¶ … scientific observation that distinguish it from our everyday observation are that scientific observation is conducted using precisely defined observational conditions; by performing the observations systematically and objectively; and through keeping careful and accurate records. Scientific observation, as opposed to everyday observations, must take place within certain well-defined parameters, whether in naturalistic or laboratory settings. Furthermore, the scientist does not choose the parameters arbitrarily but rather relies on such methods as sampling to conduct the observations and experiments in a way that renders them valid. For example, if a researcher wanted to examine the effects of watching violent video games on a child's behavior, he or she would first have to determine which video games would be deemed violent, and what specific behavioral affects to look for. An everyday observation of the same phenomenon would be far less precise.

Also, everyday observation can be passive and filtered through the individual's biases and opinions. Scientific observation, on the other hand, must be conscientiously objective and unbiased. Scientific observations are therefore often more specific and focused than everyday observations, which can be generalized and unspecific.

Scientific observation also demands the keeping of records and data, so that the results of the study can be calculated and analyzed. Such records-keeping must be as precise as possible. Everyday observations do not require such close attention to records. Even if an individual records his or her behaviors, thoughts, and feelings in a journal, those records are generally not as accurate as scientific records, which can employ mathematical computations.

2. Researchers rely heavily on sampling, either time sampling or situational sampling, in their studies. Time sampling entails observing the subjects' behavior at certain time intervals. If the behavior is a rare occurrence, such as a nightmare or a panic attack, then the researchers use what is known as event sampling, a specific form of time sampling. Situation sampling entails observing the same behavior under various circumstances. For example, a researcher could examine police brutality in different neighborhoods, or at different times of day. Time and situational sampling are often used in conjunction with each other.

Sampling usually increases the external validity of the experiment, which is one of the main reasons scientists use carefully-chosen sampling methods. One of the specific reasons to use sampling when performing empirical research is to create a representative group of subjects. Also, it would be impossible or too time-consuming to observe every possible person and every possible behavior, and therefore the scientist samples the population and can then extrapolate the data. Samples should ideally be as representative as possible to ensure external validity. The sampled group is said to represent the population as a whole. Both time and situational sampling ensure the sample's accuracy and external validity. The sampled group will demonstrate a certain behavior or characteristic that can be observed closely under the experimental settings. Researchers can replicate the findings by selecting similar samples. Sampling ensures the external validity of the study and its ability to be extended applied to various settings and situations. Without using sampling methods, the experiment would be too arbitrary, and the results would be hard to replicate.

3. Participant observation, in which the researcher or an assistant actively participates in the research situation, can be performed either disguised or undisguised. When disguised, the researcher does not want the population sample to know they are being observed, whereas in undisguised participant observation, the researcher may interact fully with the population sample.

Participant observation can significantly influence the behavior being observed. In some cases the influence is desirable, as when an anthropologist lives for a long period of time with a group of people. Participating while observing ensures that the group under observation acts as naturally as possible, and also helps the researcher better formulate hypotheses and opinions. However, in many cases participant observation can hamper the reliability of the experimental results. For example, most people behave differently when they know they are being watched. Many people might provide answers to surveys that they think the experimenter wants, rather than those that reflect the true attitudes of the subject.

Three factors in participant observation that researchers need to consider to determine the extent of the observer's influence on the behavior being observed include the following: whether the participant observation is disguised or undisguised;...

Certain types of studies would warrant disguised participant observation, while others would perform better through undisguised participant observations. The size of the group will also affect the level of influence of the participant observation. On the one hand, the observer will be less noticeable in a larger group, but on the other hand a larger group means that there might be more information that the researcher ignores or misses while observing. The role of the researcher is also a significant factor in determining the influence of the observation. For example, if the researcher is viewed as an authority figure, the subject might try to "please" the observer by acting or behaving a certain way. Age, ethnicity, and gender are some factors that might affect the role of the researcher. Other issues that researchers must consider include ethical issues such as privacy; and the degree to which the behavior might be altered in the presence of the observer.
4. Structured observation represents a compromise between naturalistic observation and laboratory experiments. The primary advantage of this compromise is that researchers can use structured observation to study behaviors that may otherwise be difficult to observe in naturalistic settings, such as studies involving sleep and dreaming. Also, using structured observations, the scientist can choose to observe behaviors in a relatively natural setting; although structured, the researcher does not necessarily need to conduct the experiment in the confines of a laboratory. Structured observations enable the scientist to study a wide range of behavioral phenomenon in ways that are empirically sound but also in ways that mirror natural behaviors.

The potential cost of this compromise is that it may be difficult for other researchers to replicate the study, especially if the research parameters were unusual or the circumstances rare or elaborately constructed. Also, the scientist may not foresee or may underestimate the impact of unknown variables on the outcome of the study. For example, a researcher does not always replicate the exact conditions of an initial study when conducting a new experiment.

5. Quantitative measures generally fall into one of four categories: nominal, ordinal, interval, and ratio. Certain variables lend themselves to being measured by one or more of these quantitative measures. In some cases, more than one measure can be used to study the same variable.

For example, if a researcher were trying to measure eye contact between pairs of people in conversation, he or she could use a nominal measurement scale by simply observing whether or not the two people made eye contact at all during the conversation. Or, the nominal measurement scale would be used for some of the independent variables, such as the gender of the individuals involved. The nominal scale is best used to measure simple, either/or variables.

If the researcher used an ordinal scale, he or she can increase the complexity of the measure somewhat. For example, the researcher could observe the number of times each pair made eye contact and assign the groups to three different categories based on frequency of eye contact. If the pair made eye contact fewer than five times during the conversation, they would be labeled as "Low eye contact," if they made eye contact between five and ten times they would be labeled "Medium eye contact," and if they made eye contact more than ten times they would be labeled "High eye contact." Using these three groups, the researcher could postulate that individuals that fell into the high eye contact group were more likely to be extroverted than individuals who fell into the low or medium eye contact groups.

If the scientist were to use an interval scale, he or she could use a similar design as used in the ordinal measure, but instead of creating three categories, the researcher would count each instance of eye contact as one interval. The dependent variable in such a study would use the number of instances of eye contact as the results: if a pair made eye contact five times, then the results for that pair would five. The results of an interval scale would yield powerful results that could be used for statistical analyses.

Finally, the researcher could measure eye contact using a ratio measurement scale. In this case, instances of eye contact would be treated in relationship to a measure of absolute zero. Absolute zero in this case would mean that the pair never once made eye contact. Each instance of eye contact would be measured as one notch above that. Alternatively, the researcher could try to measure the total amount of time the pair maintained eye contact, for one period of eye contact could last anywhere from a fraction of a second to a whole minute. Therefore,…