Download social survey statistical database

Here is complete database of questions and users's answers on the export moment, expect of question about date of birth. We do not export information about user's date of birth to increase anonymity

We have got free raw statistical data, but if you will need some free tutorials on how to research, or for those who want to understand surveys and research methods in general, you can find many useful things on Free Resources for Program Evaluation and Social Research Methods

In CSV format:

or in SQL format:

Read:

Notices about export

Question's types

Export in CSV

Export in SQL

Notices about export

• Questionnaire's ids are changes from one export to another. New questionnaire placed in random part of the list, not at the end
  
• Question's ids stay the same from one export to another and coincide with online question's ids. As they are equal you can use something like "http://ru.postyour.info/123" in your Web browser to try to get translated version of question (Sorry, but only for Russian or Spanish speaking users for this moment)
• All Information are exported on English language

Question's types

All answers are encoded into digits. To understand how to use it you should know question's type. Question type is the attribute of question and it is placed near other information about question in your downloaded files

QA_SET

This type of question let user to give only one answer from the list of predefined answers for this question

Example:

Q: Have you ever seen the sea?
Possible answers: {1=No, 3=Yes, 4=It's hard to say, 5=None of the above}

User can answer Yes or No, but not Yes and No simultaneously

See list of predefined answers and it's ids for every question in the question's reference of your downloaded files

QA_MSET

User can give several predefined answers for this type of question at the same time

Example:

Do you have favourite months of the year?
Possible answers: {1=January, 2=February, 3=March, 4=April, 5=May, ..., 12=December}

User can give answer April and May simultaneously

QA_LDIGIT

This type of question imply that user could give either a voluntary digit or one of predefined answers

Example:

Q: Do you have a favourite number from 1-100 inclusive?
Possible answers: {1,2, ..., 64000, 64001=No, 64002=It's hard to say, 64003=None of the above}

The range of voluntary digit are usually extra limited: from 1 to 100 for this question

User can answer 13 that will mean 13, or "It's hard to say" than is predefined answer for this question Predefines answers ids starts from 64001 for this type of question and mentioned as predefined answers for other types

User can give only one answers for this type of question

QA_HLEN

User can give only one answers for this type of question

It is digit means length in centimeters

Example:

Q: How tall are you?
170 = 170 cm

QA_HWGHT

User can give only one answer for this type of question

It is digit means weight in kilos

Example:

Q: How much do you weigh?
71 = 71 kg

Export in CSV

questions.csv - information about questions and predefined answers

data.csv - wide CSV file with matrix of user's answers:

• Columns are questions
• Rows are questionnaires

Here is information about how to import CSV data into SPSS: SPSS Importing CSV Files Screenshot Tutorial

See Questions's types to understand how to interpret answers for this or that question

Export in SQL

If you experienced in SQL you can also use SQL format of downloading. SQL instructions in archive designed for MySQL 5. But, very likely you can also install it to many another Database Management System after several manual correction in SQL instruction in file. Your DBMS should maintain loading SQL instruction from file. You can also upgrade our basic instructions with extra indices or precise tuning to provide maximum speed of calculating

We assume you enough experienced in SQL to overcome installation on your system, so we will brief in the rest of the paragraph

There are 3 tables in the sql file data.sql

qs {id, text, comment, type} - description of questions
id - question id coinsides with online question's id
text - text of the question
comment - more precise definition of question, or additional information about it
type - type of the question

For example: {11, "Are you male or female?", "i.e. Please indicate are you male or female:","QA_SET"} ans {qid, aid, atext} - table of predefined answers for questions
qid - points to question to which this predefined answer belong to
aid - id of predefined answer unique within the limits of every qid
atext - text of the name of the answer

For example:
{11,1,"Male"}

data {uid, qid, val} - Table which store answers for questions
uid - id of questionnaire (this only need to point that several questions belong to the one questionnaire)
qid - id of questions
val - id of predefined answer or digit means user's answer

For Example:
{1,11,1} - this shows than questionnaire with id = 1 has answer "Male" for question "Are you male or female?"

Examples

Here is several samples of SQL instruction without description for MySQL which mining statistical information from data.

0. Recieve "Can you cook?" question id.

mysql> SELECT id, text, type FROM qs WHERE text LIKE "%cook%";
+-----+------------------------------------------------------------+---------+
| id  | text                                                       | type    |
+-----+------------------------------------------------------------+---------+
|  87 | Can you cook?                                              | QA_SET  |
| 282 | What form of energy do you use to cook food in your house? | QA_MSET |
+-----+------------------------------------------------------------+---------+
2 rows in set (0.06 sec)

1. Get distribution of existing answers for question "Can you cook?"

mysql> SELECT d.val, COUNT(d.val) as Num FROM data d WHERE  d.qid=87 GROUP BY d.
val;
+------+-----+
| val  | Num |
+------+-----+
|    1 |   1 |
|    2 |   2 |
|    3 |   2 |
|    4 |  22 |
|    5 |  19 |
|    6 |  13 |
|    7 |  16 |
+------+-----+
7 rows in set (0.06 sec)

2. Get distribution of answers for question "Can you cook?"

mysql> SELECT a.atext as Answer, Stats.Num FROM  ans a LEFT JOIN (SELECT d.val,
COUNT(d.val) as Num FROM data d WHERE  d.qid=87 GROUP BY d.val) Stats ON Stats.v
al = a.aid WHERE  a.qid=87;
+----------------------------------------------+------+
| Answer                                       | Num  |
+----------------------------------------------+------+
| No, absolutely not                           |    1 |
| No                                           |    2 |
| More likely no                               |    2 |
| Yes, a little. I can cook a few simple meals |   22 |
| Yes, I cook reasonably well                  |   19 |
| Yes, I cook well                             |   13 |
| Yes, I cook very well                        |   16 |
| It's hard to say                             | NULL |
| None of the above                            | NULL |
+----------------------------------------------+------+
9 rows in set (0.05 sec)

Take note - there are no any answers "None of the above"

3. Get distribution of answers for question "Can you cook?" for womans:

mysql>  SELECT a.atext as Answer, Stats.Num FROM  ans a LEFT JOIN (SELECT d.val
 COUNT(d.val) AS Num FROM data d WHERE d.qid=87 AND uid IN (SELECT uid FROM dat
 dd WHERE dd.qid = 11 AND dd.val = 2) GROUP BY d.val) Stats ON Stats.val = a.ai
 WHERE a.qid=87;
+----------------------------------------------+------+
| Answer                                       | Num  |
+----------------------------------------------+------+
| No, absolutely not                           |    1 |
| No                                           |    1 |
| More likely no                               |    1 |
| Yes, a little. I can cook a few simple meals |    9 |
| Yes, I cook reasonably well                  |   14 |
| Yes, I cook well                             |    9 |
| Yes, I cook very well                        |   13 |
| It's hard to say                             | NULL |
| None of the above                            | NULL |
+----------------------------------------------+------+
9 rows in set (0.03 sec)

4. Get distribution of answers for question "Can you cook?" for womans who not younger than 35:

mysql> SELECT a.atext AS Answer, Stats.num FROM ans a LEFT JOIN (SELECT val, COU
NT(val) AS num FROM data WHERE qid = 87 AND uid IN (SELECT uid FROM (SELECT uid,
  COUNT(uid) AS ucount FROM data WHERE (qid = 11 and val = 2) or (qid = 10 and v
al>=35) GROUP BY uid) list WHERE ucount = 2) GROUP BY val) Stats ON Stats.val =
a.aid WHERE a.qid=87;
+----------------------------------------------+------+
| Answer                                       | num  |
+----------------------------------------------+------+
| No, absolutely not                           |    1 |
| No                                           | NULL |
| More likely no                               | NULL |
| Yes, a little. I can cook a few simple meals |    2 |
| Yes, I cook reasonably well                  |    7 |
| Yes, I cook well                             |    4 |
| Yes, I cook very well                        |    1 |
| It's hard to say                             | NULL |
| None of the above                            | NULL |
+----------------------------------------------+------+
9 rows in set (0.06 sec)

Here we use the fact that uid who has answer female and uid who has answer age>=35 appear 2 times ( ucount = 2 ) in group of uid