RMA Week 2: Non-Experimental Methods

describe 3 features of a non-experimental design

1. ‘what’ questions

2. describing what’s happening

3. test predictions

describe 3 features of an experimental design

1. ‘why’ questions

2. explanation of what’s going on

3. controlled variables

what are 3 features of a naturalistic observation? (non-experimental design)

1. mainly categorisation (observing and ticking boxes)

2. as little disturbance as possible

3. observer stays in the background

what are the two examples given for naturalistic observations?

1. feeding behaviour of fussy children, looking through a one-way mirror

2. cross-cultural eyebrow raising on greeting (ebi-eibsfeldt)

explain the reliability problem for naturalistic observers

reliability of categorisations - categories may be too ambiguous for behaviour shown. also subjective to cultural changes depending on the researcher’s background

solution to low reliability in naturalistic observations

comparing notes of more than one observer. inter-observer reliability achieved if observers have similar results

explain the reactivity problem of naturalistic observations

if participants notice they’re being observed, they may change their behaviour

solution to reactivity in naturalistic observations

participant observations - researcher’s part of the group

describe 3 features of case studies (non-experimental design)

1. observation of a single person

2. e.g. clinical (i.e. neurotic patient)

3. e.g. developmental (i.e. child with autism)

what are the two examples given for case studies?

1. freud’s oedipus and electra complexes

2. festinger, riecken, schachter: ufo cult that believed in end of world soon

name the three problems with case studies

1. generalisation - hard to generalise from one person

2. reproducibility - can it happen again?

3. cause-effects - what if the result is due to age/gender instead?

what is a solution to all three problems with case studies

similarities of two cases with one difference

name the three types of survey (non-experimental design)

1. questionnaire - e.g. five-factor personality questionnaire

2. interview - structured/unstructured

3. diary study - e.g. food diary for issues with eating

describe reactivity as a problem in surveys

participants will know their answers will be analysed, so may change answers to adhere to social norms, especially if it’s face-to-face

how would you alleviate low validity in questionnaires

use a well-tested method to increase face validity, so it measures the intended variable

how would you ensure you could compare answers across questionnaires

use similar scales

what are structured interviews (non-experimental design)

they use a fixed set of questions in a fixed order. answers will use a rating system or be multiple choice

benefits of structured interviews

1. easily quantified

2. comparability across participants as they produce quantitative data

3. all topics can be covered

costs of structured interviews

1. rigid structure

2. not adaptable to participant, cannot ask more questions to gain deeper understanding

3. surface-level information only

what are unstructured interviews (non-experimental design)

a number of topics without a fixed order or fixed questions

benefits of unstructured interviews

1. more in-depth information

2. relevent to each specific participant

costs of unstructured interviews

1. can’t generalise

2. analysis is time consuming - prior transcription and the analysis itself

advantages of descriptive research

1. sometimes it’s the only practical and ethical method of study

2. often inexpensive and flexible

3. real-life studies (e.g. support groups) need no manipulation

4. high ecological validity if in irl setting

disadvantages of descriptive research

1. researcher bias

2. reactivity

3. lack of causality

purpose of correlational research (non-experimental research, but not descriptive like the rest)

to determine the relation between two variables without manipulation

how does correlational research achieve its purpose?

by measuring two or more variables as they exist, without interference

what does correlational research do?

looks at the direction and strength of the relationship between variables

in correlational research, variables must be…

quantifiable and on some kind of scale, measurable

problems with correlarional research

1. direction of relarionship often unclear

2. confounding variables (and how to determine them)

why can correlational studies be more informative than descriptive research?

because the degree of relation is studied and predictions can be made (e.g. high blood pressure - make interventions earlier)

what can relational studies tell us?

they can identify and quantify relationships but they never have proof of causality