Why TalentMap Does Not Release Raw Data
At TalentMap, we often receive the question: “if we pay for the data, why won’t you release it to us? We only want to use the data to conduct research. Why do you refuse to release the data, even if we remove the names and e-mail addresses?”
We conduct each and every survey with a commitment to employees that we will protect their individual information. This is stated very clearly in all of our sales and contract documents, pre-survey communication, and on the questionnaires themselves. Individual responses are both confidential and anonymous.1 Breaking this commitment with employees is certain to be interpreted as a breach of trust; and most employees would no longer trust surveys conducted by their organization, even if managed by an external vendor.
These privacy standards are not only the “right thing to do”, but are also mandated by the codes of ethics of number of professional organizations which guide practices in survey research, including the (now defunct) Marketing Research and Intelligence Association (MRIA) in Canada, and the Council of American Survey Research Organizations (CASRO) in the US (which also has many Canadian members). We also adhere to the standards of ESOMAR (the equivalent body for Europe, which de facto sets the global standard).
We often get the question from universities and academics who would like to use the data to conduct further research. Even if we remove identifiers (e.g. e-mail address, names), it is possible to use demographic characteristics to identify very small groups, or even individuals (this is called ‘nesting’)2.
So, we cannot and will not release the raw data file, since if this were to happen, we would be breaking our covenant with employee-respondents. There is; however, one compromise we would consider and that we have used in the past with very insistent organizations, and that is that we will release the data if and only if we also remove all demographic and characteristic variables, or at least enough so that no combination of cross-tabulations will yield fewer employee-respondents than the agreed upon threshold (the industry minimum is five). However, this usually renders the data meaningless, because the whole purpose for requesting the raw data in the first place is to conduct analysis on these subgroups. Nevertheless, there have been a number of clients that have requested this.
So please understand that our unwillingness to release raw data is grounded firmly in our insistence that we must protect the bond of trust between the organization and its employees, as well as our obligations to uphold standard industry practice.
comments are just anonymous. Since they are can be read on an individual basis, they are not considered “confidential”.
2 For example, if someone requests to see the responses of only employees 25 or under, working in admin, in the CEOs office; this would yield a very small number of individuals, perhaps even a single individual thereby breaking the commitment to those few, or that individual, employee.
We conduct each and every survey with a commitment to employees that we will protect their individual information. This is stated very clearly in all of our sales and contract documents, pre-survey communication, and on the questionnaires themselves. Individual responses are both confidential and anonymous.1 Breaking this commitment with employees is certain to be interpreted as a breach of trust; and most employees would no longer trust surveys conducted by their organization, even if managed by an external vendor.
These privacy standards are not only the “right thing to do”, but are also mandated by the codes of ethics of number of professional organizations which guide practices in survey research, including the (now defunct) Marketing Research and Intelligence Association (MRIA) in Canada, and the Council of American Survey Research Organizations (CASRO) in the US (which also has many Canadian members). We also adhere to the standards of ESOMAR (the equivalent body for Europe, which de facto sets the global standard).
We often get the question from universities and academics who would like to use the data to conduct further research. Even if we remove identifiers (e.g. e-mail address, names), it is possible to use demographic characteristics to identify very small groups, or even individuals (this is called ‘nesting’)2.
So, we cannot and will not release the raw data file, since if this were to happen, we would be breaking our covenant with employee-respondents. There is; however, one compromise we would consider and that we have used in the past with very insistent organizations, and that is that we will release the data if and only if we also remove all demographic and characteristic variables, or at least enough so that no combination of cross-tabulations will yield fewer employee-respondents than the agreed upon threshold (the industry minimum is five). However, this usually renders the data meaningless, because the whole purpose for requesting the raw data in the first place is to conduct analysis on these subgroups. Nevertheless, there have been a number of clients that have requested this.
So please understand that our unwillingness to release raw data is grounded firmly in our insistence that we must protect the bond of trust between the organization and its employees, as well as our obligations to uphold standard industry practice.
comments are just anonymous. Since they are can be read on an individual basis, they are not considered “confidential”.
2 For example, if someone requests to see the responses of only employees 25 or under, working in admin, in the CEOs office; this would yield a very small number of individuals, perhaps even a single individual thereby breaking the commitment to those few, or that individual, employee.
Updated on: 02/13/2019
Thank you!