Psychometric (assessment and techniques). Psychometric methods involve measuring psychological characteristics, such as personality, intelligence or aptitudes. Statistical techniques are particularly important, to establish reliability (does the test produce consistent results?) and validity (is the test measuring what it is supposed to be measuring?). Another important concept is 'norms'. For example, on intelligence tests, the norm for children of a particular age is the average score gained by the children of that age on whom the test was standardised. It therefore reflects typical development for that age group. 
Personality structure and Personality development . The term 'personality' can have widely divergent meanings. It can be seen as the outcome of the factors that make a person different from others in their patterns of thought, values, beliefs, emotions, and social roles and relationships. But personality can also be seen as the 'something'- the underlying construct- that leads to people's consistency in the way they behave and what they feel, over time and across many situations. In this sense, personality may be essentially unique to each of us but it can also be seen as a profile of underlying dimensions (possibly with a biological bases) that are common to many people.\nPersonality is an important concept in clinical, psychoanalytical and psychiatric contexts, particularly in terms of diagnosing and treating people seen as having disorders of the personality (such as compulsive, psychopathic, or paranoid personality disorders).
Surveys involve the collection of much less detailed information than case studies, but include much larger samples. Survey data can be either descriptive (finding out about certain characteristics of the group being examined) or explanatory (aiming to provide information that researchers can use to look at potential cause-effect relationships). In practise, most surveys probably have elements of both the descriptive and the explanatory, though perhaps in different proportions. This is because the survey method is comparatively weak at finding out cause-effect relationships (compared with experiments, for example), and even 'descriptive' surveys will usually aim to provide some explanations for the results. Questionnaires make up the basic type of survey, which simply ask people to answer a list of questions. Often this may involve giving them a pre-printed checklist of questions, which they fill in themselves and return, though it could also involve being asked questions by the researcher (the usual form with political opinion polls, for example). The questions can be closed questions, where the range of responses is restricted to those provided by the research, or more open-ended questions which do not limit so much the kind of response that can be given (which, however, can then be more difficult to code/analyse). There could of course be a mixture of open and closed questions. Attitude surveys could be seen as a particular specialised type of questionnaire, finding out about peoples' attitudes on a given topic, e.g. levels of crime, whether drug-taking should be decriminalised etc. This method uses an attitude scale to indicate the strength with which people hold the attitude being investigated, by presenting them with a series of questions relating to the topic. For example, with the drug-taking topic, one item might be 'criminalising drug-taking does nothing to reduce the overall level of drug use in society'. Participants would then give their response to this item as either 'strongly agree/agree/disagree/disagree strongly' etc. Correlational methods\nThis approach investigates statistical relationships between two or more variables, where there is some kind of orderly association between the variables (i.e. as one changes, the other also tends to). This can be reflected mathematically in a correlation coefficient, indicating the strength of the association. For example, a researcher might look for correlations between a child's reading age and IQ score. Note that the presence of a correlation doesn't prove that one variable is acting directly to cause the other to change, so care must be taken with interpreting these data (contrast with experiments, which are designed to reveal causal connections).