A-Level Psychology - PAPER 2 - Research Methods Part 3

results could have been a fluke

all details of a study need to be published

compare results

needs to be able to be empirically tested to see if it false

numerical data

| - analysed using statistical techniques

OBJECTIVE:

does not require interpretation

less prone to bias

EASY TO ANALYSE:

computer programmes

allows larger sample size = generalisable

DOESN'T TELL US WHY:

cause and effect

hard to make PRACTICAL APPLICATIONS

NARROW:

only certain behaviour can be measured this way

reduces the SCOPE OF STUDY

detailed information

| - themes in pps responses

RICH DETAIL:

more representative of real life

CAN EXPLAIN WHY:

cause and effect

develop more accurate theories

practical applications

SUBJECTIVE

open to interpretation

bias

reduces VALIDITY

DIFFICULT TO ANALYSE

transcripts have to be written

time consuming

smaller sample size = not generalisable

average

| - adding all the scores up and dividing by number of scores

most SENSITIVE and REPRESENTATIVE

| - takes all scores into account

distorted by EXTREME SCORES

| - UNREPRESENTATIVE

middle score

unaffected by EXTREME SCORES

only looks at one or two scores

| - generally used with ORDINAL DATA

most frequent

unaffected by EXTREME SCORES

affected by the change in one score

UNREPRESENTATIVE

used with NOMINAL DATA

mean

median

mode

provides a SINGLE VALUE which is REPRESENTATIVE pf a set of numbers by implicating the most TYPICAL VALUE

range

| - standard deviation

difference between the highest and lowest scores and adding 1 (allows for any rounding that has occurred)

quick to calculate

| - takes account for EXTREME VALUES

doesn't provide an idea around the distribution of values around the centre

does account for INDIVIDUAL VALUES

affected by EXTREME SCORES

VARIABILITY of scores from its MEAN

considers ALL the scores

| - sensitive

difficult to calculate

less meaningful if data isn't normally distributed

distorted by EXTREME SCORES

vertical bars of equal distance apart

| - nominal data

shows distribution of scores

continuous scale

ordinal / interval data

plotting correlations

| - visual image

descriptive statistics or percentages

data is symmetrical

| - forms bell-shaped curve on graph

data is NOT symmetrical

mean moves to the RIGHT

| - HIGHER SCORING pps have moved mean to the right

a hard test where most students didn't score very well

mean moves to the LEFT

| - LOWER SCORING pps have moved the mean to the left

a test which was easy and most students scored well

counting frequency data

separate categories

each piece of data can only go into one category

counting whether pps are happy or sad

rating on a scale

John came first, Fred came second, Brian came third

similar to ordinal

| - has a UNIT e.g. grams

measure time in seconds

allow researchers to draw conclusions about their research

psychologists need to mathematically express the likelihood their result occurred due to chance

difference or relationship ?

experimental design type ?

type of data ?

null hypothesis is rejected when it should have been accepted

FALSE POSITIVE

significant level isn't harsh enough

null hypothesis is accepted when it should have been rejected

FALSE NEGATIVE

significant level is too harsh

strikes a balance between making type I and type II errors

DESIGN - experimental design, variables, controls

MATERIALS - any special materials

DATA ANALYSIS - reference descriptive and inferential analysis

am i looking for a DIFFERENCE or RELATIONSHIP ?

what is my EXPERIMENTAL DESIGN ?

what type of DATA do i have (nominal / ordinal / interval)

CHI-SQUARE

SIGN TEST

CHI-SQUARE

MANN-WHITNEY

WILCOX

SPEARMAN'S RHO

UNRELATED t-TEST / INDEPENDENT t-TEST

REALTED t-TEST

PEARSON'S R

COUNTERBALANCING

splits pps so they complete different levels of the IV in a different order

balance out order effects

RANDOM ALLOCATION

| - pps randomly allocated condition to distribute them evenly

PILOT STUDY

| - consider key variables that may effect the DV