-
Notifications
You must be signed in to change notification settings - Fork 0
Read this before analyses
Please read the following important information regarding data from the Norwegian Mother, Father and Child Cohort Study (MoBa) before analyzing the data.
Most of the documentation at the MoBa Wiki web site is translated into English, although some documentation is only available in Norwegian.
There are more than one version of each questionnaire (A, B, C, ..). Many questions are identical across versions and the data is stored in the same variable. Where the questions differ significantly between versions data is stored in separate variables. A few questions are version specific. In addition, there are both paper versions and a digital version (W) for the dietary (Q2), 3 year (Q6) and 7 year (Q7) questionnaire. An overview of all questionnaires can be found on the wiki page Questionnaires and on fhi.no (English versions, Norwegian versions)
The variable names consist of two or more letters and a number. In questionnaire 1 (Q1) all variables (corresponding to questions) are named ‘AA’ followed by a number. In questionnaire 3 (Q3) all variables (corresponding to questions) are named ‘CC’ followed by a number.
The variable labels include the following information:
- The questionnaire version(s) in which the question is found
- The question number(s)
- The question text
Note that the question number may differ in different versions of the questionnaire. For the majority of the questionnaires, labels have been translated into English.
It is essential to know the basic rules applied for coding of data before analyzing the data. A categorical variable is coded ‘1’ for the first category, ‘2’ for the second category, and so on. Normally, the answer category ‘No’ is coded ‘1’ and ‘Yes’ is coded ‘2’. However, in some cases it might be opposite, where ‘Yes’ is coded ‘1’ and ‘No’ is coded ‘2’. In general, one should always check the questionnaires for the correct sequence of categories. Dichotomous variables are coded ‘1’ for a tick and a blank (missing) value when there is no tick.
Variables corresponding to questions where only one answer is intended (index variables) are coded ‘0’ if more than one answer is given. However, some index variables are given valid codes for combinations of two or more ticks.
Occupational information from questionnaire 1 is coded for approximately 70 % of the participants. Occupational data is mainly available on an aggregate level (See wiki page Coding of occupation). All medicines stated in questionnaires during pregnancy and up until the child is 7 years have been coded using the ATC classification system (See wiki page Coding of medication). Data for diseases and other text fields are not coded and are not included in the data sets.
MoBa data have been through extensive quality control procedures. These are divided in two levels.
Level 1: During scanning and verification, all scanned and digitalized numbers are evaluated to check correspondence with what is written in the questionnaires. Some check boxes will also be validated, i.e. when more than one check box is marked for a group of mutually exclusive check boxes.
Level 2: Registered data are evaluated again. Rules have been defined for each variable to check values that are (slightly) extreme or not consistent with other answers. Note that limits for valid values set in the quality control procedure do not necessarily comply with plausible limits. Specifically, numbers are checked by comparing the digitalized number with the written response in the questionnaire. Where there are discrepancies, corrections are made. Otherwise data are left unchanged, even if implausible or illogical. Hence, for cut-offs, researchers must use their own considerations. Rules applied in level 2 of the quality control can be found on wiki page Quality control. Quality control is an ongoing process. Attached documentation explains what has already been done to ensure good data quality. However, if errors are discovered by use of MoBa data, please let us know. If necessary, new rules for quality control will be implemented.
Limits for valid values are incorporated in the digital questionnaires. I.e. for questions where only one answer is intended a radio button that only let the user select one option of a collection of options is used in the questionnaire.
Some questions are excluded from the data files for confidential and practical reasons. The information given in these questions is incorporated in generated variables. The most important generated variables are described in the sections below.
All date variables in the questionnaires are replaced by variables that correspond with the child’s age in days/months/years.
The generated variable ‘ALDERUTFYLT_Sx’ contains information about when the questionnaire was filled out by the respondent, given by time in days/months/years from the date of birth of the child until filling out the questionnaire. This will correspond to the child’s age in days, months or years.
The variable ‘ALDERUTFYLT_Sx’ is calculated from ”Date for filling out the questionnaire” given by the respondent in the questionnaire.
Note that for the first (Q1), second (Q2) and third (Q3) mother's questionnaire as well as the first father’s questionnaire (QF), the variable ‘ALDERUTFYLT_Sx’ corresponds to the number of days before the child is born.
Also note that we have observed some evident errors in the variable ”Date for filling out the questionnaire”. Some respondents have written their own date of birth, their child’s date of birth or other dates that are not valid. This results in errors in the generated variable ‘ALDERUTFYLT_Sx’. To control for these errors, we have added two generated variables to the dataset; ‘ALDERUTSENDT_Sx’ and ‘ALDERRETUR_Sx’. These variables are calculated in the same way as ‘ALDERUTFYLT_Sx’, but utilize the date the questionnaire was sent and the date we received it. These variables may be used for estimation of the child’s age when ‘ALDERUTFYLT_Sx’ is missing or implausible.
The variable ‘ALDERUTFYLT_Sx’ is calculated from the date when the questionnaire was submitted online by the respondent.
A category of variables has been generated, providing the total number of responses filled in on each page of the questionnaire by the respondent. The variable names consist of the questionnaire number and page number; Q1P1, Q1P2 etc. Note that a question may be placed on different pages in different versions of the same questionnaire. These variables are meant to simplify quality control of the data for users and the interpretation of, for instance, missing values (e.g. without noticing, some women have left two pages blank because two sheets have been stuck together).
38 variables are generated from the dietary questionnaire and are included in the file “Q2_calculation”. These variables are calculated by use of the Norwegian Food Composition Table 2001. Note that dietary supplements are not included in the intake calculations of vitamins and minerals. The variable ‘VERSJON_KOST_TBL1’ specify which version (KOST_A=A/B or KOST_B=C/D/W) of the dietary questionnaire the calculations are based on. Questionnaires A and B are very different from questionnaires C, D and W. In version A and B the women are asked what they have been eating the last 12 months before they became pregnant, while in version C, D and W they are asked what they have been eating since they became pregnant and until the day they filled out the questionnaire (about week 17-22). Note that data from the first two versions (A and B) of the diet questionnaire, are given in separate files only by request, since they are not comparable to the questionnaires C, D and W.
Considering the large amount of data, each questionnaire is delivered as separate data files. When merging the files for analysis, it is important to be aware of the following
- Data files based on data before birth (questionnaire at week 15-17 of pregnancy (Q1, QF), Week 22 of pregnancy (Q2), Week 30 of pregnancy(Q3)): Include an unique identifier for pregnancy called ‘PREG_ID_XX’. This variable must be used as the key variable when merging files.
- Data files based on data after birth (questionnaire at 6 months (Q4), 18 months (Q5), 3 years (Q6), 5 years (Q5Y), 7 years (Q7Y) etc.): Include an identifier for pregnancy (‘PREG_ID_XX’) and a identifier for the child (‘BARN_NR’). Both variables ‘PREG_ID_XX’ and ‘BARN_NR’ must be used when merging files based on data after birth to ensure that data is linked correctly for twins/triplets.
-
For the second Father questionnaire (QF2):
- Data file based on data from Part 1 (about the father) include an unique identifier for the father called ‘F_ID_XX’. This variable must be used as the key variable when merging files.
- Data file based on data from Part 2 (about the child) include an identifier for pregnancy (‘PREG_ID_XX’) and a identifier for the child (‘BARN_NR’). Both variables ‘PREG_ID_XX’ and ‘BARN_NR’ must be used when merging files based on data after birth to ensure that data is linked correctly for twins/triplets.
Syntax for merging files can be found on the wiki page Merge files. Some participants will have data from all questionnaires, while others will only have data from one questionnaire. It is crucial to be aware of this when interpreting missing values. The data set “SV_INFO” includes a unique identifier for each woman called ‘M_ID_XX’. This variable can be used to identify women who participate with more than one pregnancy and hence identify siblings.
Not all questions are present in every version of a questionnaire. Hence, a variable may have a missing value because the respondent has answered a version of the questionnaire that does not include the specific question. The variable label indicates in which version(s) a specific question was asked, while the variable ‘VERSJON_SKJEMAX_TBL1’ indicates which version of the questionnaire each respondent has answered. If you want to use a variable not included in e.g. version A of a questionnaire, you can use ‘VERSJON_SKJEMAX_TBL1’ to remove version A-respondents from the data set. We urge you to consider how the specific question may be interpreted by the respondent. If a respondent has answered “No” to a question, he/she may not have answered the follow-up questions and the corresponding variable(s) will have a missing value. Some respondents may also have answered inconsistently, e.g. answered “No” when asked “Have you ever smoked?” and answered “Daily” when asked “Do you smoke now?”. Therefore, it is important to check for inconsistency. For twins and triplets, the mother fills out one questionnaire for each child after birth (i.e. questionnaire 6 months, 18 months, 3 years, 5 years, 7 years and 8 years). The mothers are told to fill in the section “About yourself” for the first child only. MoBa have not duplicated this information to the other sibling(s). These variables are usually missing for one of the twins or two of the triplets.
If you have questions or comments about the data files or suspect that something can be incorrect in labels or other documentation, please contact us at [email protected].
- Read this before analyses
- Questionnaires
- Generated variables
- Dietary Data
- Coding of medication (ATC-codes)
- Coding of occupation and industry
- Quality Control
- Genetic data in MoBa
- Phenotools
- Alcohol consumption in pregnancy
- Arthritis in pregnancy
- Caffeinated beverages in pregnancy
- Diabetes in pregnancy
- Drug use in pregnancy
- Epilepsy in pregnancy
- Fever in pregnancy
- Inflammatory bowel disease in pregnancy
- Nausea and vomiting in pregnancy
- Smoking during pregnancy
- Thyroid disorder in pregnancy
- Urinary infection in pregnancy
- Vaginal bleeding in pregnancy