Science is the attempt to answer questions through the systematic collection and analysis of objective, observable data.

Clever Hans was a horse, which had received an education and seemed to be able to answer a lot of questions, including arithmetic questions. This was, in fact, not true. The horse was able to recognize visual signals to which he responded. The story of Clever Hans shows that the result of an experiment (e.g: asking a horse what 9+10 is) can be influenced by the observers (they give unintended visual signals). This is phenomenon is known as the observer-expectancy effect.

A theory is an idea or conceptual model, that is designed to explain existing facts and make predictions about new facts that might be discovered. A hypothesis is a prediction about new facts based on the theory.

Scepticism leads to more mundane explanations instead of a highly unlikely one because scepticism leads us to test, rather than accept a bizarre theory. A theory has to be parsimonious. The simpler, more sober theories are preferred over complex theories. A theory also has to be falsifiable. Experiments should be conducted in a controlled environment because the observer (and other things) can (unintentionally) influence the outcome of the experiment. It is important that the researcher controls the conditions, to make sure that no unaccounted conditions influence the outcome of the research and to exclude alternative explanations. Besides that, researchers should beware of the observer-expectancy effects.

There are several types of research determined by:

• Research Design
  Experimental research, correlational research and descriptive research
• Research Setting
  Field and laboratory
• Data-collection method
  Self-report and observation

Experimental research is used to determine a cause-effect relationship between two variables. An experiment is a procedure in which a researcher systematically manipulates one or more independent variables and looks for changes in one or more dependent variable while keeping all other variables constant. There are two types of variables in a cause-effect relationship. An independent variable is a variable that is determined and controlled by the researcher. A dependent variable is a variable that (hopefully) is affected by the independent variable. There are two basic types of experiments. A within-subject experiment, in which the participant is exposed to each of the conditions of the independent variables and is repeatedly tested. It is a within-subject experiment if there multiple conditions of the independent variable are applied to the same subject. A between-subjects experiment, in which the participant is only tested in one condition of the independent variable and there are multiple participant groups, a group for each of the conditions of the independent variable.

A correlational study is used when you cannot conduct experimental research, such as when researching personal variables that can´t be manipulated, such as divorce. A correlational study is a study in which the researcher does not manipulate any variable, but observes or measures two or more already existing variables to find a relation between them.

It is not possible to get a cause and effect relation between two variables through a correlational study, because it is never certain that the two variables affect each other the way you might think they do. It is possible that A affects B, but also possible that B affects A, or that C affects A or B and because of this it is not possible to get a cause-effect relation by conducting correlational research. Correlation does not imply causality.

A descriptive study describes the behaviour of an individual or a set of individuals without assessing relationships between variables. A descriptive study can be very similar to a correlational study, especially when a descriptive study uses numbers (e.g: the amount of mental disorders in a community), but when the goal is to describe the situation and not to correlate two variables then it is a descriptive study.

A laboratory study is any study in which the participants are brought to a specifically designated area that has been set up to facilitate the researcher’s collection of data or control over environmental conditions. A laboratory study can be anywhere, as long as the researcher has control over the experiences the participants have at that time. A field study is any research study in which the researcher does not have control over the experiences a participant has.

A laboratory study allows the researcher to collect data under more controlled conditions than in the field, but an artificial environment can cause different behaviour than usual and the one the researcher wants to study. E.g: parent-child interactions may be different in a controlled environment than when at home.

A field experiment is possible when a researcher manipulates a variable but has no control over the other variables. (e.g: putting signs somewhere and observing what happens. The researcher has control over the variable signs, but has no control over the other variables, while he wants to research the behaviour the signs cause)

There are two broad dimensions of data-collection methods. The first one is the self-report. Self-report methods are procedures in which people are asked to rate or to describe their behaviour or mental state in some way. One form of self-report is introspection: the personal observations of one’s thoughts, perceptions and feelings. Introspection is subjective, but not unreal. Another form of self-report is when someone is asked to assessments of other people. (e.g: teacher being asked to evaluate children in terms of aggression).

Observational research includes all procedures by which observe and record the behaviour of interest rather than relying on the participant’s self-report. It is possible to test behaviour or to have a naturalistic observation in which the observer doesn’t do anything but observe. Naturalistic observation has one important disadvantage to it. People may change their behaviour if they know they are being observed. This is called the Hawthorne effect. There are two ways to minimise the Hawthorne effect. The first is called habituation. If a stimulus is repeatedly or continuously present, the participants might habituate to the presence. Another way to minimise the Hawthorne effect is hiding.

Statistics are needed to determine that the likelihood that observed patterns in data are simply the result of chance. The statistic procedures used for these purposes can be divided into two categories: descriptive statistics, which are used to summarize sets of data and inferential statistics, which help the researchers decide how confident they can be in judging that the results they observed are not due to chance.

The mean is the arithmetic average. The median is the centre score. Variability refers to the degree to which the numbers in the set differ from one another and their mean. A common measure of variability is the standard deviation. The further the most individual scores are from the mean, the greater the standard deviation is.

A correlation can be positive or negative. A positive correlation (>0) means that when variable A increases, variable B increases as well. A negative correlation (<0) means that when variable A increases, variable B decreases. If the correlation is near 0, it means that the two variables are unrelated.

Inferential statistic methods are procedures for calculating the probability that the observed effects could derive from chance alone. In a correlational study, p is the probability that a correlation coefficient as large as or larger than observed would occur by chance. If ‘p’ is <0.05 then it means that the results are statistically significant; the chance that the results occurred by chance is acceptably low.

When calculation ‘p’ the following things are needed in the calculations:

• The size of the observed effect
  The larger the size of the effect, the higher the chance that the results are statistically significant.
• The number of participants in the study
  The more participants, the less likely it is that the results occurred by chance.
• The variability of data within each group
  The less variability of data within each group, the less likely it is that the results occurred by chance

Bias refers to non-random effects caused by some factor or factors extraneous to the researchers' hypothesis. Bias can lead researchers to false conclusions, while randomness (and thus error) only increases the chance of finding statistically insignificant results. With a bias, a factor, irrelevant to the hypothesis has influenced the results. Three types of bias are the sampling bias, the measurement bias and the expectancy bias.

If the members of a particular group are systematically different from those of another group or are different than the population the researcher is interested in there is a biased sample. When conducting research, participants should be assigned to a group randomly, because the initial differences will then merely be a source of error, otherwise, there will be a bias. It is not possible to draw conclusions about a population when there is a bias in the research. A sample is biased when it is not representative of the population that researchers are trying to describe.

A measure is reliable if the measurement is:

• Replicable
  It yields about the same result every time it is used with a participant under a set of conditions, also known as replicability. Low reliability decreases the chance of finding statistical significance in research.
• Interobserver reliability
  Two observers should come to the same result when measuring something and they should be able to observe the same behaviour. The behaviour should be carefully defined ahead of time. This is done by generating an operational definition.

A measurement procedure is valid if it measures or predicts what you want it to measure or predict. A lack of validity can be a source of bias. There are several types of validity:

• Face validity
  Common sense should tell us that something is valid (it should be a reasonable way to measure what you want to measure)
• Criterion validity
  Does the measurement correlate with a related variable?

In a previous experiment with autistic children who did not have language ability, but appeared to have exactly that because a facilitator would help them type things on a keyboard. This showed the observer-expectancy bias. The observers (in this case the facilitators) expected something and unintentionally helped the autistic children write exactly that.

The observer-expectancy bias reveals itself in two ways. If an observer expects something he acts (unintentionally) different towards the participants or if the observer expects something he is quicker to see what he expects. This can be prevented by keeping the observer blind from what is expected. Participants also have expectations. This is called subject-expectancy bias. This can also be prevented by conducting double-blind research in which both the observer and the participant don’t know what is expected of them. When administering drugs to participants some should receive a placebo, to make sure that the drug doesn’t only work because of the placebo effect.

In research with humans, ethical considerations revolve around three interrelated issues:

1. A person’s right to privacy
2. The possibility of discomfort or harm
   If there is a possibility of harm there must be made sure that there is no way to test what the researchers want to test with less or no chance of harm and should be outweighed by the human benefits. Besides that, participants have to be free to quit at any time.
3. The use of deception
   Not all psychologists agree on this matter, but some psychologists believe that the use of deception is unethical and undermines the possibility of obtaining truly informed consent. Special considerations, in this case, must be given when the participants are children, have limited intellectual capabilities or have limited capabilities of their own (e.g: prisoners)

When animals are used in research the discomfort and the suffering of the animal has to be outweighed by the potential benefits of the research. Animals must be well cared for and not subjected to unnecessary deprivation or pain.