This page has the documentation for the data files derived from the survey, for theory and everything else see the main page of the People Matching Project research program.
The first friend pairs survey ran February 2017 - February 2018 as part of the effort to develop an open-source interpersonal compatability algorithm. People were recruited on-line to sign up and take a survey and then get one of their best friends to sign up and take and the same survey, with the two surveys linked to together so that what questions best friends are most similar on can be determined.
The raw data (after some cleaning) from the survey can be downloaded at paired-survey-5-responses.csv (n=3,727) and is described by paired-survey-5-codebook.txt.
The raw data for download does not include the friendship pairing information due to severe deanonymization risk (participants in the survey would probably have little difficulty in locating their own record, and thus access the record of their partner). Summary statistics that should cover most of the value of this data are availible for download. The file paired-survey-5-cross-correlations.csv contains the Pearson correlations between each item computed across the friendships pairs. This file should be read as Person A's response to question X correlated with Person B's response to question Y. You will notice that this table is not symmetrical, X correlated with Y is not the same as Y correlated with X. These values should be theoretically the same (as who is the first member of each pair is arbitrary), but due to sampling error they differ in this data. In the case of missing data for an question, subjects were ommited from the calculation of a correlations for the questions they were missing only. Resulting in each correlation being based on ~1,300 observations. This number is lower than would be expected from the number of records in the raw data because the raw data includes surveys that were never successfully paired.