Psychology Notes > Intro to Psychological Research (1st year) Notes

This is an extract of our **Into To Psych Research 1st Year ** document, which
we sell as part of our **Intro to Psychological Research (1st year) Notes** collection written by the top tier of
Durham University students.

* The following is a more accessble plain text extract of the PDF sample above, taken from our Intro to Psychological Research (1st year) Notes.
Due to the challenges of extracting text from PDFs, it will have odd formatting:
*

INTRO TO PSYCHOLOGICAL RESEARCH

3 Experimental methods

1 or 2 tailed

Confounding variables 4

Validity and Reliability 5

Within and Between subjects 6-8 Experiment types

True

Field

Quasi

6 Non-equivalent

Time series

Time seires w/ non-equivalent: multiple time series

6 Single case

A-B

o 7 Multiple baseline

Alternating treatments

Simultaneous treatments 9

Non-Experimental methods

Correlations 10-12 Quantitative data

Questionnaires

10 Surveys

Observations 13

Qualitative data

Interviews

Pictures

Ethnographies

Data displays 14-15 Descriptive Stats

Function of stats

Measures of central tendency and dispersion

IQR

SD

Types of data (Nominal, Ordinal, Interval, Ratio and Cont, Discrete, Dichotomous)

16-18 Tables and Diagrams

Bar charts, histograms, freq polygons and cum freq polygons

Distributions

16 stem-and-leaf diagrams

17 box-and-whisker plots 19

Probability 20

Sign test 21

Wilcoxon Matched Pairs Signed Ranks Test 22

Mann-Whitney U test 23

Parametric tests

1 24

1 sample T test 2-matched sample T test

Independent samples (equal groups)

Independent sample (unequal groups ie pooled variances)

Correlation

Pearson's R

Spearman's Rho

Chi-Square

Chi-Square- 1 sample

Chi-Square - 2 categorical variables sample (Contingency X 2)

25 26 27 28 29 30 31

FOR A STAT TEST, WRITE OUT:

Design = IV (state general and then what levels), DV; within-Ps, between-Ps, correlation, freq/chi-sq

Question type = relationship, differences

Data type = qualitative, quantitative

Data type = NOIR, parametric, non-parametric, variances

STANDARD DEVIATION ON CALCULATOR

Mode 3: Stat 1: 1-VAR (if 1 set of data) 2: A+BX (if 2 sets of data only use if equal number data)

Enter values than press AC

Shift 1: STAT

4: VAR

SX/SY for SD of x/y values

Mode 1: Comp for normal mode

Parameter: value of population - no uncertainty

Statistic: value of a sample - has uncertainty stats used to estimate parameter

EXPERIMENTAL METHODS

2 Involves some manipulation

True experiments

Field experiments

Quasi experiments

Single case experiments

Hypothesis = prediction of what you expect to find in your study 1 tailed = direction of the rel is fixed in advance ie there is a direction +ve or -ve 2 tailed = direction not specified ie there is a diff in a yet unknown direction

Allows investigating the casual rel

Control IV

Better control of extraneous variables

Random sampling and random allocation

Increases internal validity

At risk of decreasing external and ecological validity

CONFOUNDING VARIABLES

Order effects

Performance in 2nd half may be better due to learning or worse due to tiredness

Participant Bias

Demand charac - tendency of Ps to respond in certain ways because they know they are being observed & believe that they know what the researcher wants

Volunteer Bias

When you seek volunteers, you will get a sample that is not rep of the larger pop

Experimenter Bias All stages from designing to analysing can be affected. Bias towards a result expected by experimenter

Minimising confounding variables

Effective randomization

Single blind = info that could introduce bias or skew the result is withheld from Ps, but the experimenter knows all facts experimenter bias

Double Blind = attempt to eliminate subjective bias on the part of both experimental subjects and the experimenters. The key that identifies the subjects and which group they belonged to is kept by a third party and not given to the researchers until the study is over.

Standardised procedures

Stats

VALIDITY

= measures what it claims to measure. Whether a study scientifically answers the qs it intends to answer

3 Criterion Validity

Compares the test with other measures or outcomes already held to be valid

IQ test validated against academic performance

Content Validity

Does an IQ questionnaire have items covering all areas of intelligence discussed in literature?

Construct Validity

Involves the empirical and theoretical support for the interpretation of the construct

To what extent is an IQ questionnaire actually measuring "intelligence"?

External Validity

Can results of a study be held to be true for other cases e.g. to different people, places or times?

Can findings can be validly generalized?

Ecological Validity

To what extent can research results be applied to real life situations outside of research settings?

To be ecologically valid, the methods, materials and setting of a study must approximate the reallife situation that is being studied

RELIABILITY

= a test is reliable when it gives consistent results of the same measure

Internal reliability = consistency of a measure within a test (ie all items measuring the same things)

Split half method (same participant do both halves of the test. If both provide similar results =

test has internal reliability)

External reliability = ability to replicate the results and get the same/similar results

Test-retest method (testing the same participant twice over a period of time on the same test.

Similar scores = test has external reliability)

Inter-rater reliability (comparing the ratings of 2 or more observers)

Ceiling and floor effects

Ceiling effects

Test is too easy and many Ps score near the top

Test can't distinguish between individuals

Floor effects

Test is too difficult

WITHIN-SUBJECTS (REPEATED MEASURES)

= each P takes part in each level of the IV

4

Controls for indiv diff between Ps

Reduces number of Ps needed - higher consistency

Allows to follow Ps over time

Time constraints

Exp mortality

Order effects - practice/fatigure effect

Demand charac

Solution:

o Randomize order of conditions

Counterbalancing

Counterbalancing

Balance effects or order of conditions

Split each group in half - group 1 does A then B, group 2 does B then A

Order effects balanced out as they occur equally in both groups

Complete Counterbalancing:

Requires that all possible condition orders are used

Eg if 6 levels = 6x5x4x3x2x1=720 orders

Latin Square:

Ensures that each level of the IV appears equally in each position

A B C

C A B

B C A

Matching - treated as within-subjects

Matched groups = ensures that each group is matched on a particular charac or variable eg age

Matched pairs = each member of a group has a corresponding 'pair', eg that there is an individual in another group who is the same in a variable

BETWEEN-SUBJECTS (INDEPENDENT GROUPS)

= 2+ separate groups receive different levels of the IV

Generally easier on the P

Hypotheses not guessed (P bias ie demand charac)

Avoids order effects

Indiv diff between Ps in the groups may affect results

More Ps needed

Solution: Random allocation or Matching

EXPERIMENT TYPES

True experiments

5

Sample groups must be assigned randomly

Ps must be randomly assigned to either control or experimental group

There must be a variable control group

Only one variable can be manipulated and tested

Results can be stat analysed

Much easier to replicate and validate the results

Usually gives a Y/N answer

Almost too perfect - conditions under complete control ie not rep

Difficult to exclude other factors that may affect the manipulated variable

Field experiments

Experiment in a natural setting/performed outside the lab

Situational variables are allowed to vary as they normally would eg noise, presence of other people

A manipulation is introduced

Less control - more ecologically valid

Ps may be unaware of the aims of the experimenter

Looking at a behaviour in its context is more meaningful

Less control - confounds

Normally more time consuming/expensive

Might not end up with optimal settings due to other restrictions

More difficult to replicate

Any generalization is tenuous

Quasi-experimental design

= still includes manipulation but lacks random assignment (i.e. uses pre-existing groups)

Random allocation cannot be used in all research eg looking at the effectiveness of treatments or comparing teaching methods across schools

Allows for research where it is not practical/ethical/possible to randomly assign Ps to groups

Cannot control for some diff between groups that could cause diff in DV

3 types:

Non-equivalent control group designs

Time series designs

Time series with non-equivalent control group designs

Non-equivalent control group designs

You can't assume that groups are equivalent before testing occurs

must test both groups before and after the manipulation (Pre- and Post-testing)

6 It's often likely that the groups are not equivalent, therefore NECG

Selection bias - initial advantage of one group

Selection/maturation interaction - groups mature at diff rates which creates illusion of a program effect when there is not one

Time series designs

One sample

You make a few observations (measurements of the DV) to establish a baseline, do the intervention, and then make a few more measurements

Any change could be due to something other than the treatment

Testing effects Repeated testing: are we measuring X or the skill with the instrument?

Instrumentation effects Does the measurement instrument decay and gain/loose sensitivity?

Experimental mortality Ps drop out

Time series with non-equivalent control group designs (Multiple Time Series Design)

Combination of the previous two designs

Time Series Design with the addition of a comparison group

Measures of DV taken on many occasions for an intervention group and non- equivalent control

Provides some info about what might have happened to the experimental group had the experimental treatment not been applied

Expensive and need more Ps

In longitudinal studies - Ps in the 2 groups may discuss intervention

Single Case Experimental Design

A special case of the time series design

Measurements taken repeatedly (eg 10 times) before and after an intervention on one or a few Ps

P serves as his/her own control, rather than using another individual/group 4 types:

A-B designs

Multiple baseline designs

Alternating treatments design

Simultaneous treatments design

A-B designs

A is a baseline phase during which the natural occurrence of the target behave is monitored

In B the treatment is introduced

Is it a behav change that follows the onset of treatment due to treatment or some other reason?

- Overcome by using ABA design BUT ethical issues and effects of medication persist

Multiple baseline designs

Multiple aspects of behaviour are identified and measured over time to provide baselines against which changes can be measured

7

*Buy the full version of these notes or essay plans and more in our Intro to Psychological Research (1st year) Notes.*