RESEARCH METHODS

1
RESEARCH METHODS

• This topic is examined as part of unit 1 and the
  exam questions are linked to the topics of Memory
  and Early Social Development

2
Research Aims

• It is important when doing research to be clear
  about what you are trying to find out. This is
  known as the Aim of the research.
• The Aim should be written as a clear expression
  of what you are trying to find out.
• For example, you might write something along
  these lines-
• The researcher wanted to find out/ to see if /
  to see whether/ to investigate . . .

3
Example-

• A teacher thinks that the children in her class
  who eat breakfast concentrate better than those
  who dont.
• She carries out an investigation to investigate
  this.
• What would you write as the AIM of the study?
• Have a go

4
You may have written something like-

• The researcher wanted to find out whether eating
  breakfast affects the level of concentration in
  children.
• OR
• The researcher wanted to investigate the effect
  of eating or not eating breakfast on the
  concentration levels of children.

5

• Think about Sperlings study.
• What was his aim?

6
Hypotheses

• A hypothesis is similar to but more specific than
  a research aim.
• It is a precise statement about the expected
  outcome which is to be investigated in a
  psychological study.
• Hypotheses are generated from the findings of
  previous research or from observations made in
  every day life.

7

• A hypothesis is expressed in the form of a
  statement that can be tested when the test is
  done we can decide whether it is true or not.
• It must be expressed in a clear way, in terms of
  things that can be measured. This is known as
  being OPERATIONALLY DEFINED.

8

• Think about this hypothesis-
• People who are deprived are less intelligent
  than people who are not deprived
• Why is it not operationally defined?
• What do we have to think about in order to
  operationally define it?
• How can we make it operationally defined?

9

• Examples might be-
• There will be a positive relationship
  (correlation) between peoples IQ scores and
  their household income.
• There will be a difference in the number of A-C
  GCSE passes of children whose parents have manual
  occupations and children whose parents have
  professional occupations.

10
Types of hypotheses

• The NULL HYPOTHESIS The symbol used is Ho
• This is used so that the researcher can show her
  prediction to be wrong it states that the
  findings of the investigation are not the result
  of the predicted effect but in fact are due to
  chance.
• It is in effect, the conclusion we come to if the
  prediction is not found to be true.
• It is needed as a starting point for statistical
  analysis of the data collected during a study,
  but for AS level you will not be required to
  carry out any statistical significance tests.

11

• The null hypothesis states that either
• There is no significant relationship between the
  two variables being studied, any relationship
  found is the result of chance.
• or
• That one variable has no significant effect the
  other, any effect found is the result of chance.

12
Examples-

• There will be no significant relationship
  (correlation) between peoples IQ scores and
  their household income, any relationship found is
  due to chance.
• There will be no significant difference in the
  number of A-C GCSE passes of children whose
  parents have manual occupations and children
  whose parents have professional occupations, any
  difference found isdue to chance.
• You have a go at writing one for A level grades
  and hours spent revising.

13

• The ALTERNATIVE HYPOTHESIS
• this is the prediction you make and are testing
  in your investigation.
• Sometimes called the experimental hypothesis or
  research hypothesis.
• The symbol used is H1

14

• The alternative hypothesis states that either
• There is a relationship between the two variables
  being studied
• or
• That one variable has an effect the other
• The examples given earlier about intelligence
  and social deprivation are examples of alternate
  hypotheses. Here is a reminder-

15

• There will be a positive relationship
  (correlation) between peoples IQ scores and
  their household income.
• There will be a difference in the number of A-C
  GCSE passes of children whose parents have manual
  occupations and children whose parents have
  professional occupations.

16
There are two different types of alternative
hypotheses-

• Directional (one-tailed)
• This states that one variable will affect the
  other in a specified way
• or that the relationship between the two
  variables is a particular type of relationship.
• Here are some examples-

17

• Eating 3 pieces of fruit a day will reduce the
  risk of cancer.
• Putting a list of words into categories will
  increase the number of words recalled in a recall
  test.
• IQ scores are positively correlated (related) to
  head circumference

18

• Non-directional (two-tailed)
• This states that one variable will affect the
  other but does not specify in what way
• or that there is a relationship between the two
  variables but does not say what type of
  relationship.
• For example-

19

• Eating 3 pieces of fruit a day will affect the
  risk of cancer.
• IQ scores are correlated (related to) with head
  circumference
• N.B. Remembering the difference between
  directional and non-directional hypotheses can be
  difficult imagine the hypothefish

The one tailed fish swims in one direction only
whereas the two tailed fish can go either way.
20
Ways of testing the alternative hypothesis

• Experimental methods
• True quasi experiments
• Laboratory, field natural experiments.
• Non-experimental methods
• Observation studies
• Case studies
• Content analysis
• Surveys
• Interviews
• Questionnaires

21
Experimental Methods

• Experiments are probably the most often used
  method of research.
• In the experimental method the researcher is
  specifically looking at the possible effect on
  one variable (the Dependent Variable DV) which
  might be caused by changing an other variable
  (the Independent Variable IV)
• Hence an experiment is an investigation in which
  the IV is manipulated (altered/changed) in order
  to cause a change in the DV.
• Only the experimental method can draw conclusions
  about cause and effect.

22
Non-experimental methods

• Research methods in which the the independent
  variable is not manipulate in order to see its
  effect on the dependent variable.
• These methods can not draw conclusions about
  cause and effect but can provide us with a lot
  of information about behaviour. The information
  can be used to support a theory and provide ideas
  for experimental testing.
• There are some situations where an experiment
  would not be possible, and non experimental
  methods are the only option for gathering data.

23
Lets look at Variables a little more
closelyMatch the correct term to the description

• The variable that is manipulated (or changed) to
  see if it has an effect on another variable.
• The variable that is not manipulated, but is
  simply recorded to see if it has been affected by
  the manipulated variable.

• Dependent Variable
• Independent Variable

24

• Like a hypothesis, the IV DV should be written
  in a very clear way.
• They should be written in terms of how they will
  be measured, so that someone knows exactly what
  you are counting, timing etc.
• As with hypotheses this is known as OPERATIONALLY
  DEFINING the variables.

25
Extraneous Variables

• a general term referring to any variable other
  than the IV that might have an effect on the
  measured DV. An extraneous variable may or may
  not have been allowed for and/or controlled.
• If the extraneous variables have not been
  controlled the unwanted effects they have (on the
  DV) are often known as errors.
• E.g. the distraction of unwanted noise in a room
  where a memory test is taking place.

Have a go at the worksheets on Hypotheses,
Independent variables Dependent variables
26
The Case Study Method

• A case study is an in depth study of one
  individual or of a small group of individuals or
  of an organisation. Usually the participant(s)
  has a defining feature that is of interest to the
  psychologist.
• A case study can provide insight into the
  behaviour of the individual(s).
• When carrying out a case study consent must
  always be gained from the participant before any
  data is gathered.

27

• A case study involves gathering data about
  and/or from the participant. This can be done in
  several ways by using, interviews,
  questionnaires, tests and observations on both
  the participant and friends or relatives of the
  participant.
• It may also involve looking at various written
  records on the participant e.g. medical records.
• Case studies have strengths and weaknesses just
  like any other method of research.

28
In your Research Methods Workbook fill in the
section on Case Studies use page 121 in the
textbook and the handout with the Memory Case
studies on.
29
Analysing Data from Research

• Psychologists gather qualitative and
  quantitative data.
•  
• QUALITATIVE data focuses on thoughts
  feelings about experiences expresses in words.
•  
• QUANTITIATIVE data focuses on how much there
  is of something expressed in numbers.
•  

30

• Analysing Quantitative Data
• Raw Data. This is the list of numbers you gather
  when you measure your dependent variable. It
  usually contains the individual data from each
  participant. Generally trends and differences are
  not easily identifiable from this list. You need
  to summarise the raw data to be able to do this.
  Descriptive Statistics allow you to summarise
  what you have found.

31

• Levels of Measurement
• In order to decide on the appropriate descriptive
  statistics to use on your data you must identify
  the level of measurement of the dependent
  variable.
• There are three levels of measurement that you
  need to be aware of.
• Each level gives more information about the
  dependent variable than the previous level
• Nominal
• Ordinal
• Interval/Ratio

32

• Nominal frequency counts into named categories
  e.g Tall or Short, Large or Small, Read
  more than one book in last 6 weeks or not read
  more than 1 book in last 6 weeks

Raw Data - More than 1 4, Not more than 1 1
33

• Ordinal - values on this scale represent
  rankings, ratings or placing e.g put people in
  order of their heights from shortest to tallest,
  put people in order of how many books theyve
  read in past 6 weeks

34

• Interval/Ratio these scales are more
  sophisticated and measure quantities or numbers
  of fixed units with equal distances between all
  points on the scale e.g. the actual heights
  in cm of each person. The number of hours spent
  reading.

35

• Data can also be classified as discrete or
  continuous.
• Discrete data a scale where there are only
  separated values of the variable being measured.
• E.g. number of children in KES whole children
  can only exist
• Continuous data a scale on which it is always
  (theoretically) possible to subdivide units of
  measurement
• E.g. length of boundary fence round KSE
  theoretically it would be possible to measure to
  the nearest thousandth of a cm.

36
Descriptive Statistics

• Measures of central Tendency
• Used as a summary of the sample of data using one
  typical score.
• Mean
• Median
• Mode

37

• Mean
• Arithmetical average.
• Its sensitivity is both its strength and its
  weakness it can be distorted by extreme scores.
  
• Best used on interval/ratio data, when
  distribution is normal.
• Best not used when there are extreme scores, or
  when the distribution is skewed.
• How to calculate the mean . . . .
• Add up all of the scores and divide by the total
  number of scores.

38

• Median
• The central value when scores have been placed in
  order of size.
• Its advantage is that it is not affected by
  extreme scores but it is affected by small
  changes in the number of scores, therefore is not
  useful when only a very small no. of scores are
  available.
• Best used on ordinal data or when distribution is
  skewed.
• How to calculate the median . . . .

39

• Put all scores in ascending order.
• If there are an odd no. of scores then the median
  is the middle score.
• If there are an even no. of scores then take the
  middle two scores add them together and divide
  them by 2 to get the median score.

40

• Mode
• The most frequently occurring score.
• It is immune to extremes, good for relatively
  large amounts of homogenous (similar) data.
• Best used on nominal data.
• Good when data is bi-modal (two modes).
• Not useful when there are several different
  modes.
• How to calculate the mode . . .
• Find the most common score in the set of scores.

41
Experiments

• True experiments
• The IV is manipulated by the researcher
• Participants are randomly allocated to (groups
  or) the conditions of the experiments
• Quasi experiments
• When the above are not met the experiment is
  known as a quasi experiment, example would be
• Comparing groups of people where the groups
  already exist e.g. males females

42
Laboratory experiments.

• Laboratory carried out in carefully controlled
  environments. e.g. participants are taken to a
  small classroom set up with the equipment and
  noise level kept constant in a study of memory
  recall.

43
Advantages Disadvantages of Lab.
ExperimentsIn pairs you have 5minutes to
think of as many advantages and disadvantages as
you can using a lab. experiment

44
(No Transcript)
45

• In your Research Methods workbook fill in the
  section on Laboratory Experiments

46
Doing a Laboratory Experiment

• State the alternative hypothesis to be tested.
• Identify operationally define the IV and the
  DV.
• Choose the most suitable Type of Experimental
  Design.
• Identify and implement any controls associated
  with your chosen design
• Identify the Extraneous variables that might
  affect the DV if they are not kept under control.
• Decide how to control the extraneous variables
  do this by using Standardised Procedures
  Standardised Instructions.

47
Experimental Designs

• The design of an experiment must be decided on,
  the choice will depend on
• The nature of the topic of the experiment
• The availability of resources (time,
  participants, materials, cost etc)
• There are three designs to choose from and you
  need to weigh up the pros and cons of each when
  making your decision.

48
The Three Experimental Designs are
49
Have a go at the Review worksheet on Hypotheses,
Variables Designs
50
Advantages and Disadvantages of the designs.

• Work in groups
• You have 10 minutes to think of as many
  advantages and disadvantages as you can for each
  design.

51
Repeated Measures design (RM)

• Disadvantages
• Order effects can be a problem
• It may be necessary to use different stimulus
  material
• A time lag between conditions may be necessary
• Demand Characteristics are more likely

• Advantages
• Participant variables are less of a problem
• Fewer participants are needed

52
Independent Groups design.(IG)

• Advantages
• No order effects
• Same stimulus material can be used for each group
• No time lag needed between conditions
• Demand Characteristics are less likely

• Disadvantages
• Greater problem with participant variables
• Need more participants

53
Matched Pairs design. (MP)

• Advantages
• No order effects
• No time lag is needed
• The same stimulus material can be used in each
  condition
• Demand Characteristics are less likely

• Disadvantages
• It is difficult to find well matched pairs
• How do we know which characteristics are
  important to match?
• The matching process is time consuming.

54
Controlling Extraneous Variables and dealing with
the problems of the various designs.

• What can you do about Extraneous Variables in
  general?
• Use standardised procedures (all ps do the
  experiment in the same way)
• Use standardised instructions (all ps are given
  the same instructions)
• Keep the environment the same for all ps

55

• What can you do about Participant Variables (a
  problem with IG design)?
• Random allocation to the different conditions of
  the experiment. This should equally distribute
  the differences between people across the
  conditions of the experiment.
• Use the Repeated Measures design instead of the
  independent groups design.

56

• What can you do about Order Effects (fatigue and
  boredom or practice can be a problem in the RM
  design)? There are two options-
• Counterbalancing this involves both conditions
  being experienced as both a first and a second
  task.
• E.g. Task A press buzzer in response to a light.
  Task B press buzzer in response to the sound of
  a bell.
• To counterbalance half of the ps do Task A
  followed by task B and the other half do Task B
  followed by Task A. (the ABBA design).

57

• Randomisation - this involves the different
  conditions of the experiment being presented in a
  random order to the participants.
• E.G. In the above example two pieces of paper one
  with A and one with B written on could be placed
  into a hat. Each participant could pick out one
  paper. Whatever letter is on the paper is the
  condition that participant experiences first.

58

• What can you do about
  Demand Characteristics ?
• If the ps pick up cues during the experiemnt
  they might try to guess what the study is about.
  They might then alter their behaviour to either
• please you (the experimenter) by giving you the
  results they think you want
• screw you this means deliberately trying to
  mess up your results
• act in a socially desirable way and be very
  compliant
• So, what can you do about it?

59

• Use the single blind control
• Try to make sure that the participants do not
  know what the experiment is about. This may
  involve deception. Deception raises ethical
  issues which we will consider later.
• Use the double blind control
• Sometimes the experimenter can give clues about
  what the experiment is about (remember the study
  with the dull bright rats by Rosenthal).
  It is a good idea if the person carrying out the
  study doesnt know the aim of the study as well
  as the participants. Again Deception is a problem
  here.

60
Analysing Data from Research continued

• Measures of dispersion
• These indicate how varied the scores are around
  the measure of central tendency.
• Range
• Standard Deviation

61

• Range
• The difference between largest and smallest value
  (1).
• Easy to calculate.
• Is distorted by extreme, freak scores.

62

• Standard deviation
• more complex measure which takes into account
  every score and its deviation from the mean. (The
  mean deviation from the mean).
• It gives an indication of how spread out the
  scores are around the mean value in the
  distribution.
• Best used on interval /ratio data.
• More accurate measure of dispersion than the
  range.
• Takes longer to calculate than the range.

63
 How to calculate the standard deviation .

• Draw a table with 3 columns.
• In the first column put the scores.
• Calculate the mean for the set of scores. (mean
  4.4)
• Take each score in the first column and minus the
  mean value from the score.
• In the second column put the values of the
  scores minus the mean.

64

• Square each value in the second column.
• In the third column put the squares of the
  second column.
• Add up all the values in the third column.
• Total 40.4

65
(N 1) (10-1) 9 40.4 / 8
4.4889 v4.4889 2.1187 SD 2.1187

• Calculate N-1 (N being the number of scores in
  the first column)
• Divide the result of step 8 (total of 3rd column)
  by the result of step 9 (N-1).
• Find the square root of the result of step 10,
  this is the standard deviation.

66

• Choosing an appropriate diagram, graph or chart
  to summarise and display your data.
• The choice depends on knowing whether data are
• a)     Nominal, ordinal or interval /ratio.
• b)    Discrete or continuous
• Nominal data discrete data - use bar charts,
  pie charts.
• Ordinal data may use histograms, frequency
  polygons scattergrams for correlation.
• Interval/ratio data may use histograms,
  frequency polygons scattergrams for
  correlation.

67

• Histograms Frequency Polygons

The y axis always shows the frequency that is
the number of times a particular score appears in
the data.
68

• Important notes about histograms-
• Empty values with in the range must be shown
• Columns must be the same width for the same sized
  intervals
• The width of the column indicates the size of the
  interval.
• Columns always touch each other to make a
  continuous curve (except where there are empty
  values)

69

• Important notes about frequency polygons-
• It is similar to a histogram, but instead of
  drawing a column, the mid points of the columns
  are joined with a continuous line.
• All zero frequencies must be shown within the
  data as well as at the beginning and end of the
  set of data.
• You can plot more than one set of data on the
  same polygon.

70
Bar Charts
71

• Important notes about Bar Charts-
• The y axis (column height) can represent among
  other things, totals, percentages, mean values.
• Columns are always separated by a gap(although
  sub groups can be clustered together) because
  the categories on the x axis are discrete.
• It is important that all graphs and charts are
  given an informative title and that the axes are
  clearly labelled. Bar charts should also be given
  a key describing the bars.

72
These are used for nominal discrete data. Each
segment of the chart is calculated by dividing
the frequency for that category by the total
frequency and then multiplying that value by
360. This tells you the size (in degrees) of that
section of the chart.

• Pie Charts