - consent ETC
• QUASI EXPERIMENT -> the independent variable isn't being manipulated by the experimenter but is changing/naturally occurring in the environment

  STRENGTHS -> depends on whether it takes place in a lab or field setting

  -> greater ecological validity as it often occurs in a natural environment

  WEAKNESSES -> depends on whether it takes place in a lab or field setting

  -> impossible to replicate exactly due to the lack of manipulation of an IV
• ETHICAL ISSUES- RESPECT

  ● INFORMED CONSENT -> all participants should agree to participate in research and must know what they're agreeing too

  *Sign an informed consent sheet!*

  ● CONFIDENTIALITY -> must remain anonymous

  *Must conform to DATA PROTECTION ACT ,, participants referred by letter/number*

  ● WITHDRAWAL -> participants must be made aware that they can leave at any time

  *allow them to leave at any time*
• ETHICAL ISSUES - RESPONSIBILITY

  ● PROTECTION OF PARTICIPANTS -> keeping participants safe from physical & mental harm

  *should be assessed for any factors which may create a risk ,, participants should know they can contact the researcher*

  ● DEBRIEFING -> participants must be debriefed ,, it they've been deceived the true nature of the study should be explained to them

  *experiment should be discussed ,, should be offered a write up of the study & results *
• ETHICAL ISSUES - INTEGRITY

  ● DECEPTION -> shouldn't be deceived about the true nature of the study

  *should be justified to an ethics committee (decides whether research can go ahead)*
• ETHICAL ISSUES - COMPETENCE

  ● AWARENESS OF PRO ETHICS -> psychologists should work within their own capabilities
• ● ALTERNATIVE HYPOTHESIS -> statement of difference between 2 (or more) conditions

  ● NULL HYPOTHESIS -> statement of no difference between 2 (or more) conditions

  ● ONE TAILED -> predicts direction of difference ,, whether it's increasing ,, decreasing ,, fast or slow

  ● TWO TAILED -> ONLY predicts difference ,, NO direction
• ● NOMINAL DATA -> where DVs measured in 2 categories/frequencies

  EG -> tall & short ,, pass & fail

  STRENGTHS -> some data can be recorded as nominal

  WEAKNESSES -> only labels with no arithmetic properties - only know the frequency

  ● ORDINAL DATA -> different values which can be placed in an 'order'/rank - rating scales

  EG -> 1st John ,, 2nd Anna ,, 3rd Tim ,, 4th Hamzah

  STRENGTHS -> allows us to judge magnitude

  WEAKNESSES -> can't make meaningful judgements about individual pieces of data
• ● INTERVAL DATA -> measured on a scale ,, equal intervals

  EX -> measure your height in cm

  STRENGTHS -> allows us to judge magnitudes

  -> can make meaningful comparisons between individual pieces of data
• SOCIALLY DESIRABLE BIAS -> the participant You are not giving the 'correct' answer to the question but an answer which makes you appear to be a better person than you are
• FULLY OPERATIONALISED NULL HYPOTHESIS

  1 -> state that there will be NO difference between the 2 conditions/groups

  2 -> state the two (or more) conditions of the IV

  3 -> state the DV

  4 -> mention how the DV has been operationalised
• FULLY OPERATIONALISED ALTERNATE/EXPERIMENTAL HYPOTHESIS

  1 -> state there will be a difference between the 2 conditions/groups

  2 -> state the two (or more) conditions of the IV

  3 -> state the DV

  4 -> mention how the DV has been operationalised
• TYPES OF INTERVIEWS

  STRUCTURED -> general topic is clearly defined ,, but set questions presented in a pre-determined order

  -> like a face-to-face questionnaire

  SEMI-STRUCTURED -> general topic clearly defined ,, SOME set questions but also be able to follow up on responses

  UNSTRUCTURED -> general topic clearly defined but NO set questions

  -> interviewees can talk as freely as they like

  -> RARELY USED!
• INTERVIEWS vs QUESTIONNAIRES - WEAKNESSES

  -> interviews requires some contact between interviewer & interviewee so this means interpersonal variables in the interview situation could affect how the interviewee behaves.

  -> they may be more likely to give socially desirable answers when face-to-face with the 'live' interviewer

  -> gender ,, ethnicity & personalities of both interviewer & interviewee can also affect the responses given in an interview
• INTERVIEWS vs QUESTIONNAIRES - STRENGTHS

  -> a strength - participants can ask for clarity if there are any ambiguous questions they dont understand

  -> the interviewer can determine the validity of a persons answer by assessing their tone of voice ,, body language ETC

  -> interviews (SPECIFICALLY SEMI & UNSTRUCTURED ONES) can be tailored to the individual respondents
• MEASURES OF CENTRAL TENDENCY

  mean -> add the numbers together and divide it by the amount of numbers there is

  median -> put the numbers in order and find the middle

  mode -> the number that repeats the most
• INDEPENDENT MEASURES DESIGN -> there are different participants in each condition of the experiment - performance of 2 groups of people are being compared

  STRENGTHS:

  -> no problems with order effects (boredom etc) as they take part once

  -> may be used where a repeated designs impractical/possible to use

  -> less likely to figure out what the experiments about so no demand characteristics

  WEAKNESSES:

  -> influence of participant variables - different people take part in each conditions of the experiment

  -> need twice as many participants as a REPEATED MEASURES DESIGN
• REPEATED MEASURES DESIGN -> ALL participants take part in each conditions of the experiment - you are comparing each participants performance in one condition with their performance in the other

  STRENGTHS:

  -> Participant variables are removed as same people take part in each condition of the experiment

  -> fewer participants are required

  WEAKNESSES:

  -> participants are more likely to work out the aim of the study so they may display demand characteristics

  -> participant will do the same task twice which may lead to order effect
• MATCHED PARTICIPANTS DESIGN -> 2 separate groups are again being tested - here each participant in one group is matched on certain characteristics as a participant in other groups

  STRENGTHS:

  -> no problems with order effects as participant only takes part in ONE condition

  -> Less likely to figure out what the experiment is about so no demand characteristics

  WEAKNESSES:

  -> difficult to achieve fully matched pairs, as participants would be need to be matched on all the variables that could potentially affect their performance in the experiment

  -> it is time consuming to do this properly
• PARAMETRIC ASSUMPTIONS

  In order to carry out a parametric statistical test ,, data must meet the following requirements:

  -> data must be interval level

  -> data should follow a normal distribution

  -> data must be drawn from groups with similar variances
• CHI SQUARE CHECKLIST

  CHECKLIST:

  -> DV produces nominal data

  -> independent measures design

  -> exploring a difference between conditions (levels of the IV)
• CHI SQUARE CALCULATIONS


  CALCULATION:

  1 -> add the totals for each column & row

  2 -> calculate the expected frequency - row total x column total / overall total

  3 -> x^2 = (observed-expected)^2/expected *ADD ALL CELLS UP*

  4 -> calculate the degrees of freedom - (no of rows - 1) x (no of columns - 1)

  5 -> use table of critical values to find critical value (0.05)

  6 -> the calculated value has to be greater than/equal to critical value
• BINOMIAL SIGN TEST CHECKLIST

  CHECKLIST:

  -> DV produces nominal data

  -> repeated measures design

  -> exploring a difference between conditions (levels of the IV)
• BINOMIAL SIGN TEST CALCULATION

  1 -> work out the flow of direction ( + ,, - ) ,, if responds are the same in each conditions ignore them

  EG -> positive , negative + .. negative , positive -

  2 -> count total positive and negative signs you have

  3 -> smallest score is the calculated value

  4 -> critical value - total number of positive and negative

  5 -> for a binomial sign test ,, the calculated value has to be LESS THAN/EQUAL to the critical value

  -> greater than ,, not a significant difference
• MANN-WHITNEY U TEST CHECKLIST

  CHECKLIST -> DV produces ordinal or interval data

  -> independent measures design

  -> exploring a difference between conditions (levels of the IV)
• MANN-WHITNEY U TEST CALCULATIONS

  1 -> rank ( from low to high) - if the numbers are the same - add the ranks up divide by the amount of numbers

  2 & 3 -> add up the ranks for the 1st (RANK 1) & 2nd (RANK 2) column

  4 -> use the smallest total in this formula!

  U1 = R1/2 - n1/2 (n1/2+1) /2 - N - number of participants in a condition

  5 -> use table of critical U values for Mann-Whitney U Test

  n1/2 -> number of participant in condition 1/2

  6 -> calculated value is LESS THAN / EQUAL TO the critical value of U

  -> LESS THAN ,, accept alternate ,, null hypothesis can be rejected
• TYPES OF INFERENTIAL TESTS - 1

  CHI SQUARED TEST -> nominal data

  -> difference

  -> independent measures design

  BINOMIAL SIGN TEST -> nominal data

  -> difference

  -> repeated measures design

  MANN WHITNEY U TEST -> ordinal / interval data

  -> difference

  -> independent measures design
• TYPES OF INFERENTIAL TESTS - 2

  WILCOXON SIGNED TEST RANKS -> ordinal / interval data

  -> difference

  -> repeated measures design

  SPEARMEN'S RHO TEST -> ordinal / interval data

  -> relationship

  -> correlation