Panel data

Panel data consist of several units of a socio-economic panel survey, each unit is observed over at least two time periods. In the panel data analysis, various statistical models and methods are used. Exemplary panel data are the survey results from the same individuals over several years. To distinguish it from the panel data are cross-sectional data in which the units are recorded at a single time, and time-series data in which a single unit is observed over several time periods.

  • 2.1 Classification by examination subject
  • 2.2 Classification according to panel design

Assessment

Benefits

  • Compared with cross-sectional surveys, in which all relevant variables are collected simultaneously with a single measurement, panels have the advantage that causal questions can be handled much better, because changes at the individual level data can be found on the same object of investigation.
  • Opposite trend surveys, where for each new study participants will be selected, panel studies have the advantage that the behavior of the respondents can be observed over several measuring points in time. Thus, time-series studies are possible on individual data level can be any grounds for concluding causal relationships between multiple variables collected in the panel. Of course, the panel participants may additionally participate in additional ad hoc investigations.
  • Compared to traditional ad- hoc methods can be used in the household / consumer panels otherwise - about classical screening methods - often difficult or impossible -to-find ( special) target groups (eg households that have purchased a particular product in the last quarter ) - and thus select representative surveys market research economically. The basis for this form ( shopping ) data from the continuous tracking, which take place in the panels. Because the data for each household can be representatively selected " push-button " survey relevant target groups.

Disadvantages

  • Constancy of the measurement instruments: The socio-cultural change means that changes the meaning of terms. Therefore, it may happen that the same question 10 years later, no longer measures the same expression of the same construct. At the same time, new and sometimes better, more social scientific measurement and selection are developed. This resulted in a conflict between the desire to use measurement methods with the highest possible quality and consistency of the measuring instruments is to ensure.
  • Panel attrition: In the course of panel studies, it often happens that participants eg not actively participate or withdraw due to illness, relocation or loss of motivation to participate longer time in the investigation. Of particular interest for the study, therefore, are those participants who were active during the whole period of the panel. This part of the panel group is referred to as a continuous mass. The solid mass is smaller, the longer the period of investigation. This problem is referred to as panel mortality or panel mortality. It is especially problematic because it can lead to systematic distortions: Frequently who retired persons differ with respect to one or more features systematically from the remaining subjects in the panel. To counteract the problem of panel attrition, an attempt is made to keep the failure rate of panel maintenance low and to select the sample sufficiently large.
  • Panel Effects: Due to the repeated questioning the investigated objects can change. It may happen that the people interviewed only develop their attitudes on a particular topic by the survey. In addition, existing attitudes and behaviors through the repeated questioning may be changed or strengthened. Both distortions are summarized as panel effects. In addition, knowledge of the investigation process can alter the response compared to the first survey. In interview studies, the formation of friendly relations between interviewer and respondent influence the response behavior also.
  • Panel solidification: Even though the survey unit remains the same, to personal changes of the persons interviewed, such as lead the birth of children, or a professional or financial advancement to the fact that the panel loses its representativeness for the population.
  • Panel selection effects: P. denote the systematic distortion of the result of a panel study by selective recruitment process and selective failures. This problem occurs especially when the sample of individuals is formed, which voluntarily agree to participate in the panel investigation prepared as happens with access panels in commercial market research.
  • Human and financial effort: Another drawback of the panel research, although not a methodological nature, is in the higher human and financial resources management and supervision of the participants.

Special panels

View the object of investigation

  • In operation panel, the same companies are questioned repeatedly. The most important operating panel is the IAB Establishment Panel.
  • In individuals panels, the same people are repeatedly interviewed. Depending on the objectives of the investigation, the panel may be an individual panel, in which it comes to individuals, a household panel in which all the people of the household are of interest, or a specialized panel such as motorists panel, children's panel, etc. the most important person panel in the Federal Republic the " Socio-Economic Panel " (SOEP ) and
  • The micro-census ( this is also imposed in Austria, and Switzerland ). Are surveyed every year 1% of private households in Germany, which is about 390,000 households with 830,000 people.

View the panel design

In the simple panel study the whole sample is interviewed in each survey. To reduce the impact of panel attrition and panel effects, other panel designs have been developed.

  • In the alternating panel, the sample is divided into subgroups, which are alternately interviewed. For example, only the first sub-group could be interviewed in the first survey to the second survey, only the second, third in the survey again, the first and so on. In this way the burden on test subjects is reduced, and the intervals longer the risk of the panel effects is reduced. This method, however, should be applied only at sufficiently large samples.
  • Another design variant is the rotating panel differs in each survey wave a subgroup of and is replaced by a new one. This, however, takes each subgroup often different part in the survey. The group participated only once, must be excluded. The panel mortality of the group which has participated in all surveys is particularly large.
  • The divided panel ( "split panel survey" ) has the advantage that for each survey wave, a control group is available. In this way it can be determined whether panel effects occur and how strong they are. To do this, forming two groups. With the first group, a simple panel study is performed ( see above). The second group goes through either a rotating panel or repeated cross-sectional study.
  • Balanced data represent an ideal-typical data set in which all individuals have all the data for all time points. In a balanced panel ( "balanced panel data " ) are available for all the individuals observations at the same times. They are statistically somewhat easier to describe, but due to the above reasons rarely available. In reality, the data are usually incomplete, then one speaks of unbalanced panels. The use of unbalanced data sets in the considered models is not a problem, provided that the data are missing at random and enough consecutive observations are available. Other hand, systematically missing observations lead to a selection bias and make special estimation procedures necessary. A further advantage is the increase in the data base. Most are unbalanced ( " unbalanced" ) panel data in which individuals retire earlier from the survey or viewing later. We distinguish between balanced and unbalanced data.

Important panels

  • Survey of Health, Ageing and Retirement in Europe
  • IAB Establishment Panel
  • German Mobility Panel
  • Household Panel
  • National Education Panel
  • Pairfam
  • Saxon longitudinal study
  • German Ageing Survey
  • Transition Panel of the German Youth Institute
  • CSA Panel - The Open bus access panel for the social sciences of CSA
Hausman-Test
631808
de