Research Methods

Aim =

A general statement about the purpose of the investigation

• starts with ‘To investigate…’

Hypothesis =

A precise testable statement about the outcome of an investigation

• Must operationalise

Directional Hypothesis

= That predicts a change AND the direction the results are expected to go

• ‘One-tailed’

Non-Directional Hypothesis

= That predicts a change but NOT the direction of the change.

• starts with “There will be a difference…”

• ‘Two-tailed’

When do researchers choose directional hypothesis?

• Researchers choose directional when previous research suggests a particular outcome 

When do researchers choose non-directional hypothesis?

• Researchers choose non-directional when there is no previous research, or findings from earlier studies are contradictory.

Null Hypothesis

= A statement of no difference or no relationship which the researcher will try to disprove

• starts with “There will be no difference…”

Independent Variable

= The variable that the experimenter manipulates and controls

• What you change

Dependent Variable

= The variable that alters as a consequence of the IV being manipulated

• What you measure

Control Variable

= Variables that are held constant or regulated by the researcher

• Variables you keep the same

Operationalise

= Making variables (both IV & DV) measurable, defined and specific that can easily be tested 

‎

• e.g. ‘Problem solving ability’ becomes ‘Time taken to complete a puzzle’

Extraneous Variables =

= Any variables other than the independent variable that MIGHT affect the dependent variable if its not controlled

• Do not vary systematically between the different conditions

Confounding Variables =

• A type of extraneous variable but varies systematically between the different conditions

• Changes in the dependent variable may be due to confounding variables rather than the IV

• DEFINITELY affects DV

Kinds of Extraneous Variables

Order Effects (Confounding) = How the positioning of tasks influences the outcome, particularly in the second condition.

• e.g. Practice, Boredom, Fatigue

‎

Participant Variables = Individual differences between participants

• e.g. Age, Gender, Personality, Intelligence, Motivation, Concentration, Experience in task

‎

Situational Variables = Features of the experimental situation.

• e.g. Noise, Weather, Temperature, Time of day, Instructions

‎

Investigator Effect = Investigator’s behaviour (conscious or unconscious) that may affect the DV.

• e.g. Design of the study, selection of participants, interaction with the participants

Mundane Realism =

• The extent to which a study mirrors the real world. 

• How realistic/similar a study (procedures & environment) is to everyday life to reflect normal behaviour.

Types of Participant Reactivity that weakens internal validity

• Demand characteristics

• Social Desirability Bias

• Hawthorne Effect

Demand Characteristics =

= Any cue from the researcher or situation that may be interpreted by participants as revealing the purpose of the investigation → Leads to participants changing their behaviour (extraneous variable)

• ‘please-U’ effect: Participants act in a way they think is expected to please the experimenter

• ‘screw-U’ effect: Participants deliberately under-perform to sabotage results

Sources of Demand Characteristics

• Setting

• Communication during study

• The way the participant is asked to volunteer

• Type of person the researcher is

Social Desirability Bias =

Distortion in the way people answer questions - they tend to answer questions in such a way that presents themselves in a better light, which may be dishonest.

• occurs mainly in questionnaires

Hawthorne Effect =

When participants become aware that they are being observed and in response, change their behaviour.

Ways to deal with Extraneous Variables

Randomisation = The use of chance methods to reduce researcher’s unconscious biases when designing an investigation.

• reduces investigator effect

‎

Standardisation = Making procedures the same for all participants in order to control situational variables, make the study replicable, and reduce investigator effects.

‎

Single Blind Design = Participants don’t know the true aims of the research

• reduces demand characteristics

‎

Double Blind Design =  Both participants and person conducting the study don’t know the true aims of the research 

• reduces demand characteristics, researcher bias/investigator effect

‎

Experimental Realism = Make the task sufficiently engaging so the participant pays attention to the task and not that they are being observed

• reduces hawthornen effect, demand characteristics

‎

Pilot Studies = Small-scale trial run of a research design before conducting the real full-scale study to test its effectiveness and make improvements on certain aspects of the design that don’t work.

• can reduce hawthorne effect, social desirability bias, demand characteristics, investigator effect

Pilot Studies =

=  Small-scale trial run of a research design before conducting the real full-scale study to test its effectiveness and make improvements on certain aspects of the design that don’t work.

Aims/Benefits of Pilot Studies

• Identify any potential issues early and modify the design of the procedure

• Saves time and money in long run

What to check/test during Pilot Studies

• Procedure - whether it’s effective

• Instructions - whether it’s too complicated/whether it’s standardised/whether any vital steps are left out

• Validity of measure - whether it measures what it’s supposed to

4 Types of Experiments

• Laboratory

• Field

• Natural

• Quasi

Laboratory Experiment =

IV manipulated by researcher in highly controlled environments

Strengths of Laboratory Experiment

• High internal validity - High control over confounding and extraneous variables - can ensure any effect observed on DV is caused by manipulation of IV (causal relationship)

• Replicable

Limitations of Laboratory Experiment

• Low external validity - lacks generalisability as artificial environment cannot be applied to everyday life

• Participants know they are being tested on in a lab experiment (may show demand characteristics, hawthorne effect)

• Tasks don’t represent everyday life - low mundane realism

Field Experiment =

IV manipulated by researcher in a natural, more everyday setting

Strengths of Field Experiment

• High mundane realism - the study’s environment represents everyday life so more likely to reflect normal behaviour.

Limitations of Field Experiment

• Loss of control of confounding & extraneous variables

• Ethical issues - lack of consent

Natural Experiment =

= IV is NOT manipulated by the researcher - no control over IV, someone/something else (naturally) causes IV to vary

• + Conducted in natural/real-life environment

Strengths of Natural Experiment

• High external validity - involves the study of real-world issues (e.g. effect of natural disasters on stress levels), so findings more relevant to real-life experiences

• Allows research where IV can’t be manipulated for ethical or practical reasons

• High mundane realism - the study’s environment represents everyday life so more likely to reflect normal behaviour.

Limitations of Natural Experiment

• Participants are not randomly allocated to experimental conditions - confounding variables not controlled so less sure whether IV affected DV

• Reduced opportunities for research - natural events may only occur rarely

Quasi Experiment =

= IV based on pre-existing difference between people - no one has manipulated it

• (e.g. age, gender)

Strengths of Quasi Experiment

• Highly controlled conditions (extraneous & confounding variables)

• Replication

Limitations of Quasi Experiment

• Participants are not randomly allocated to conditions - confounding variables not controlled so less sure whether IV affected DV

• Can’t for certain establish causal relationships - researcher doesn’t manipulate IV so can’t be certain that any change in DV was due to IV

Control Group =

= A group used in an experiment where everything is kept constant (no manipulation/change in IV).

• Used to compare the results of the experimental group to the control group.

Types of Experimental Designs

• Independent Groups

• Repeated Measures

• Matched Pairs

Independent Groups Design

• When more than one group is used

• Participants are allocated (usually randomly) to one of the groups 

‎

• Each group is exposed to a different level of the IV (different conditions):
  - One group does condition A (e.g. task with TV on)
  - One group does condition B (e.g. task with no TV)

• The performance (DV) of the groups are then compared

Strengths of Independent Groups Design

• No order effects - as different participants do each condition, participants only do one condition

• Less chance of demand characteristics as each participant only does one condition - less likely to guess aims of study

• Saves time - no need to leave a gap between conditions as there’s different participants for each condition

  
  

Limitations of Independent Groups Design (+ ways of dealing with them)

• The participants who occupy the different groups have different characteristics and are not comparable → participant variables such as age - extraneous and confounding variables - reduced internal validity

  • Can be dealt with by random allocation to each condition

• Need more participants (twice as many for 2 conditions) - more expensive

Repeated Measures Design =

• All participants are exposed to all levels of the IV (participate in every condition)

• They do BOTH condition A and condition B then the results are compared

‎

e.g.

• Participant eats chocolate cake (condition A) then gives it a taste score

• Same participant then eats lemon cake (condition B) then gives it a taste score

• Taste scores from each condition are compared

Strengths of Repeated Measures Design

• No participant variables (individual differences) between conditions - so participant variables do NOT act as confounding variables

• Fewer participants needed to produce same amount of data as independent groups - less time and money spent recruiting (cheaper)

Limitations of Repeated Measures Design (+ ways of dealing with them)

• Order effects - pp’s may perform better/worse in second condition due to boredom, practice, fatigue

  • Counterbalancing

• Demand characteristics - pp’s more likely to guess aim of study as do multiple conditions

  • Single & double blind design

• One condition may be more difficult - so pp’s may do better in one condition because it was easier (extraneous variable) rather than because of the IV

  • Standardisation

• Takes more time than independent groups because a gap may be needed between conditions to counter order effects

Counterbalancing (definition + procedure)

= Order effects are equally distributed across conditions so they can’t become confounding variables (each condition is tested first or second in equal amounts)

‎
Way 1 (AB or BA):
- Divide participants into two groups:
- Group 1 - does condition A, then condition B
- Group 2 - does condition B, then condition A

‎

Way 2 (ABBA):
- All participants take part in each condition twice
- They do condition A, then B, then B again, then A again
- Scores in condition A are added and averaged
- Scores in condition B are added and averaged

Matched Pairs Design =

• Pairs of participants are matched on some participant variables (e.g. IQ, gender) that may affect the DV

• They are ranked in order and top 2 are a pair, the next 2 are a pair and so on

• Then one member of the pair is randomly allocated to Condition A and the other to Condition B

• (Each participant only does one condition)

Strengths of Matched Pairs Design

• Some participant variables are controlled - there should be less chance of participant variables affecting results

• No order effects - as different participants do each condition, participants only do one condition

• Less chance of demand characteristics as each participant only does one condition - less likely to guess aims of study

Limitations of Matched Pairs Design (+ ways to deal with them)

• Impossible to match all variables between participants - some unmatched variables might vitally important in affecting the DV

  • Conduct a pilot study to identify key variables that may be important when matching

• Time consuming to match 

• Need more participants than repeated measures - more expensive & time consuming to recruit

• If one part of the pair withdraws from the study, the whole pair’s data is lost

Target Population =

Group of individuals a researcher is interested in

Sample =

Smaller group of people from target population that is studied

• Ideally the sample will be representative of the target population so that generalisations can be made. 

Sampling Frame =

List of people in target population from which the sample is taken 

Random Sampling Procedure

• A complete list of all members of the target population is obtained.

• All the names on the list are assigned a number

• The sample is generated using a random number generator selecting the people who’s number is chosen
  OR if a researcher is using this method for random allocation of pp’s to groups/conditions, first number generated goes into condition A, second number generated goes into condition B and so on.

Strengths of Random Sampling

• Unbiased sample - all members of the population have equal chance of selection (leads to representative sample)

• Free from researcher bias - researcher has no influence over who is selected

Limitations of Random Sampling

• Sample can still become biased if some participants decline to take part

• Takes more time and effort - because a complete list of target pop. (sampling frame)is required  then contact all who are selected 

• May be difficult to obtain a complete list of target pop. (sampling frame)

Systematic Sampling Procedure

• A sampling frame is produced

• A sampling system is nominated (e.g. every 3rd person or every 5th person)

• Begin from a randomly determined start point

• Researcher then works through sampling frame until the sample is complete

Strengths of Systematic Sampling

• Objective (free from researcher bias) - once system for selection is established, researcher has no influence over who is chosen

  
  

Limitations of Systematic Sampling

• Sample could become biased if some participants decline to take part

• Takes more time and effort - because a complete list of target pop. (sampling frame) is required  then contact all who are selected 

Stratified Sampling Procedure

• Strata (subgroups) within a population are identified (e.g. age groups: 10-12, 13-15 yrs etc.)

• The proportions needed for the sample to be representative of the pop. are worked out

• Participants that make up each stratum (subgroup)  are selected using random sampling

Strengths of Stratified Sampling

• Generalisability of findings more likely as it produces a representative sample because it reflects the proportional composition of the population

• Avoids researcher bias as once the target population has been sub-divided into strata, the participants are randomly selected

Limitations of Stratified Sampling

• Complete representation of the target pop. is impossible

• Very time consuming to identify subgroups, then randomly select participants and contact them.

• Can’t reflect all the ways that people are different 

Opportunity Sampling Procedure

• Researcher asks anyone that happens to be willing and available at the time of the study

Strengths of Opportunity Sampling

• Less costly (cheap)

• Convenient

• Less time taken to create your sample

Limitations of Opportunity Sampling

• Researcher bias - researcher has complete control over the selection of participants

• Unrepresentative of the target pop. as sample taken from specific area & small part of pop. - so findings can’t be generalised

Volunteer Sampling Procedure

• Researcher may place an advert in a newspaper or on a noticeboard or on the internet

• Volunteer sample selected by participants self-selecting themselves to be a part of the research 

Strengths of Volunteer Sampling

• Less time consuming - as participants come to researcher to volunteer

• Participants are more engaged - as they selected themselves so are willing to participate

Limitations of Volunteer Sampling

• Volunteer bias - may attract a certain ‘profile’ of people who are curious or motivated and more likely to try please the researcher

  • so generalisability is limited

6 Ethical Issues

Informed Consent - Participants should be aware of the aims of the study, what is going to happen to them (procedures), know their rights: right to withdraw & what their data is used for

‎

Right to Withdraw - Participants should be aware that they can leave a study at any time, and can withdraw their data after the study has finished

‎
Deception - Participants are not told the true aims of a study (e.g. what participation will involve), so cannot give truly informed consent

‎

Protection from Harm - Participants should not experience excessive physical or psychological harm (such as injury, embarrassment)

‎

Confidentiality - Participants kept anonymous and any personal data protected

‎

Privacy - Shouldn’t invade people’s personal spaces

Ways of dealing with Lack of Informed Consent

• Get participants to sign an informed consent form - which contains all the details of the study, like the purpose of the research and their role in it & right to withdraw, so pps can decide whether they wish to participate

• Presumptive consent - Ask a similar group if they would participate. If this group agrees, consent of original group is presumed.

• Prior general consent - Participants agree to be deceived without knowing how they will be deceived.

Ways of dealing with Right to Withdraw

• Participants should be told at the beginning of a study that they have the right to withdraw

• Written in the informed consent document

• At end of study, allow participants an option to withdraw their data

Ways of dealing with Deception

• Debriefing - After the study, all details of a study should be explained to participants, such as true aims

• Participants given the right to withdraw their data from the study afterwards

• Participants should be reassured that their behaviour was typical/normal

• Researchers should offer counselling for participants who have been subjected to stress or embarrassment

Ways of dealing with Protection from Harm

• Right to withdraw from study at any time

• At end of study, participants should be reassured that their behaviour was typical/normal in debrief

• Full debrief at end to make aware of true aims of study

• Researcher provides counselling for participants 

Ways of dealing with Confidentiality

• Protect any personal details collected

• Maintain anonymity by not recording any personal data/details of participants (don’t record names), but using pseudonyms/initials/numbers.

• Debrief reminding participants that personal data will be protected and not shared with others

Types of Data

• Quantitative vs Qualitative

• Primary vs Secondary

• Nominal vs Ordinal vs Interval

Quantitative Data =

Data that occurs in numerical form, i.e. measured in numbers or quantities

Strengths of Quantitative Data

• Simple & easy to analyse statistically - therefore comparisons/conclusions between groups can be easily drawn

• Data in numerical form tends to be more objective and less open to bias

• Graphs/charts can be easily drawn and averages can be calculated

Limitations of Quantitative Data

• Much narrower in scope and meaning than qualitative data

• May oversimplify and fail to represent reality - for example, a questionnaire with closed questions may force people to tick answers that don’t represent their feelings - therefore conclusions may be meaningless

Qualitative Data =

= Data that is expressed in language/words and is descriptive

• Can’t be counted but can be turned into quantitative data by placing the data in categories then counting frequency

Strengths of Qualitative Data

• Info much richer in detail & allows participants to develop their thoughts & feelings - greater external validity

• Can provide unexpected insights into thoughts, feelings and behaviours

Limitations of Qualitative Data

• More subjective and more open to bias

• More difficult to analyse, therefore comparisons/conclusions between groups are harder to be drawn

Primary Data =

Original, first-hand data collected specifically for the purpose/aim of the investigation by the researcher.

Strengths of Primary Data

• Relevant data for investigation - researcher can design data collection procedures to fit aims of the particular study

  
  

Limitations of Primary Data

Time consuming, expensive & takes effort

Secondary Data =

Second-hand data; Data collected by someone else other than the researcher and not initially collected specifically for the aim of the study.

Strengths of Secondary Data

Simpler, cheaper, less time consuming, & easily accessible

Limitations of Secondary Data

• Data may be of poor quality, outdated or incomplete

• May not quite fit the needs of the study (irrelevant data)

Strengths of Mean

• Most sensitive because it uses all the values when making the calculation - therefore more representative of the data as a whole

Limitations of Mean

• Easily distorted by extreme values - may end up unrepresentative of the data as a whole

• Most likely to lead to a score which is not an actual score in the data set / doesn’t make sense for discrete data
  (e.g. average number of shoes owned)

Strengths of Median

• Not affected by extreme scores

• Easy to calculate

Limitations of Median

• Less sensitive than mean as not all scores are used in the calculation - less representative of whole data set as a whole

Mode =

Most frequently occuring value

Strengths of Mode

• Only measure to use when data is nominal (in categories)

• Less prone to distortion by extreme values

• Easy to calculate

Limitations of Mode

• Less sensitive than mean as not all scores are used in the calculation - less representative of whole data set as a whole

• Data can end up with multiple modes or no modes - so not useful for those data sets

Strengths of Range

• Easy to calculate

Limitations of Range

• More affected by extreme values as only takes into account two most extreme scores - more affected by anomalous data - may be unrepresentative of whole data set

• Fails to take account of the distribution of numbers - whether most numbers are closely grouped around the mean or spread out evenly 

Standard Deviation =

Measures how much scores deviate from the mean, amount of dispersion/spread/variability of data around mean

• The larger the SD, the greater the dispersion within a data set.
  (e.g. larger SD suggests that not all participants were affected by the IV in the same way, results are inconsistent)

Strengths of Standard Deviation

• More precise because it includes all values within its calculation

• More representative than range because it includes all values within its calculation

Limitations of Standard Deviation

• Can be distorted by extreme values (but much less so than the range)

Tables (when to use & what it is)

• Contains descriptive statistics (measures of central tendency and dispersion)

• Summary paragraph beneath the table explaining the results/numbers and drawing conclusions 

Bar Charts (when to use & what it is)

• Used for nominal (discrete, categorical) data

• Bars are separate & not touching each other

• Can easily visually see differences in values to make comparisons