What is ODA?

ODA stands for "Online Data Analysis." It is a tool that was developed by the Caucasus Research Resource Centers (CRRC) in order to publish and visualize survey data collected in the three countries of the South Caucasus - Armenia, Azerbaijan and Georgia. The platform allows users, who may not be familiar with statistical software and lack sophisticated data analysis skills, to create charts, display frequency and contingency tables, and analyze cross-country and time-series data.

What can I do with ODA?

ODA datasets are organized in three different ways. How you use ODA will depend on what kind of information you want to receive.

In the single-dataset mode, you can analyze a particular wave of the survey in one country (for example, the Caucasus Barometer in Armenia, 2013). After selecting the dataset, you can choose the variable(s) of interest and generate a frequency table and/or chart. Single-dataset mode does not show differences across countries or trends in time. However, you can analyze particular population groups (e.g., by age, by settlement type) by building contingency tables (crosstabs). Crosstabulation is a powerful tool; however, you should remember that when choosing a very small population group (e.g., less than 5% of the sample), the results might be irrelevant.

Cross-country mode allows you to see the differences between Armenia, Azerbaijan and Georgia within one particular survey (for example, Caucasus Barometer, 2013). After selecting regional datasets, you can choose the variable and generate frequency tables and/or charts. This mode lets you see differences between countries. However, it does not allow you to build crosstabs.

Time-series mode is a perfect way to observe how attitudes change over time. After selecting the dataset, you can select the country and the variable of interest, and generate table(s) and/or chart(s). The charts display frequencies for each year and corresponding trend lines. As in cross-country mode, you cannot build contingency tables in time-series mode.

What is the codebook?

The codebook is a list of all the variables included in the particular dataset. Information about each variable consists of the text of the questions asked to the respondent and the answer options. When relevant, specific notes are also provided. For example, when the answer options were recoded for ODA and hence are different from the ones in the questionnaire, a note is given. The notes also show whether the variable was asked to a particular sub-group of the respondents and not the entire sample. The actual questionnaires are also available from the downloads section of ODA.

How to search in ODA?

If you need to find the questions asked on the topic you are interested in, the Search function will help. It is available in all ODA modes. The system finds the search term either in the text of questions, in answer options, or in variable labels - for example, if you look for "employment," the query will yield a list of variables which have "employment" in any of the categories.

What is the factsheet?

Factsheets accompany each dataset and give basic facts about the survey - sample size, response rates, geographic coverage, fieldwork dates, and so on. For more detailed information, you can download survey documentation and datasets in STATA and/or SPSS.

Can I export data from ODA?

ODA allows you to download charts as PNG images and CSV tables for further analysis or visualization. You can also embed charts in your blogs or other online materials (copy and paste the chart code into your blog or website HTML).

Can I download source datasets?

Of course. You can download datasets and other essential documents from the downloads page.

What is the difference between source datasets and ODA?

The datasets used for ODA differ in a number of aspects from the original datasets in STATA. Some of the answer options are grouped together for ODA. This is done mostly in order to ease visualization. Specifically, it sometimes occurs in the following cases:

  • In the case of open-ended questions (e.g. "How many children do you have?"), the options are grouped in a more sensible way (e.g. instead of number of kids, ODA displays whether the respondent has children);
  • Sometimes the scales from the original dataset are reduced in order to be displayed on ODA more properly. For instance, 10-point scales are recoded to 5-point scales - options 1 and 2 are coded as 1, options 3 and 4 as 2 and so on;
  • If the variables have many answer options, the options with low frequency are grouped and labeled as "Other";
  • The answer options "Don't know" and "Refuse to answer" are often grouped together.

If the variable in ODA is different from the variable in the original dataset, a special note will accompany the codebook, the chart and the respective table.

Should I cite ODA?

Yes, you should cite ODA and CRRC in your works as the source of the data. The following citation should be used:

The Caucasus Research Resource Centers. (dataset year) "[dataset name - e.g. Caucasus Barometer]". Retrieved through ODA - on {date of accessing the database here}

Please also inform us about your work where CRRC data was used. You can find our contact information in the about section.

What if I have other questions regarding ODA?

You should definitely write us - we will try our best to follow up with your questions. Please see contact information in the about section.

Analyze a single dataset

Analyze a cross-country dataset

Analyze a time-series dataset