Research methods

Aims

Aims are general statements that describe the purpose of an investigation

Examples

  1. To investigate whether energy drinks make people more talkative

  2. To investigate whether sunshine increases levels of happiness

  3. To investigate whether a brain abnormality causes schizophrenia

What is an experiment?

An experiment involves the manipulation (changing) of an independent variable to see what effects it has on the dependent variable, while at the same time trying to control the other variable.

Variables

When conducting a simple experiment you will need to have two variables.

A variable is things which can vary or change. In psychological experiments, there are two different types of variables.

IV-manipulated or changed by researchers, not by other variables in the experiment

DV-the variable that changes as a result of the independent variable manipulation

Hypothesis

When carrying out experiments it is expected that the researcher will start a hypothesis.

The Alternate hypothesis is a testable statement which predicts how one variable will effect another. It is a statement which predicts a difference between conditions in an experiment.

The Null hypothesis says that there will be no difference between conditions and any difference will be due to chance and not from the manipulation of the IV.

directional or non-directional hypotheses

Directional hypotheses are predictions that state the direction the results will go in. For example, people who have consumed alcohol will concentrate less well than those who have not consumed alcohol

Non-directional hypotheses are predictions that do not state the direction that the results will go in. For example, there will be a significant difference in the concentration levels after consuming alcohol

Levels of the IV

In order to test the effect of the IV we need different experimental methods. The controlled condition is where there is no manipulation of the IV. The experimental condition is where you have manipulated the IV.

For example, the group that drinks energy drink (experimental condition) will be chattier than the group who does not drink an energy drink (controlled condition)

Experiments Hypotheses

Null- there will be no significant difference between the participants who [do something IV] and those who [do something else IV] when asked to complete [the task DV]

Directional- participants who [do something IV] will be significantly [faster better quicker etc] at [something DV] than participants who [do something else IV]

Non-directional- there will be a significant difference between participants [do something IV] and those who [do something else IV] when asked to complete [the task DV]

Pilot studies

A pilot study is a small-scale trial run of the actual investigation. This may only involve a small number of participants. The pilot study allows the research to identify any potential issues and modify the design or procedure, saving time and money in the long run.

Controls

Controls are used by the researchers to make sure the experiment is reliable. A control is something that is kept the same for each participant doing the experiment.

Researchers try to keep all variables (WHICH ARE NOT IV AND DVS) the same for all participants, so they do not affect the results.

The variables that need to be controlled in the research are called extraneous variables

Extraneous variables

An extraneous variable is any variable that you're not investigating that can potentially affect the outcomes of your research study.

Cofounding variables

If extraneous variables are not controlled, such as testing participants at different times of the day, in different environments ect, then these variables can become what we call confounding variables.

A cofounding variable is something that is not controlled throughout the experiment and has an effect on the result (DV)

So, if there are two conditions with caffeine (IV), and they influence the number of words recalled (DV). The extraneous variables will be the time of day, if its tea if coffee and if there’s noise.

there are two types of extraneous variables to control: Participant and Situational

Participant- gender, age, personality, intelligence

Situational- time of day, demand characteristics, light, heat

Psychologists try to control extraneous variables so that they dont become confounding variables. This means experiments can show cause and effect.

Demand characteristics

A feature of a procedure that influences a participant to try to guess what a study is about and look for clues as how to behave.

Controls for demand characteristics

  • deception- distractor questions, lying about the aim

  • single blind- the participant isn’t aware of which condition they are in

Investigator effects

The investigator effect is a term used to describe subtle cues or signals from an experimenter that affect the performance of participants in studies. The cues may be unconscious non verbal cues, such as muscular tension or gestures. They may be vocal cues such as tone and voice.

Controls for investigator effects

  • double blind- neither the researcher or the participant knows which condition they are in

  • Randomisation- using chance wherever possible to reduce the researchers influence.

Standardisation

Standardise- keep everything the same for each participant

Standardised procedure and standardised instructions.

Experimental design

This refers to the way in which participants are arranged in relation to the different experimental conditions.

  • independent group design

  • repeated measure design

  • matched pair design

Independent group design

Pps only take part in one condition of the experiment (two separate groups)

Strengths

  • no order effects (practice, fatigue, boredom) as they only take part in one condition

  • less chance of demand characteristics

limitations

  • any differences between the conditions could be due to participant variables, eg one group could do better because they are more intelligent- control by randomly allocating groups

  • less economical as twice as many participants are needed in comparison to repeated measures

Repeated measures design

Pps take part in both conditions of the experiment (one group)

Strengths

  • any differences between conditions are likely to be due to changes in the IV and not due to participant variables

  • fewer participants are needed to be recruited as they are used twice

Limitations

  • order effects may occur (e.g. practice, fatigue, boredom) as participants take part in all conditions

  • may see more demand characteristics as participants are more likely to work the aim if they take part in both conditions

Matched pair design

Pps are matched on variables relevant to the experiment, eg IQ, gender. One pps from each pair are then allocated a different condition.

Strengths

  • there are no order effects as pps take part in only one condition

  • Participant variables between conditions are reduced as participants have been matched.

Limitation

  • It is time consuming and expensive to match up participants

  • they can never be matched exactly.

What are order effects?

Carrying out a task repeatedly leads to change in performance

  • boredom effects- deterioration of performance across conditions as PPs become tired or bored

  • Practice effect- improvement across conditions through familiarity of the task or environment

This is a problem with repeated measure design

  • There can be a long gap between conditions.

  • use independent groups or matched pairs

  • counterbalanced design

Counterbalancing

Important control when using repeated measures. It reduces order effect as half pps do condition A, then B and the other half do B and A.

Types of experiments

lab experiment- Involves the manipulation of an IV and conducted under highly
controlled conditions. A laboratory is any environment where variables
can be well controlled

Field experiment- Involves the manipulation of an IV but the experiment is conducted in
‘the field’/a real world situation. Participants are often unaware that
they are being studied.

Natural experiment- The researcher takes advantage of a pre-existing IV which would
have changed even if the experimenter was not studying it.
Participants may be tested in the field or a lab.

Quasi experiment- Have an IV that is based on an existing difference between people,
e.g. age or gender. No one has manipulated this variable, it simply
exists.

Statistical testing

Levels of measurement

There are a lot of ways that variables can be measured and psychologists group measurements into one of the four scales.

  1. nominal

  2. ordinal

  3. interval

  4. ratio

The simplest level of measurement is nominal data (frequency data) then followed by Ordinal data (scores in rank order) then interval (a continuous scale with no absolute zero) and finally ratio ( a continuous scale with an absolute zero)

Qualitive data

Ordinal, nominal

Quantative data

Ratio, interval

Descriptive and inferential statistics

Descriptive statistics refers to graphs, tables, summary of data such as mean, mode and median etc, and measure of dispersion.

Inferential statistics refers to the use of statistical tests which tell a psychologist if the difference or relationship they have found is significant or not.

They help decide which hypothesis to accept and which to reject

Standard deviation

Standard deviation is a measurement of dispersion that shows that spread of scores around the mean. The greater the standard deviation the great the spread of scores around the mean.

Choosing a statistical test

Carrots Should Come With Swede Under Roast Potatoes

Probability

The accepted level of probability in psychology is p<0.05

This means that there is an equal to or less than 5% probability that the result in a particular study occurred by chance and 95% probability that the results were due to the change in IV.

In some circumstances, researchers employ a stricter significance level such as 0.01, 1%

Statistical testing

In order to make valid conclusions that can be generalised to the target population we have to:

Use statistical (inferential) tests that have been designed to work out the probability (p) of whether a particular set of data could simply have occurred by chance.

Establish whether the difference between the two groups we tested is big enough to be of significance

When to use a sign test

To determine whether the difference we have found is significant, we can use the sign test.

To use the sign test:

  1. we need to be looking for a difference than an association

  2. we need to have used a repeated measure design

  3. we need data that is organised into categories (nominal data)

How do a sign test

A study was conducted into whether people are happier after they had a holiday. Participants rated their happiness on a scale of 1 to 10, where 1 is very unhappy and 10 is very happy. They do this before and after the holiday (repeated measure design)

Step one- state the hypothesis

People will be happier after going on holiday than they were beforehand (directional hypothesis, so is a one tailed test)

If the hypothesis was non directional then a two tailed test is used.

Step two- record the data and work out the sign

On the table of results, you’ll put a + if the PP was happier after their holiday and a - if the PP was happier before their holiday.

Step three- find the calculated value

S is the symbol for the test statistic we are calculating. It is calculated by adding up the minuses and selecting the smaller value.

Step four- find the critical value of S

The calculated value now needs to be compared to a critical value to decide whether the results are significant or not.

You need to do the following instructions

  1. the significance level is desired

  2. the number of PPs in the investigation (The N value which ignores any zero values)

  3. Whether the hypothesis is directional, or non directional

Correlations

A correlation is a statistical technique that can tell us if there is a relationship between two continuous variables

Correlations can establish whether there is a relationship between two co variables.

They use Quantative data.

Correlations can be

  • positive- as the scores on one variable increase, so do the scores on the other variable

  • negative- as the scores on one variable increase, so do the scores on the other variable

  • no correlation- when there is no relationship between two variables they are said to be uncorrelated.

Correlations coefficient

This refers to a number between -1 and +1 and states how strong a correlation is. If the number is close to +1 then there is a positive correlation, but if its close to -1 then there is a negative correlation. If the number is close to 0 then the variables are uncorrelated.

  • +1.0 perfect positive correlation

  • +0.8 strong positive correlation

  • +0.5 moderate positive correlation

  • +0.3 weak positive correlation

  • +0.1 very weak positive correlation

  • 0 no correlation

  • -0.1 very weak correlation

  • -0.3 weak negative correlation

  • -0.5 moderate negative correlation

  • -0.8 strong negative correlation

  • -1.0 perfect negative correlation

Operationalising variables

You must think about how you are going to measure each variable so that you can correlate your data.

For example- Is there a relationship between stress and illness?

  • variable one- scores on a stress questionnaire

  • variable two- number of days absent from college since September

Hypotheses for correlations

The research hypothesis for correlations is called an alternate hypothesis. Hypotheses for correlations predict a relationship and not a difference.

Therefore you should never include the words “difference” “cause” or “effect” in a correlational hypothesis. (there is NO manipulation of the IV in correlation and so they can not establish a causal link)

Hypotheses for correlations must show both variables

Directional hypothesis for correlations

Say what kind of correlation is predicted, positive or negative.

For example; ‘there will be a significant positive correlation between the amount of independent study students do and their A level grades’ (the more they study the better their grades will be, and the less they study the lower their grades will be)

Non-directional

Non directional hypotheses for correlations just say that a correlation is predicted, no direction is predicted.

for example; ‘there will be a significant correlation between the number of hours paid work a week students do in term time and their a level result’

Null

The null hypothesis is the no relationship hypothesis, so for correlations the null will always say that there will be no relationship between variables

for example; ‘there will be no correlation between the amount of independent study students complete and the A level result

Correlation studies summary

  • can establish whether there is a relationship between two variables.

  • use Quantative data gathered by self report, observation, experiment or physiological measures.

  • allow us to assess the strength of that relationship and whether it is a positive or negative relationship.

  • Use descriptive stats known as scattergrams to represent the relationship between two variables.

  • are unlike experiments because it does not allow us to establish causal relationships (cause and effect)

Difference between correlations and experiments.

In an experiment the researcher manipulates the IV in order to measure the effect on the DV, meaning we can establish cause and effect.

In a correlation the variables are simply measured, no deliberate change is made, meaning no conclusion can be made about cause and effect.

Strengths of correlations

  • No manipulation of the variables so is useful when manipulation is practically or ethically impossible

  • Can be used as a starting point to an investigation and may suggest ideas of future research

Weaknesses of correlation

  • Can not determine cause and effect as only show a relationship between variables. This is because another untested variable could be causing the relationship between the two co-variables, known as intervening variables.

  • Technique is subject to any problems associated with the method used to collect data