The following items should be included in this section: The Hypothetico-deductive (scientific)method, types of psychological research methods, research designs, sampling, reliability, validity, and triangulation.
Research into mind can be traced back to Ancient Greece. However, empirical psychological research has its roots in investigations into cognitive functions such as introspection and memory. While early psychological researchers attempted to bring the same standards of rigor and control to their investigations as physical scientists enjoy, psychological research poses unique obstacles. Psychological research investigates mind. Only recently the contents of the mind become observable since the advent of neuro-imaging technologies such as EEGs, PET scans, and fMRIs, thus early psychological research was focused on disagreements between different schools or generations of researchers that used varied approaches toward their investigations into the invisible mind. For example, cognitive researchers rely on inferences made from activities aimed at employing cognitive functions such as memory as opposed to examining how or where actual memories are laid down. Conversely Behaviorist researchers employed a more empirically rigorous method seeking only to make generalizations about phenomena that were directly observable and replicable in controlled settings.
Contemporary psychological research is derived from these disparate traditions and perspectives. It utilizes the hypothetico-deductive or scientific method:
1. observation and data gathering
2. inference of generalizations
3. construction of explanatory theories
4. deduction of hypothesis to test theories
5. hypothesis testing
6. support or challenges to existing theories and commensurate adjustments.
Theories and HypothesisEdit
Two key steps, theory construction and hypothesis deduction/testing pose special problems for researchers. Theories are sets of related generalizations explaining a specific mental phenomena e.g. schema and memory organization and hypotheses are specific predictions for research investigations. These steps are derived from empirical data, but are heavily influenced by an individual researcher’s perspective. Thus, researchers seek to clearly articulate operational definitions in an effort to make their research easily replicable. Additionally, controls are implemented to ensure credibility of results and subsequent conclusions. Finally, published research contributing to knowledge in the discipline is peer reviewed and usually rigorously scrutinized. Psychological research can take many forms ranging from: controlled laboratory true experiments (involving the manipulation of independent variables and controls for confounding variables) to field research (involving deliberate manipulation of independent variables in natural uncontrolled environments) to naturalistic/quasi experimental method (involving observation and analysis of independent variables changed by natural incidence). No matter which research method is employed, controls are carefully implemented to ensure the credibility of research. Key issues surrounding controls are: research design, sampling, reliability and validity.
The underlying structure of an investigation. It involves how psychologists use subjects/participants in their experiments. The three most common designs are:
1.Repeated Measures: using the same subjects in the experimental and control conditions
2.Independent Measures :using different subjects/participants in the experimental and control conditions
3. Matched Pairs :using different subjects/participants in the experimental and control conditions with each sample having similar characteristics.
The process of selecting participants/subjects to examine derived from a target population (a specified subpopulation of all humans). The results of a study are inferred from examination of the sample’s performance on a given measure, thus the sample is key in the line of reasoning from initial design to examination of results. Several methods can be employed when choosing a sample: random, stratified and convenience. Random sampling provides the best chance for the sample group to be representative of the target population. Stratified samples reflect similar proportions of various sub-groups within a sample. Convenience sampling involves choosing participants/subjects that are available at the time of data collection. Convenience samples do not control for possible biases that may within certain subgroups of a population and thus the results and conclusions from a convenience sample must be analyzed with caution and triangulated.
A study is reliable if it is replicable and the same results are achieved repeatedly. There are four types of reliability in regard to psychological study:
- Test-Retest Reliability (also called stability reliability)
- Interrater Reliability
- Parallel Forms Reliability
- Internal Consistency Reliability
To judge for reliability in this case, the test is administered two different times to the exact same or similar subjects. This judges for consistency of results across time, and to make sure the results were not affected by context of the time. Reliability is higher if the retest is close in chronological proximity to the original test.
Research psychologists tend to replicate older studies to generate theories or to amend findings to account for reliability. In attention for example, Treisman consistently retested findings to amend the attention models.
Two or more judges score the test. The scores are then compared to determine how much the two raters agree on the consistency on a rating system.
An example of interrater reliability would be that of teachers grading essays for an AP or IB exam. If a scale from 1 to 5 was used (where 1 is the worst and 5 is the best), and one teacher gave an essay a score of 2 and another gave a score of 5, then the reliability would be inconsistent. Through training, practice, and discussion, the individual raters can reach a consistent level of assessing an experiment, test, or result. Often, the raters are moderated by a higher rater who will assist in reaching consistency.
Parallel Forms ReliabilityEdit
A large set of questions that are related to the same construct are generated and then divided into two sets. The two different sets are given to the same sample of people at the same time. Therefore, the two tests that study the same content are judged against each other for consistency.
An example would be a pretest-posttest, where the two groups would either receive form 1 or form 2, and in the posttest situation, the groups would be switched.
Internal Consistency ReliabilityEdit
In this case, the tool is used as the tool to determine reliability. Thus would be a test situation in which the items on the test measure the same content. Often, questions can be strikingly similar, which shows that the test is also a measure for internal consistency reliability. Therefore, the similar questions should be answered in the same way. There are different ways to measure internal consistency reliability:
- Average Inter-item Correlation
- Average Itemtotal Correlation
- Split-Half Reliability
- Cronbach's Alpha (a)
Quantitative versus Qualitative MeasuresEdit
Coolican, H. (2004). Research methods and statistics in Psychology. Cambridge University Press.
1. In what ways has new technology changed the science of psychology? Provide three examples.
2. How does the importance of validity and reliability change depending on the type of study?
3. In what ways will the different aspects of an experiment (sampling, methods, reliabilty, and validity) affect the results and conclusions of an psychology study?