Research methods - PSYB1



Research methods - PSYB1

Read the study then answer all the small mark questions

Question worth 20 marks

[pic]

1. Planning research: Aim, Hypotheses & Variables

Psychologists first aim to make predictions about a common sense explanation of behaviour. For example, “hmm, does eating fish really make someone smart like the advertisements say?”

Therefore the first thing a psychologist must come up with is an aim of her ‘what will be’ investigation:

1.1 AIM

“The researcher wanted to investigate whether eating fish could actually help improve a person’s intelligence”.

This is still quite a general statement so psychologists must make refine the aim and make it more testable, i.e. a hypothesis. Something that reads that can be tested in a laboratory experiment.

First, in order to make sure your hypothesis is EXACT is to know your Independent and Dependant variables. (See notes below hypotheses)

1.2 HYPOTHESES - There are two major types of hypothesis:

The experimental (or alternative) hypothesis. It is experimental because it predicts a RELATIONSHIP between the IV and the DV and it is alternative because it can be of two types:

Alternative Hypothesis

One – tailed Hypothesis

■ A statement that predicts there will be a difference (increase/decrease) in the DV, i.e. a specific direction

■ “Participants who eat fish at least three times a week will have higher IQ scores than participants who never eat fish”.

Two - tailed hypothesis

■ A statement that predicts there will be a difference between the IV and DV but it doesn’t say in which direction….

■ ”Participants who eat fish three times a week will have different IQ scores than participants who do not eat fish”.

1.3 The second type of hypothesis is the Null hypothesis

■ A statement that predicts no difference between the IV and the DV that will be tested in the investigation.

■ “Participants who eat fish at least three times a week will not have different scores from participants who never eat fish at all”.

1.4 Correlational hypothesis

Are used to establish a relationship between two variables. The wording of a hypothesis for correlational studies differs from previous hypotheses:

■ ‘There will be a correlation between the number of portions of fish eaten by participants and their IQ score’s. (2-tailed)

■ There will be a positive correlation between the number of portions of fish participants eat and their IQ scores. (1-tailed)

■ Can you think of another 1-Tailed hypothesis for this scenario?

1.5 Variables

A variable is any object, quality or event that can change or vary in some way. For example, aggression, intelligence, time, temperature, eye-colour, amount of alcohol or attraction etc…. All experiments have two variables:

■ Independent variable (IV): The variable that is manipulated – changed (cause)

■ Dependant variable (DV): The variable that is measured (effect)

■ What are the IV and DV in the experimental hypotheses above?

1.6 Extraneous Variables (EV)

‘Any variable other than the IV that could affect the outcome of the study (DV)’. These variables must be controlled or kept the same in both conditions….so that we can say for sure the IV was the only variable that affected the outcome.

1.7 Confounding variables

Any EV not controlled will develop into a confounding variable and these will affect the outcome. This will make it difficult to establish a cause effect relationship with certainty..

Confounding Variables can be classified as two types Subject variables: Anything to do with the individual participant, i.e. fatigue, tired, boredom, IQ, age, gender, lack of motivation etc

Situation variables: Anything to do with the situation, for example, too warm, cold, bright, dark, noise, distractions etc…

2. Populations and sampling techniques

2.1 Target Population

The group of people whose behaviour we are interested in measuring. A small group of people from that target population must be selected for investigation who will be representative of that population.

When the results of that investigation are analysed, it is possible to generalise from that small group and say these results apply to the target population.

A sample is a small group of people who are gathered to take part in the investigation. We must refer to these people as ‘participants’.

2.2 Sampling Bias

If the sample is not representative of the target population, its is said to be biased. For example, some characteristic of the sample was either over or under-represented.

For example, if a sample of 20 teachers consisted of 16 males and 4 females, the females would appear to be under-represented in that sample.

2.3 Six main methods of collecting participants – You must know at least 4. However, Self-selecting and cluster methods are also possible

Random sampling

Definition

Every member of the identified target population has an equal chance of being selected or the sample. This could be achieved by putting all names in a hat, mixing it up and picking names out.

Example

For example putting all AS year 12 students studying psychology in the UK into a database, number each person then pick random numbers.

|Strengths |Weakness |

|The sample is unbiased, so the participants should be |It can be time consuming and costly particularly if participants |

|representative of the target population and therefore we can |refues to take part. |

|generalise their results to the population. |The sample may not be representative of the original target |

| |population, for example a random sample of 20 students might end |

| |up exclusively female. |

| | |

| | |

| | |

Opportunity sampling

Definition

The sample selected consists of people who are available and willing to take part. Or imply those available at the time. This is not a random sample, because the researcher has chosen them.

Example

University lectures will use their students as participants for a study.

|Strengths |Weakness |

|It is quick, convenient and often the most economical method of |Opportunity sampling gives very unrepresentative samples and is |

|sampling. |often biased on part of the researcher who may choose subjects |

| |who will be ‘helpful’ and so again it would be difficult to |

| |generalise these findings to all people in the population. |

| | |

Systematic sampling

Definition

Means the every nth member of the target population is selected. For example every fifth person on a register.

Example

For example every 4th baby born in a maternity hospital in one week.

|Strength |Weakness |

|This sampling method is relatively unbiased since and it is only |It is less quick and convenient than opportunity sampling, for |

|those who are in the relevant position can be selected. Second |example participants do not stand an equal chance of being |

|it is faster than random sampling. |selected if the starting point of the sample is not conducted at |

| |random. |

Stratified sampling

Definition

It involves dividing the target population into important sub-categories (strata/types) and then selecting people from that stratum (group) in the proportion that they occur in the target population.

Example

For example, if a target population contained 75% women and 25% men, a sample of 100 people would contain 75 women and 25 men.

|Strengths |Weakness |

|It produces a representative sample of key groups in the target |It is time consuming, since pre-research studies are required to |

|population. |establish the necessary proportions of each group. |

METHODS OF INVESTIGATION I

3. The experimental method

3.1 Laboratory Experiment

An environment where all variables are controlled (IV and DV)…. FOR EXAMPLE Bandura’s Bo Bo dolls :

1. Indicates cause and effect

2. Easy to replicate

3. Lacks ecological validity

4. May involve demand characteristics

3.2 Field Experiment

The researcher manipulates the IV in the participant’s natural environment…. For example Pivilian train station study:

1. High ecological validity

2. Avoids demand characteristics

3. Difficult to replicate

4. Ethical problems of consent, deception, invasion of privacy etc.

3.3 Quasi-experiment

The IV is changed by natural occurrence. The researcher just records the effect on the DV. For example study in Canada where TV was introduced and their behaviour was measured before and after it was introduced.

1. High ecological validity

2. No demand characteristics

3. Difficult to infer cause and effect

4. Ethical problems of consent, deception and invasion of privacy etc

4. Experimental Design

In an exam, once you have read the study provided, you may be asked to specify what ‘experimental design’ was used by the researcher. There are three types of experimental design and all it means is ‘How participants are allocated into the different conditions of an experiment. That is how participants are randomly allocated into the experimental or control condition. You could then be asked to give an advantage or disadvantage to the design used. The three are as follows:

4.1 Independent measures design

Different participants are randomly allocated to each condition. So each participant only takes part in one condition (experimental or control).

4.2 Repeated measures design

The same participants are used in both (or more) conditions. So each participant is used in all conditions. (They first do the experimental condition, then do the control condition)

4.3 Matched pairs design

This is similar to independent measures but this time the participants who are randomly allocated to each condition are matched on certain variables i.e. age, gender, background etc.

|  |Strengths |weaknesses |

|Independent groups design (unrelated) |No order effects (such as boredom, fatigue or practice)|Can’t use same materials in both conditions |

| |Reduced demand characteristics |Some participant variables may be uncontrolled |

| | | |

|Repeated measures design (related) |Can use same materials in both conditions |Order effects (such as boredom, fatigue or |

| |Participant variables kept constant |practice) |

| |Fewer participants needed |Demand characteristics |

| | | |

|Matched pairs design |Some participant variables kept constant |Loss of one participant means loss of pair |

| |No order effects (such as boredom, fatigue or practice)|Can’t use same materials in both conditions |

| |Reduced demand characteristics |Time consuming |

5. Controls

5.1 Counterbalancing is used to prevent order effects (fatigue, boredom and practice) from distorting the results in the repeated measures design.

For example half the participants are given condition A then B and half are given condition B then A. This means the chances of boredom, fatigue or practice are halved because we can say with mire certainty that the results are due to the IV, not the practice etc…

5.2 Demand Characteristics may occur because the participant appears to be more helpful or a researcher who talks or behaves in a way that might influence the participants behaviour and so jeopardise the results. Demand characteristics can be reduced by using a single blind procedure. The participants do not know what condition they have been placed, i.e. they do not know if they are in the alcohol or no alcohol condition.

5.3 Investigators can also jeopardise the results by experimenter expectancy or observer bias. For example, experimenter expectancy is when the experimenter behaves in a way that pushes the participant to act in a certain way. The observer bias is when a researcher will ignore certain behaviours of the participants or interpret their behaviours in a certain way or even corrupt the results according to their expectations. Using a double blind procedure can reduce these problems. Neither participants nor experimenter know what conditions the participants are in or the hypothesis being tested.

6. Questionnaires

Questionnaires are written methods of gaining data from subjects that do not necessarily require the presence of a researcher. For example, a researcher might be interested in finding out about people’s thoughts and feelings about happiness. There are three types of questionnaires

|Open questions: |Closed questions |

| | |

|Can you tell me about how happy you feel right now? |Do you feel happy right now? Yes ____ No ____ |

|Likert attitude scale question: |

| |

|On a scale of 1 – 50, where 25 represents your normal level of happiness and 50 is the most happy you could feel, estimate how |

|happy you feel right now. |

|Strengths |Weaknesses |

| | |

|Collates large amounts of data very quickly and conveniently |Self report data may be biased by the motivation levels of the |

|Highly replicable and easy to score into quantitative form |participants responding |

|Provides rich detailed data - new insights into further research|Social desirability may affect the responses (ticking yes to all |

| |the questions to be viewed in a socially accepting manner) |

| |Poor response rate |

| | |

7. Interviews

These essentially gather the same information as questionnaires but are conducted face to face. They may be used if the researcher has more time available and or wants to obtain more personalised views.

There are two main types of interviews

7.1 Structured Interviews

Used more in research (market research) and uses a standard set of fixed predetermined questions with certain ways of replying (yes/no). This means data from all the participants can be collated and summarised statistically as they were all asked the same questions.

7.2 Unstructured interviews

Used more in therapeutic situations (psychodynamic therapy) and will contain a topic area for discussion but will not have any fixed questions or ways of answering.

The strengths and weaknesses of unstructured interviews

|Strengths |Weaknesses |

|Provides rich detailed data |Self report - biased |

|Extremely flexible |Experimenter expectancy - Demand characteristics |

|Participants input is unrestricted and so new ideas can be |Difficult to quantify and analyse |

|explored |Not replicable, generalisable |

|Clarification of issues can be explored | |

Now try to think of the strengths and weaknesses for structured interviews!!!

|Strengths |Weaknesses |

| | |

| | |

| | |

| | |

| | |

8. Correlations

8.1 Correlations are used to see if there is a pattern or relationship between two sets of variables. It aims to find a cause and effect relationship BUT IT DOES NOT ESTABLISH A CAUSE AND EFFECT RELATIONSHIP in the same way an experiment can…..

None of the variables have actually been manipulated in any way, but are simply recorded. Correlations investigate naturally occurring phenomena and situations that would be unethical to create in an experiment. For example, the relationship between TV and violence.

The hours of violent TV would be recorded as one variable and the number of violent acts would then be recorded as the second variable. A visual description of these measurements would be displayed on a scattergram.

A strong positive correlation is where the score for one variable increases, so does the other. For example hours of study, better the grade on a test.

A strong negative correlation, is where the score for one variable decreases, the other variable increases. The older the car, the cheaper its value for sale.

Copy a zero correlation: This is where there is no relationship between the two variables, as one increases the other neither decreases or increases. For example the colour of hair and intelligence.

8.2 Please note – you may be asked to either interpret the data on the graph or to actually draw a graph. Label axes clearly and be as accurate as possible.

For example, a scattergram to show the table to show the time taken to complete a card sorting task according to age.

|Strengths |Weaknesses |

| | |

|No manipulation of variables/behaviour is required |No cause and effect relationship can be inferred – relationship |

| |may be due to other variables |

|Strong correlations can suggest ideas for future experiments to | |

|determine a cause and effect relationship | |

| | |

9. Observational studies

| |Participant observation |Non-participant |Overt |Covert |

| | |observation | | |

|Definition |The experimenter becomes |The experimenter watches|Overt means obvious – |Covert means concealed – |

| |part of the observation. |the behaviours from |the participants know |the researcher does not |

| | |afar. |they are being watched |tell the participants she |

| | | | |is a researcher |

| | |High ecological validity| | |

|Advantages | |if done in a natural |Can give an insight |High ecological validity if|

| |High ecological validity |setting |into future areas of |done in a natural setting |

| |if done in a natural | |study | |

| |setting | | |Can give an insight into |

| | |Can give an insight into|Can provide richer data|future areas of study |

| | |future areas of study |than experiments | |

| |Can give an insight into | | |Can provide richer data |

| |future areas of study |Can provide richer data | |than experiments |

| | |than experiments | | |

| |Can provide richer data | | | |

| |than experiments | | | |

| | | | | |

| | | | | |

|Disadvantages |No control of variables so|Open to observer bias |Demand characteristics |Ethical problems – invasion|

| |difficult to determine | |if participants know |of privacy |

| |clear cause and effect | |that they are being | |

| | | |watched | |

| |Open to observer bias | | | |

| | | | | |

Data recording techniques in observational studies

A category system is used to quantify behaviour. If I wanted to decide on what exactly is ‘greeting behaviour at airports’ why would I first perform a pilot study? Can you think of any styles of greeting behaviours you might see at an airport?

For example:

|Flowers |Kissing |Hugging |Back-slapping |

| | | |II |

|IIIII IIIII |IIII |IIII | |

|III | |IIII | |

| | |IIII | |

| | |IIII | |

| | | | |

Observer bias – the researchers expectations may lead to mis-interpretations – so to ensure reliability (i.e. measure consistently what supposed to measure) researchers should work in pairs. They would decide before hand what exactly back slapping constitutes as etc…

This means they have high inter-observer reliability

What is being observed?

■ Continuous recording – all target behaviours are recorded.

■ Time sampling – The observation time is split up into intervals, i.e. every 20 seconds…

What type of recording method would be best for the scenarios below?

Play behaviour

Aggressive language at a football match

10. Quantitative vs. Qualitative data

Quantitative data is in the form of numbers/charts, i.e. laboratory experiments, structured interviews etc produce this kind of information.

Qualitative data is in the form of written description i.e. case studies, unstructured interviews etc that produce lots of written information

Some argue human behaviour should be expressed in the form of numbers so that we can scientifically establish a cause effect relationship. However, qualitative research provides rich detailed data which is also helpful when making conclusions. Usually combinations of both methods are used.

11. Pilot study

A pilot study is a small practice scale study conducted before the actual main study/experiment. Its aim is to check and make sure everything involved in the study has been appropriately dealt with, including control of all extraneous variables. For example it will discover any ambiguous questions, anything that might be offensive, too difficult to understand, check how long it takes to do the experiment etc…

12. Case Study

A case study is a detailed in-depth study of an individual/small group or organisation. Case studies have many roles. The information gathered will then be used to find new themes/ideas/patterns and these will then be used to develop new theories. It can be used to support/refute previous theories. Or it could be used to decide which therapy is appropriate for an individual. Little Hans – is a famous case study.

|Strengths |Weaknesses |

| | |

|Produces rich detailed information |Difficult to generalise |

|May produce new insights and ideas |Retrospective data may be unreliable |

|Is one person centred |Too subjective |

| | |

Content analysis

The investigation of communication through content. Content of media, advertisements, TV, movies, magazines, newspapers, diaries etc…

Evaluations:

■ It enables an effective analysis of a wide range of information

■ No ethical problems of deception as there are no participants as such.

■ The findings are limited by the researchers expectations as categories beforehand

■ Although the researchers are trained their interpretations may still be subjective

13. Representing data

This section introduces descriptive statistics and graphs to display and summarise the results of an investigation. This will show the reader any patterns or differences that may be present.

When data is in numerical form the researcher must convert that raw data into a statistical result which will summarise the results. The summary of or descriptive statistics commonly used are called measures of central tendency and measures of dispersion.

13.1 Measures of central tendency

These provide a single value to describe a set of raw scores:

|Measure of central tendency |Advantages |Disadvantages |

|Mean (add all the scores then divide by the|It is a sensitive statistic – using all the|If one score is extremely high or low it |

|number of scores) |data |may distort the mean value |

|Median (put all scores in order from lowest|Is not distorted by extreme values |Can be distorted by small samples and is |

|to highest, then identify the middle value.| |less sensitive |

| |Is an actual score | |

|Mode (score which occurs most often) |Not influenced by extreme scores |Does not use all scores |

13.2 Measures of dispersion

Sometimes the mean, median and mode may be the same for both conditions, which would suggest there are no differences between those two sets of scores. However looking at the raw data one could easily see that in the first condition the values were very similar to each other and in the second condition, the values were very spread out. In order to summarise these sets of data, we need a statistic that displays this difference in spread or dispersion.

|Measures of dispersion |Advantage |Disadvantage |

|Range (The largest score minus the lowest |Simple to calculate |Distorted by extremely high or low scores |

|score +1) | | |

|Standard Deviation (the difference between |A sensitive statistic - uses all the data |Time consuming |

|each score in a condition and the mean |available | |

|value for each condition) | | |

Ethical guidelines

In Britain the BPS, British Psychological Society (1993) published the ‘Ethical principles for Conducting research with Human participants’ which guides psychologists to consider the implications of their research for the participants concerned.

Make notes on the following ethical guidelines:

|Informed consent |Deception |

|Psychologists must get full consent from their participants |Participants must not be deceived about the nature of the |

|before starting their experiment. This can be verbal or written.|experiment. However sometimes it is necessary so that genuine |

|If under 16yrs o age consent must be gained from a guardian. |behaviour can be observed. So long as the experimenter can argue|

| |the ends justify the means. |

|Debriefing |Withdrawal from the investigation |

|Participants must be told the full purpose of the experiment at |Participants should be told from the very start they have the |

|the end and have time to ask any questions at the end of the |right to leave at any time if they do not feel comfortable with |

|experiment. |the experimental conditions. |

| | |

| | |

|Confidentiality |Protection from harm |

|Participants should be told that their results will be kept in |All participants should leave the experiment in the exact same |

|the strictest of confidence and can be destroyed at their |physical and mental state in which they entered the experiment. |

|convenience. | |

Correlations

Correlations are used to see if there is a pattern or relationship between two sets of variables. It aims to find a cause and effect relationship BUT IT DOES NOT ESTABLISH A CAUSE AND EFFECT RELATIONSHIP in the same way an experiment can…..

None of the variables have actually been manipulated in any way, but are simply recorded. Correlations investigate naturally occurring phenomena and situations that would be unethical to create in an experiment. For example, the relationship between TV and violence.

The hours of violent TV would be recorded as one variable and the number of violent acts would then be recorded as the second variable. A visual description of these measurements would be displayed on a scattergram.

Copy a strong positive correlation:

As the score for one variable increases,

so does the other.

Copy a strong negative correlation:

As the score for one variable decreases,

the other variable increases.

Copy a zero correlation:

This is where there is no relationship

between the two variables, as one increases

the other neither decreases or increases.

For a correlation coefficient of +1 every rise in variable A is reflected in a rise in variable B

For a correlation coefficient of -1 every rise in variable A is reflected in a fall in variable B

For a correlation coefficient of 0, there is no relationship between variable A and variable B

Copy diagram from board…..

In the example earlier, there may be a strong positive correlation between TV and violent acts, giving a coefficient of 0.95, but this does not mean watching TV makes people violent. The hypothesis needs to worded in such a way as to not indicate a cause and effect relationship.

Two tailed alternative (correlational hypothesis)

There will be a correlation between the hours of TV watched and the number of violent acts committed by children aged 6-10 years.

One tailed alternative (correlational hypothesis)

There will be a positive correlation between the hours of TV watched and the number of violent acts committed by children aged 6-10 years.

|Strengths |Weaknesses |

| | |

|No manipulation of variables/behaviour is required |No cause and effect relationship can be inferred – relationship |

| |may be due to other variables |

|Strong correlations can suggest ideas for future experiments to | |

|determine a cause and effect relationship | |

| | |

| | |

1. Draw a scattergraph to display the data from study C of the life events and stress study. (3 marks)

Label a title for the scattergram

Label both Axes

Approximately give the correct scale and location of the data points on the scale

2. Identify the type of correlation shown in your scattergram and state what it shows about the relationship found between stress and illness. (2 marks)

3. State one strength and one limitation of the correlational method. (2 marks)

4. State a hypothesis for this study (2 marks)

Representing data

This section introduces descriptive statistics and graphs to display and summarise the results of an investigation. This will show the reader any patterns or differences that may be present.

When data is in numerical form the researcher must convert that raw data into a statistical result which will summarise the results. The summary of or descriptive statistics commonly used are called measures of central tendency and measures of dispersion.

Measures of central tendency

These provide a single value to describe a set of raw scores:

|Measure of central tendency |Advantages |Disadvantages |

|Mean (add all the scores then divide by the|It is a sensitive statistic – using all the|If one score is extremely high or low it |

|number of scores) |data |may distort the mean value |

|Median (put all scores in order from lowest|Is not distorted by extreme values |Can be distorted by small samples and is |

|to highest, then identify the middle value.|Is an actual score |less sensitive |

|Mode (score which occurs most often) |Not influenced by extreme scores |Does not use all scores |

Activity

Using the data from the reaction times and alcohol study, calculate the mean, median and mode for both conditions. Draw a table to show these averages with an accurate and detailed title below. Use the table below to help the layout your table…..(3 marks)

| | | |

| |Condition A |Condition B |

| |Memory rehearsal |Mental image |

| | | |

|Mean |14.67 |9.13 |

| | | |

|Median |15 |9.5 |

| | | |

|Mode |16 |10 |

Graphical displays – see handout

Activity

Draw a bar chart to display the data from the alcohol and reaction times study. Use phrases from the stimulus material to make sure you label the axes and give an accurate title:

| |

|A bar chart to show the mean ……. |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

| |

Past exam questions

1. What might be the psychologist’s interpretation of the data shown in the bar chart

(2 marks)

2. State one limitation of using the mean as a measure of central tendency. (1 mark)

Measures of dispersion

Sometimes the mean, median and mode may be the same for both conditions, which would suggest there are no differences between those two sets of scores. However looking at the raw data one could easily see that in the first condition the values were very similar to each other and in the second condition, the values were very spread out. In order to summarise these sets of data, we need a statistic that displays this difference in spread or dispersion.

|Measures of dispersion |Advantage |Disadvantage |

|Range (The largest score minus the lowest |Simple to calculate |Distorted by extremely high or low scores |

|score +1) | | |

|Standard Deviation (the difference between |A sensitive statistic - uses all the data |Time consuming |

|each score in a condition and the mean |available | |

|value for each condition) | | |

[pic]

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download