Note for users of the IAB establishment panel concerning the change in value coding for answers of questions
Variables ba06a, ba17, ba18, ba72, ba73, ba86a, ba86b, wave 2019
In wave 2019, some values for answer options to certain questions changed. However, the underlying question, answer options and value label are still identical to previous waves. The affected variables and values:
- ba06a:
Answer option „Not sure at present“
New value: -1
Value previously (z06a, wave2018): 4 - ba17:
Answer option „Annual result is not available, indication therefore not possible“
New value: -1
Value previously (z16, wave 2018): 4 - ba18:
Answer option „Don’t know at present“
New value: -1
Value previously (z17, wave 2018): 4 - ba72:
Answer option „Difficult to say in general, varies from case to case“
New value: -1
Value previously (y70, wave 2017): 4 - ba73:
Answer option „Difficult to say in general, varies from case to case“
New value: -1
Value previously (y71, wave 2017): 4 - ba86a:
Answer option „Unknown“
New value: -1
Value previously (z86a, wave 2018): 3 - ba86b:
Answer option „Unknown“
New value: -1
Value previously (z86b, wave 2018): 5
The IAB recommends to recode the mentioned values from 3 / 4 / 5 to -1 or vice versa when using those variables in panel analyses. Thereby, these variables become consistent in their value coding over time. If using scripts for cross-section analyses that are based on wave 2018 or earlier waves, the mentioned values should also be adjusted when using one of those variables.
Agreed average working time
Variable z64, wave 2018
In the case of the variable on agreed average working time (z64), a significant increase in the proportion of missing values from approx. 4.7% to 8.3% can be observed in 2018. This increase is due to an error in data processing in the CAWI sample (web-based interview).
Trust-based working time agreements
Variable n62b, wave 2006
The selection of the answer “the whole company” in question n62b indicates that all employees within the establishment have the opportunity to make use of trust-based working time agreements. (There is no information available on the number of employees using these regulations.) From a critical point of view, it can be questioned if this actually applies for all employees in one establishment. It is reasonable that establishment respondents also thick the answer “the whole company” when trust-based working time agreements are implemented on a company level, but do not apply for single divisions within the company. Depending on the interpretation of question n62b, this may lead to an overestimation of the share of workers employed on trust-based working time agreements.
Working time accounts
Variables: Wave 2009 (Question 30), Wave 2010 (Question 73), Wave 2011 (Question 61), Wave 2012 (Question 56), Wave 2013 (Question 80), Wave 2014 (Question 51) Wave 2015 (Question 51), Wave 2016 (Question 67)
We would like to inform users of a problem in module G “Operational Working Hours”. Since 2009 only basic questions on this subject are asked in uneven years (basic module). In even years the questionnaire comprises a further set of questions (basic and extension module). This leads to considerable variation in the share of establishments claiming to use working time accounts with higher shares in uneven years.
The IAB explains this as follows:
In even years, the questionnaire contains detailed questions about overtime and their forms of compensation located in front of the questions on working time accounts whereas in uneven years the compensation of overtime is not addressed.
We assume that the problem arises with establishments that manage some kind of flexibility by using overtime but do not consider themselves per se as using working time accounts. In even years, these establishments can choose the appropriate forms of overtime compensation and refrain from ticking working time accounts. In uneven years, this procedure is not possible. Here the solution seems to be to present themselves as establishments with working time accounts as the option coming closest to what is happening in the firm (flexibility by building up and reducing overtime). This assumption is supported by longitudinal analysis revealing that a larger proportion of establishments switch back and force between using some forms of overtime compensation (and not working time accounts) in even years and using working time accounts in uneven years.
The IAB recommends to consider this in analyses on working time accounts form 2009 on. For analyses over time is seems to be more appropriate to use the data of every other year (either even or uneven years).
Existence of opening clauses in industry-wide collective wage agreements
Wave 2005 (question 55), 2007 (Question 84) and 2011 (Question 55)
Bei der Auswertung der oben genannten Fragen sollte berücksichtigt werden, dass hier zwar ein objektiver Tatbestand (TV enthält Öffnungsklausel) abgefragt wird, die Antworten jedoch nicht immer den tatsächlichen vertraglichen Sachverhalt widerspiegeln, sondern vielmehr die betriebliche Wahrnehmung dieses Sachverhalts, die stark vom Kenntnisstand der Auskunftsperson geprägt ist. Das ist zwar generell so in Betriebsbefragungen wie dem IAB-Betriebspanel. Anders als bei allen anderen Fragen wird hier aber nach einem quasi betriebsexternen Faktum gefragt, zu dem es objektive Informationen (Tarifverträge) gibt.
Abhängig vom Erkenntnisinteresse und dem Design empirischer Studien kann die im IAB-Betriebspanel vorliegende Information zum Wissen über die Existenz von Öffnungsklauseln in den Tarifverträgen zu Problemen führen, und zwar dann, wenn Informationen zur tatsächlichen Existenz solcher Regelungen benötigt werden.
Recruitment of staff for qualified and menial jobs
Variables u36, u38, wave 2013
In some cases the sum of hires indicated in questions 36b (number of hires for qualified jobs = u36ber+u36uni) and 38b (number of hires for menial jobs =u36einf) differs from the number of total hires indicated in question 35 (u35ges). This affects 240 cases; most of them (70 percent) differ by one or two hires.