Experimental Method: Aims and Hypotheses

Experimental Method

• The experimental method involves manipulating an independent variable(IV) to observe its effect on a dependent variable (DV).
• The DV's measurement and results are recorded.
• Experiments can be conducted in a lab, field, or occur quasi or naturally.

Aims

• An aim is a researcher's general statement defining the study's purpose.
• Aims are based on existing theories and research.

Hypotheses

• A hypothesis is a precise statement of the relationship between the variables under investigation.
• Hypotheses can be directional or non-directional.
• Directional hypotheses specify the relationship's direction, while non-directional ones do not.
• Example of a directional hypothesis: "The more sleep a participant gets, the better their memory performance."
• Example of a non-directional hypothesis: "The amount of sleep a participant gets will affect their memory performance."
• Directional hypotheses are suitable when existing research suggests a particular outcome.
• Conversely, non-directional hypotheses are appropriate when little prior research exists or findings are contradictory.

Independent and Dependent Variables

• The independent variable (IV) is manipulated or naturally changed by the researcher.
• The dependent variable (DV) is the factor measured to assess the IV's effect.
• Other variables affecting the DV should be controlled to confidently attribute changes solely to the IV.
• Testing this effect requires different conditions: an experimental condition and a control condition.
• Multiple experimental conditions enable comparisons of varying IV levels' effects.

Operationalisation of Variables

• Operationalisation involves precisely defining variables for measurement, ensuring they are measurable.
• Hypotheses should demonstrate operationalisation.
• For example, a directional hypothesis should state: "Participants who get at least four hours of sleep will show better performances on the memory test, shown by them achieving higher scores than the participants that got less than four hours of sleep."
• Investigations can be further detailed by including the test questions or the maximum achievable score for the investigation.

Control of Variables

• Extraneous variables are any variables other than the IV that affect the DV but do not vary systematically with the IV.
• Extraneous variables are labeled as nuisance variables.
• Confounding variables are variables other than the IV that affect the DV and changes systematically with the IV.
• An example from a sleep study would be time of day affecting tiredness and obscuring results; potential confounding variables should be identified and controlled, e.g., testing participants at the same time.

Demand Characteristics and Investigator Effects

• Demand characteristics are cues from the researcher or the research setting that allow participants to guess the investigation’s aim.
• Demand characteristics can cause participants to behave unnaturally; this is known as participant reactivity.
• Participants may change behavior to support the perceived aim (Please-U effect) or sabotage results (screw-U effect).
• Investigator effects are unwanted influences on the DV from the researcher's behavior, consciously or unconsciously.
• Investigator effects can be from the design of the study to participant selection, and interactions.

Randomisation and Standardisation

• To reduce the effects of extraneous or confounding variables different steps can be taken by the researcher like randomisation and standardisation.
• Randomisation uses chance to reduce investigator effects, aiding material design, condition order, and participant or material selection.
• Standardisation uses identical procedures and instructions for all participants to eliminate non-standardised instructions as potential extraneous variables.

Experimental Method Types

• Experiments can be field, laboratory, quasi, or natural.
• High degree of control is the strength for laboratory experiments because experimenters control all variables, but it has low ecological validity.
• Quasi experiments cannot randomly allocate participants so there may be confounding variables present.
• Natural experiments provides opportunities for research that would have otherwise been impossible due to reasons practical or ethical.

Sampling

• The researcher needs to decide how they select participants to take part in their investigation.
• The population is a group of people from whom the sample is drawn.
• If the sample of participants are taken from the sixth formers going to schools in London, the results of the study can only be applied for that certain group of people and not all the sixth formers in the UK.

Sampling Techniques

• Opportunity sampling recruits conveniently available participants but may not represent the broader population and it is subject to researcher bias.
• Random sampling gives all population members equal selection chances, avoiding researcher bias, but it can be time-consuming and may result in a volunteer bias.
• Systematic sampling selects every nth member from the sampling frame, which avoids researcher bias and is fairly representative.
• Stratified sampling reflects the composition of subgroups (strata) within the wider population and reduces researcher bias, but it is time consuming and may not be completely representative.
• Volunteer sampling relies on participants self-selecting which is quick but may attract a particular profile of individuals.

Experimental Design Types

• The 3 types of design include independent groups, repeated measures and matched pairs
• With an independent group design, each participant only performs one condition of the IV. This eliminates order effects, and demands characteristics, but can also lead to not having control over participants' variables.
• Repeated measure design each participant take part in all conditions of the IV. Fewer participants are needed but this can lead to order effects.
• Matched pair involves matched participants to remove aspects of demand characteristics. It also required a large potential pool of participants and knowledge of which aspects to match the participants one for the other.

Pilot Studies

• A pilot study is a small-scale version of an investigation which is done before the real investigation is undertaken.
• Pilot studies identifies potential problems and modify procedures, saving time and money.

Single-Blind and Double-Blind Procedures

• A single-blind procedure does not tell participants whether they are receiving a test or control treatment to ensure that participants do not bias the results by acting in ways they "think."
• A double-blind procedure means that neither the participants nor the experimenter knows who is receiving what treatment.
• Double blind studies are particularly useful for inhibiting bias from both demand characteristics and the placebo effect by unconsciously giving participants clues; this reduces the investigator effects.
• A control group/condition sets a baseline. This allow the results from the experimental condition to be compared for the cause of effect of the IV

Observational Techniques

• These include naturalistic, controlled, overt, and covert
• Naturalistic observations record behavior in its normal setting, increasing ecological and external validity, also increases risk of ethical ramifications and replicability.
• Controlled observations structures the environment which provides extra control.
• Overt observations are when participants are aware but this may impact natural reactions vs covert observations where participants are unaware.

Observational Designs

• The issue with observational designs is observer bias, what they expect to see influences their reporting.
• A check on the inter observer reliability of the observation can alleviate.
• Unstructured design is when continuous recorded is conducted.
• This lends itself to richer and detailed observation, but makes it harder to analyze.
• Structured design is when researcher quantifies what behaviours they are observing, and there is limited details

Time and Event Sampling

• Time sampling records behaviour within pre-established timeframe, reducing observations but may not be representative.
• Event sampling records each time a behavior is carried but there may be counting error involved and difficult judging whether the conduct matches appropriate conditions.

Correlations

• Correlation is a mathematical method for investigating the association between two variables, called co-variables, without manipulation like in experiments.
• There is only association found and DV & IV terms aren't used, so no cause-and-effect can be inferred.
• The study is not simply cause and effect like in some other areas.
• The correlation can be negative or positive depending on the linear relationship.
• A curvilinear relationship means that as one variable increases, so does the other to a point, then as the other variable increases the other begins to decrease.

Strengths/Limitations of Correlations

• Correlations can assess patterns before experimental studies but has difficulty with cause and effect.
• "The third variable problem" and misuse/misinterpretation are other limitations.

Types of Data

• Qualitative data are displayed in words (non-numerical) have richness and participant-driven insights, but can be harder to analyze the results.
• Quantitative data are that which are displayed numerically, converted to graphs that makes comparisons easy.
• Primary data is new, first-hand target the study questions, but time and effort are required.
• Secondary data is affordable and data is easily accessible. However, reliability can be an issue since original study conditions can't be reviewed.

Meta Analysis

• It's possible to increase generalisability with a more studied dataset through the use of Meta-analysis.
• This combines data and results across different studies, but carries publication bias risk when negative results are excluded.

Data Analysis - Descriptive Statistics

• Descriptive statistics use tables, graphs, and summary statistics to analyse data.

Measures of Central Tendency

• Measures that calculate an average value within a data set.
• They include mean, median and mode.
• The mean use all available values but it can be greatly effected by irregular data with few data points.
• The median use a narrower focus, the central value, but is not influenced as such by extreme scores.

Measures of Dispersion

• Dispersion measures calculate variation in datasets like range and standard deviation (SD).
• The range calculation uses extreme data points while SD offers precision by factoring in how to account for variation.

Presentation and Display of Quantitive Data

• Use tables to summarise data
• Bar charts are a way to allow quantitative differences to be seen more clearly
• Histograms allow us to compare a dataset relative to one aspect
• Line graphs show changes of values

Scattergrams

• Used to view associations between co-variables vs differences
• Each point on the graph corresponds to X and Y value for a point to show correlation

Distributions

• Normal distributions in data form a symmetrical pattern of frequency.
• Skewed distributions don't.
• Positive skew means the distribution is concentrated on the right
• Negative skew means the distribution is concentrated on the left

Peer Review

• Peer review is a way to measure scientific publication that requires submission to experts within the same field.
• Anonymity makes it possible for reviewers to be unfair, that said it is difficult to find experts in some rare fields.
• A concern exists that mainstream science actively supresses challenging or otherwise non-conforming points of view
• Fraudulent research may be long-lasting

Implications of Economic Psychological Research

• Research has economic implications because the economy needs individuals to function and contribute.
• Research in social influence causes union strikes and influence minorities with appeals to NSI
• Memory research increases safety awareness and lessens imprisonment.
• Mental health research through behavior and therapy lessens workers' absences and makes them able to return to function.
• Attachment research enables parents to care Case Studies

Case Studies

• Examination into subjects to see what the past and present implications mean as well as possible areas for further research.

Content Analysis

• Data that comes from the sources that we produce like TV adverts, which then yields insight to potential prejudices or structured beliefs.
• To generate it, identify hypothesis, make coding, and data collection as well as a report on the data's potential.
• Limitations are potential for observer and interpretative types of biases.

Levels of Measurement

• Quantitative data with different levels of meansurements: ordinal, nominal, or interval.
• Nominal focuses on categorical types of data values.
• Ordinal has ranking but less precision; a subjective opinion.
• Interval data based on numerical scales and a precisely defined size.

Reporting Psychological Investigations

• Psychologists use a writing format that is a scientific report.
• Each details all parts like report summary, intro of past research, the process and its conditions before closing

Statistical Testing

• Allows for determinations on if hypothesis can be reliably tested
• This is done by analyzing potential effects over a study
• State the hypothesis for negative/positive value associations before testing.
• Step through the analysis and potential variables by the steps of the calculations
• A test of the values allows tests of certain hypothesis where a difference is found.

Inferential tests & Statistical Analysis

• Inferential tests and statistical analysis are determined by type of data(interval, ordinal nominal) and its difference.
• Is the calculation for either test and the determined critical value for the result to be significant